UNSWIPABLE

Over the past three years marketing has shifted for good from “buying impressions” to “earning attention.” Distribution became free, and attention became the scarcest resource. Whoever can capture and hold it wins, even on a smaller budget.

Three shifts that changed the rules

1. The algorithmic feed instead of subscriptions. Reach used to depend on your follower count. Now TikTok, Reels, Shorts and even YouTube serve content by interest, not by subscription. That means any video can blow up in front of a cold audience, and conversely, a large follower base guarantees nothing. Marketing became a competition of content, not a competition of budgets for followers.

2. Attention is measured in seconds. The average user flips through the feed faster than you read this line. The first 1–3 seconds decide whether the video gets watched at all. Hence the cult of the hook, retention and editing, and all of Part IV is devoted to them.

3. Trust beats reach. The abundance of advertising has bred “banner blindness.” What works now are formats people believe: expert content, UGC (user-generated content), reviews, personal brands, “showing it as it is.” Selling head-on to a cold audience yields less and less.

Content-led marketing: content as the engine

The model “build the product → buy ads → pump traffic” is giving way to the content-led approach: content simultaneously attracts (reach), warms up (trust) and sells (conversion). The very same video can be an entry point into the funnel, a warming tool and a purchase trigger.

This does not mean “drop paid traffic.” It means: paid traffic amplifies good content but does not replace it. Ads pointed at a weak creative and a weak offer only burn the budget faster.

The Pirate framework: AARRR as a growth map

Dave McClure’s classic model describes the customer lifecycle in five stages. It will serve us throughout the book as a common language.

Positioning is the answer to “why you, specifically?” inside the customer’s head. Until it’s formulated, all your content and ads run idle: you’re competing on price and luck.

The positioning formula

Customer profile vs value map

Jobs: what the person is trying to get done. Product/services: what you offer.

Pains: what obstructs, annoys, scares. Pain relievers: how you remove those pains. Gains: what outcome would delight them. Gain creators: how you amplify the outcome. A strong value proposition appears where your value map lands precisely on real pains and desires from the customer profile. If the product relieves a pain the customer doesn’t think about, the proposition is weak.

Three levels of differentiation

• Product-level difference: features, technology, quality. Easy to copy.
• Experience-level difference: service, speed, packaging, communication. Harder to copy.
• Meaning/identity-level difference: values, mission, “by our people, for our people.” Nearly impossible to copy.

Remarkability: a difference built into the product

There’s a level of differentiation deeper than positioning statements: a product people talk about on their own. The classic image is the purple cow: nobody notices the hundredth ordinary cow by the road, but a purple one is impossible not to discuss. The key idea: remarkability is a property of the product, not of the advertising. If there’s nothing to tell about the product, no budget will make people talk, and in a world of banner blindness (ch. 72) “safe, like everyone else” means “invisible.”

Marketing without numbers is fortune-telling. You don’t need hundreds of metrics; you need to understand a dozen key ones and how they connect. Then you can see where the funnel leaks and where to put the next dollar.

Traffic and advertising metrics

Customer value metrics

Content metrics

For short video and social platforms the key reference points are: hook rate (the share who didn’t swipe past the first seconds), retention / average watch time, engagement rate (reactions ÷ reach), share rate (reposts: the main virality signal), save rate (a usefulness signal), and follows per video. Shares and saves are exactly what algorithms weigh above likes.

If reach keeps growing and revenue doesn’t: congratulations, you’re a media company. An unprofitable one.

90% of marketing failures are not “bad creative” but addressing the wrong people about the wrong thing. Audience analysis answers three questions: who these people are, what they actually want, and where to reach them.

Demographics vs psychographics

Demographics (gender, age, geography, income) answer “who.” Psychographics (values, fears, desires, lifestyle, beliefs) answer “why they buy.” Advertising built on demographics alone sounds identical to every “woman 25–34.” Psychographics let you speak to a person in their own words.

The customer avatar

JTBD: Jobs To Be Done. People “hire” a product to do a job. The classic: “nobody wants a drill, they want a hole in the wall” (or more precisely, a shelf and a tidy home). JTBD shifts the focus from the product to the progress the person wants.

Example: “When I can’t unwind after work in the evening, I want to distract myself quickly, so that I feel better”: this is a job served equally by a game, a TV series or a walk. Those are your real competitors, all of them.

Segmentation. “One audience” is an illusion. Usually there are 3–5 segments with different jobs, pains and triggers. It helps to segment by: problem awareness level (per Eugene Schwartz: unaware → problem-aware → solution-aware → product-aware → most aware), by job, by usage situation, by “temperature” (cold/warm/hot).

Customer development: how to extract the truth

Customer development means talking to your audience before and during product creation. Rules for in-depth interviews: ask about past experience, not a hypothetical future (“Tell me how you last chose X” instead of “Would you buy Y?”).

Don’t sell and don’t suggest answers. Stay quiet and listen.

Dig down to emotions and specifics: “And how did that feel?”, “And what happened next?”

Look for pain people pay to relieve: if it’s only “would be nice,” it isn’t pain.

The avatar answers “who this is.” The journey map answers “what the person lives through step by step, and where we lose them.” It’s one of the most underrated tools: it turns an abstract “audience” into a concrete sequence of moments, emotions and touchpoints, each of which can be improved.

Three maps people confuse

What a CJM is made of

The technique for building any journey map is the same: it’s a graph/table with stages across the horizontal axis and, for each stage, layers down the vertical axis: the journey stage (e.g. awareness → search → comparison → purchase → onboarding → loyalty)

• The customer’s goal at this stage: what they want to get here
• The actions they take
• Touchpoints: site, ads, sales rep, reviews, courier, support
• Thoughts and questions: “what if they cheat me?”, “is this right for me?”
• Emotions: a mood curve, where the joy is and where the irritation is
• Barriers and pains: what stops them from moving on
• Improvement ideas: what to do to remove the barrier

The five classic customer journey stages

Choose a persona (one avatar = one map; different segments need different maps).

Describe their goal and their expectations of the process.

Gather data: interviews, analytics, screen recordings, “walk the path yourself” (especially for a new product where you have no experience yet).

Break the path into stages and touchpoints.

Plot the emotions, build a mood curve. The dips are your growth points.

Find the barriers and the “gray zones” with no owner.

Generate and score improvement ideas, put the best into action, and update the map regularly.

YJM: the path inside the product

Once the customer is already “inside” (site, app, account), the YJM kicks in. What matters here is not the emotions of choosing but the frictionlessness of action: how many steps to the goal, where the person gets stuck, where they quit. A single CJM can contain several YJMs (e.g. “registration,” “first purchase,” “repeat order”).

CJM and YJM by niche

A product hypothesis is a bet: “if we build X for segment Y, we’ll get outcome Z, because…” Mature marketing is a stream of hypotheses tested quickly and cheaply, not one big all-in bet.

How to formulate a hypothesis

A good hypothesis is falsifiable (it can be disproven) and measurable. “We’ll improve the product” is not a hypothesis. “Adding a free trial lesson will raise course-payment conversion from 3% to 5% within 2 weeks” is a hypothesis.

MVP thinking. An MVP (minimum viable product) is not a “stripped-down version” but the cheapest way to test the main hypothesis. Often the MVP isn’t a product at all: a landing page with a “buy” button (demand test), a manual service instead of automation, a pre-signup, selling before building. The goal is to learn whether people pay, before you’ve invested.

Prioritization: ICE and RICE. There are always more hypotheses than resources. To decide what to test first, use scoring.

Problem–Solution Fit and Product–Market Fit. Problem–Solution Fit means you’ve confirmed the problem is real and your solution addresses it. Product–Market Fit (PMF) means the market “pulls” the product itself: people buy, come back, recommend. PMF signals: organic growth, low churn, people would be upset if the product vanished (the Sean Ellis test: >40% of users answer “very disappointed” without it), word of mouth, sales without constant “pushing.”

Selling an expensive product head-on to a stranger is hard and costly. The product matrix (a.k.a. the product line) is a ladder of offers that gently walks a person from a free introduction to large, repeat purchases. It becomes the “skeleton” of the entire funnel.

The five elements of the classic matrix

1. Lead magnet. Free value in exchange for a contact, a follow or just attention. Important: the price isn’t always money; the person “pays” with the time to consume it, with their data, or with a repost. A good lead magnet closes one specific pain quickly and leads logically to the paid product.

2. Tripwire. An inexpensive, high-quality product whose job is not to earn but to turn a follower into a buyer. Selling the core product to someone who has already given you money (however little) is far easier. The tripwire is often sold at a steep discount or as “too good to refuse.”

3. Core product (flagship). The thing everything is built for and the main profit source. It sells easily after warming up with the lead magnet and tripwire: the person has already seen the value and trusts you.

4. Profit maximizer. Raises the average check. Two mechanisms: cross-sell (a complementary item from another category, a screen protector for a smartphone, feed styling added to an SMM service) and up-sell (a pricier version/tier). The maximizer has to be offered: people rarely take it on their own.

5. Return path and VIP. Products for repeat purchases (memberships, subscriptions, new cohorts) grow LTV. The top of the line is the VIP/premium product: expensive, exclusive, for a narrow segment that wants a turnkey result and maximum attention.

A funnel is a guided path from “saw it for the first time” to “bought it and came back.” Its job is to remove one objection at each step and make the next step as easy as possible. A good funnel turns cold traffic into money predictably, not “if you’re lucky.”

The three funnel levels: TOFU / MOFU / BOFU.

The content funnel: a wide reach at the top narrows toward the hot buyers at the bottom, each level with its own content.

The classic sales funnel

The warming (content) funnel

On social platforms and in personal brands the funnel is often “invisible”: the person doesn’t click through links, they get warmed up by content. The scheme: hooking content (reach) → useful/expert content (trust) → personal/values content (closeness) → selling content (offer). The key is balance: sell only, and you lose reach; entertain only, and you don’t sell.

Auto-funnels and closed funnels

An auto-funnel is a funnel that runs without your constant involvement: a sequence of automated touches (emails, chatbot messages, pages) triggered by a user action. A “closed” funnel is when you move a person out of the noisy open feed into a controlled space (an email list, a Telegram bot, a private channel, an account dashboard) where touches don’t depend on platform algorithms and reach nearly 100% of the audience.

Lead magnets and tripwires: what hooks

A strong lead magnet: a specific, fast result that closes one pain (“a template,” “a 7-point checklist,” “a 5-minute calculation,” “a free on-site measurement”). A weak one: “subscribe to the newsletter” with no promised value. A tripwire should lead logically to the flagship, being its “sample,” not a random cheap item.

Conversion benchmarks (reference points).

These figures are averages: check them against your own niche and calculate them on your own funnel (ch. 3, 41). They are order-of-magnitude numbers from open practice, not a guarantee: your results depend on niche, geography, traffic quality and offer, count your own.

A funnel with no holes exists. In pitch decks for investors.

These two concepts get confused constantly, and the difference is fundamental. Traffic channels bring attention. Sales channels get the product to the buyer and convert interest into a deal. The first is responsible for the flow, the second for the sale.

What’s the difference

Classifying traffic channels

How to choose channels

• Follow the audience: where they already spend time (see chapter 4).
• Account for the type of demand: existing demand (“looking to buy”) → search, marketplaces; latent demand → targeted ads, content, seeding.
• Don’t spread yourself thin: launching everything at once is a sure way to burn the budget with no conclusions. Test 1–2 channels, double down on what pays off.

Count by the chain, not by the last click.

The two main paid channels of digital advertising solve different jobs. Search ads catch those already looking. Social/targeted ads find those who aren’t looking yet but fit the profile. Understanding the difference saves budgets.

The key difference

When to use which

There’s existing demand (people google your product) → start with search: fast sales, clear payback.

Demand has to be created, the product is new or emotional → targeted ads + content to warm up.

Catch those who visited the site → remarketing/retargeting (in both search and social): usually the cheapest and highest-converting traffic.

Display advertising (banners, video) to initiate demand and build awareness for major launches.

Other important channels

Seeding, native integrations in themed channels/communities (text+media): cheap and with the channel audience’s trust. Telegram Ads, targeted ads inside channels; developing fast, with a falling barrier to entry. Blogger/influencer ads: the trust of an “opinion leader.” CPA/affiliate marketing, pay per action, with the risk on the partner’s side.

A hook is the barb that decides whether the thumb stops on your video or swipes on. In a feed where content changes faster than thought, the hook is the single most important second in all of marketing. A weak hook nullifies even a brilliant idea: nobody will see it.

What makes a hook strong

A hook works when it creates a gap, between what the person expected to see in the feed and what they actually saw. The brain can’t swipe past an unclosed question, a threat, a promise or something strange. Three levels of the hook work at once: the visual hook, the first frame stops the eye (movement, an unusual object, a close-up, on-screen text, a “pattern interrupt”).

The meaning hook: the first phrase promises value, intrigue or conflict.

The audio hook: the first words/sound/track catch the ear.

Hook categories

Hook formulas (plug in your own).

Hooks by platform

Reels/TikTok/Shorts: visual+text+sound in the first second, high tempo, the hook often duplicated as on-screen text. YouTube (long): the hook is the first 15–30 seconds + a strong thumbnail and title (they’re the “hook from outside”); you need to promise what the person gets and create a loop for the whole video. Posts/text: the hook is the first line, visible before “show more.”

You have three seconds. You already spent two on “Hey everyone!”

An angle is the perspective from which you show the same product. There’s one product, but dozens of angles. Switching the angle is exactly what lets you make hundreds of different videos about the same thing and find the take that “fires” for a specific segment.

Why you need angles

Different people buy the same thing for different reasons. The same English course can be sold through “career and money,” “travel,” “watching shows in the original,” “the fear of looking stupid,” “emigration.” Each angle hooks its own segment and its own pain. Without changing the product, you multiply your reach and find the combos that work.

How to find fresh angles

• Change the segment: one product → different avatars → different pains.
• Change the emotion: the same fact through fear, through humor, through inspiration.
• Change the format of truth: advice → anti-advice; success → failure; how-to → a breakdown of a mistake.
• Take them from comments and objections: every frequent question = a new angle.

The hook brings the viewer. Retention decides whether they watch to the end and whether the algorithms “reward” the video with reach. Watch-through (retention) is the number-one metric for short video: platforms promote what holds people.

The retention curve

The retention curve shows what percentage of viewers stays at each second. Reading it is simple: sharp drops = places where people got bored or confused. The first “step” after the hook is the biggest churn (those the hook misled or failed to catch). The task is to make the curve as flat as possible and add “hills” of interest.

The break happens where you haven’t started saying the main thing yet.

The difference in the first seconds is multiplied in reach not by percentages but many times over: the algorithm moves what people hold on to.

Retention techniques

Any retaining structure is promise → build-up → payoff. Don’t give it all at once: portion out the value, lead to a climax, pay off the hook’s promise at the very end (or nearly), so the person watches through.

This chapter is a working one. Below are ready storyboards (script frameworks) with second-by-second timing. Take a template, plug in your topic and shoot. All of them are assembled from proven short-video structures.

The universal short-video structure

Template 1. Hook → Value → CTA

Template 2. PAS: Problem → Agitate → Solution

Template 3. Transformation (before → after)

Template 4. Listicle (N ways / mistakes)

Template 5. Tutorial (how to)

Template 6. Story-based

The long-video framework (YouTube, 8–15 min).

Designing from metrics: the video control panel

A storyboard is not a creative ritual but a control panel for metrics. Every element of the plan exists to move a specific number, and when a number drops, you know which part of the video to fix instead of “shooting differently in general.”

Carousel storyboard: two ready templates

Everything above was about video, but carousels have their own dramaturgy: swiping instead of retention. Two frameworks that cover most needs.

A “shoot it and see” script has a name: “it won’t.”

A take (from the English “take,” an opinion or point of view) is a clearly stated position on a topic in your niche. Takes build expert authority, spark discussion and separate you from the “faceless” accounts that just retell others’ material.

Why takes work

Neutral content is easy to forget. A position is remembered, provokes agreement or argument, and argument is reach and engagement. A take shows there’s a person with experience and a viewpoint behind the content, not a feed of facts. Thought leadership, the status of someone whose opinion is heeded, is born precisely from takes.

Hot and cold takes

How to voice opinions safely and strongly

Lean on experience and data: then even a controversial take is defensible.

Attack ideas, not people, criticize the approach, don’t make it personal.

Be ready to discuss in the comments, that’s part of the mechanic.

Don’t provoke for provocation’s sake, cheap hype hits your reputation.

Copywriting isn’t “pretty words” but the engineering of persuasion. Proven formulas give structure: they lead the reader from attention to action without missing a thing. Master four and you’ll cover 90% of tasks.

The four base formulas

Headlines by the 4U formula

A strong headline is at once: useful (what I get), urgent (why now), unique (how it differs), ultra-specific (numbers, timeframes). “Lose weight” → “Drop 4 kg in 30 days without a gym, using a method that works after 40.”

Bullets that sell

A bullet is a mini-promise with intrigue: not a “feature” but a “benefit + curiosity.” Not “14 lessons” but “Lesson 7, after which clients message you first.” Feature → benefit → emotion.

The call to action (CTA).

Two lessons from the direct-response classics

Lesson one: the headline is 80% of the money. Headlines are read many times more often than the body copy: once you’ve written the headline, you’ve spent most of the advertising budget, and only a minority will see the rest. Hence the discipline: spend as much time on the headline as on the whole text; put a specific benefit and facts in it, not wit for wit’s sake. And the principle of respect, formulated back in the print-advertising era: the consumer isn’t an idiot, don’t try to win them with empty bombast, give them information they can decide on. Long copy sells no worse than short, on one condition: every line is interesting to the reader, not the author.

Lesson two: the slippery slide. The first sentence of ad copy has one single job, to get the second one read. The second’s job is the third. The text is built like a slide you can’t get off: a short intriguing opening, “seeds of curiosity” at the ends of paragraphs (“and here’s what happened next”), a build-up. The practical corollary: don’t load the main argument into the first paragraph, draw them in first. And sell the concept, not the object: not “a watch with sapphire glass” but “the last watch you’ll ever buy.” A separate technique, name the product’s weakness before the customer does and close it right away: honesty about a flaw buys trust in all the strengths (this echoes the trust objection, ch. 49).

An offer is neither the price nor the product but the ratio of value to cost in the customer’s head. A strong offer sells even with average marketing; a weak one won’t be saved by the best traffic. As practitioners say: make an offer so good that refusing would be foolish.

The offer value formula

Value stacking. Instead of “the product for $N,” assemble a stack: the core product + bonuses, each with its own value. The total perceived value of the stack should exceed the price several times over. Then, the price, against which the stack looks like “too good a deal.”

Five offer amplifiers.

• Guarantee: transferring the risk onto you (“doesn’t work, we refund” / “we’ll get you to the result”).
• Scarcity: a limit on quantity (spots, batches).
• Urgency: a deadline, a price window.
• Bonuses: they close specific objections.
• Naming: the offer has a name and “packaging.”

Price psychology

A 90% discount on something nobody needs still leaves something nobody needs.

Strategy answers the question “why each piece of content exists and how they add up to growth.” Without it you produce scattered videos; with it, a system where every element works the funnel and reinforces the rest.

Content pillars (recurring themes).

Pillars, 3–5 permanent thematic supports for the blog. They create recognizability and take away the “what do I post today” problem. The classic balance by content goal: Content type, task, share.

The content matrix.

Matrix = pillars × formats × angles. The intersections yield dozens of ready ideas. For example, the pillar “mistakes” × the format “listicle” × the angle “against the grain” = “5 trendy tips that ruin everything.” The matrix turns “no ideas” into “ideas for months.”

Atomization and repurposing (COPE).

The COPE principle, Create Once, Publish Everywhere: one “big” piece is broken into many units for different platforms. One interview/livestream → dozens of Reels, a series of posts, an article, an email, a carousel, quotes. This multiplies the return on a single idea and solves the volume problem.

Distribution and multichannel

Creating content is half the job; the other half, distributing it. Publish on several platforms (adapting to each), link them into a funnel, move the audience into your “own” channels (email, a bot) so you don’t depend on a single algorithm.

Growth loops and virality

A growth loop is a mechanic where the result of an action feeds the next acquisition (a referral loop, a UGC loop, a content loop). Unlike a funnel (which is linear), a loop is self-reinforcing. Virality lives on emotion and practical value: people share what gives them status, helps others, or triggers a strong emotion. The main signal, share rate.

UGC and influencers

UGC (user-generated content), content from the perspective of customers/creators: it’s perceived as “real” and works great in ads too. Influencer marketing, access to a ready loyal audience; choose by relevance and engagement, not just follower count. Micro-bloggers often deliver a warmer response.

Email and messengers

Your own base (email, a Telegram bot), an asset that doesn’t depend on algorithms and gives the cheapest repeat contact. Use it for warming, launches, reactivation and retention. This is exactly where the audience from reach platforms should “flow.”

In a world where AI produces tons of faceless content, personality has become the main uncopyable asset. A personal brand is the reason one expert is chosen out of a thousand identical ones, and the reason to forgive imperfections that a faceless company would never be forgiven.

What a personal brand is made of

• Expertise: what you’re genuinely strong at (one clear zone, not “everything”).
• Point of view: your takes and principles (see ch. 15).
• Story: the journey, the failures, the transformation, the things people identify with.
• Character and delivery: manner, humor, values, “how” you speak.
• Visual and verbal style: recognizable from the first frame/line.

Tone of voice (the brand’s voice).

Tone of voice is a consistent manner of communication: how the brand/expert sounds across every touchpoint. It makes you recognizable and consistent. Define it along axes: Axis, Pole A ↔ Pole B.

Choose a point on each scale and hold to it across every touchpoint.

The final layer of the system, measure, test and don’t fall behind. Marketing has stopped being an art of guesswork: the winner is whoever tests hypotheses fastest and spots a platform shift earliest.

What and how to measure

Set up end-to-end analytics: from ad impression to payment (pixels, UTM tags, CRM). Look not at “likes in general” but at metrics tied to money and watch-through: CAC, LTV, ROMI, step-by-step funnel conversions, hook rate, watch-through, saves and shares. The 2026 rule: reach for reach’s sake doesn’t work, count ROI and LTV, and among content metrics prioritize saves and reposts over likes.

A/B testing.

What to test first (in descending order of impact): offer → audience → hook/creative → landing page → CTA. Often changing the hook yields a bigger lift than remaking the whole video.

AI in marketing: where it genuinely helps

In 2025–2026, AI covers volume and routine: generating hook and headline variants, script drafts, transcribing streams, clipping and subtitles, ideas and research, personalization, chat assistants and AI agents for standard support/sales scenarios.

Key trends 2025–2026

Despite platform differences, the recommendation logic is similar: the system shows a video to a small test audience and watches the signals, watch-through, replays, saves, reposts, comments, click-throughs. If the signals are strong, reach expands in waves. Hence the practical takeaway: optimize for retention and “shareability,” not for likes. Publish regularly (one viral hit is luck, a system is consistency) and use current formats/tracks the platform is promoting right now.

Testing is a century older than the internet

The A/B test seems like an invention of the digital age, but the method is over a century old: print-era advertisers embedded coupons with different codes and, from the returns, knew exactly which headline, which offer and which price brought orders. From that same era come three principles that haven’t aged a day. Advertising is salesmanship multiplied by print (today, by the screen): anything you wouldn’t say to a live buyer shouldn’t go into the ad. Specifics beat boasting: “cuts consumption by 23%” is verifiable and therefore persuasive, “best in class” is noise that gets filtered out (ch. 72). And the third: almost any question about advertising is cheaper to settle with a test than with an argument, opinions, including your own, always lose to the coupon (today, the metric).

An A/B test: a civilized way to lose an argument with reality. In writing.

What sells isn’t “a game” but emotion, challenge and belonging to a community. The key points, the first trailer (attention), the first 15 minutes (retaining the player) and life after launch (LTV via seasons and purchases). Audience and pains.

Segments: casuals (want to unwind), hardcore (want challenge and mastery), collectors/series fans, streamers/creators. Pains: “boring,” “nothing to play with friends,” “hate to waste money on a bad game.” Job: emotion, status, socializing, escapism.

Offer and line

Demo / free-to-play (lead magnet) → a starter pack or the base game (tripwire/flagship) → in-game purchases, a battle pass, DLC (maximizer) → seasons, new chapters (return). Amplifiers: early access with bonuses, “refund within 2 hours.”

Funnel and channels

TOFU: trailers, gameplay, streamer/creator content, targeted ads. MOFU: a Steam page, wishlists, a demo, a Discord community. BOFU: launch with a first-week discount, reviews. Sales: Steam, console stores, mobile stores.

People buy not “a course” but a transformation: a new profession, income, confidence, a result. The main battle: for trust (“will it work for me specifically”) and for getting them to the first result (otherwise, churn and bad reviews).

Audience and pains

Segments: complete beginners, career-changers, practitioners who “want to level up,” hobbyists. Pains: “I study but see no result,” “afraid I won’t manage / won’t have time,” “too much free information but no system.” Job: income, status, self-realization.

Offer and line

A free webinar / checklist / mini-lesson (lead magnet) → a low-priced mini-course (tripwire) → the core course (flagship) → mentorship / a one-on-one tier (maximizer) → the next step, an alumni club (return). Amplifier: “we’ll get you to your first result or refund you,” work reviews, a community.

Funnel and channels

TOFU: Reels/Shorts with fast value, targeted ads, bloggers. MOFU: lead magnet → email/bot warming → webinar. BOFU: an offer with a deadline, student reviews, case studies. Sales: landing + bot/sales team. A closed auto-funnel in a bot is a must.

Dramaturgy of the free part: the series effect

In education the free part isn’t “a slice of the product” but the first episode of a series. Its job is not to tell everything but to make it impossible not to want the sequel. The same mechanic that keeps people glued to screens works here: cutting off at the most interesting moment.

Three more mechanics from the same logic. Splitting: a big course is cut into modules and short “boosters”, targeted add-ons sold separately that lower the first-purchase barrier (tripwire logic, ch. 16). Splitting by level: an explicit “beginner / experienced / pro” sorting raises both conversion and satisfaction, people don’t pay for what they already know. Product on demand: requests and questions with no ready course are collected into a suggestions section; once enough similar requests pile up, a new course is born from a mass one-off consultation, demand confirmed with money before production (this echoes ch. 27 on research and ch. 66). Meanwhile alumni are sold continuations and updates, an “add-on course for graduates” is almost always cheaper to promote than a new cold customer (Retention, ch. 41–43).

In services what sells is removing a problem without hassle or risk. The client can’t see the result in advance, so trust, response speed and social proof decide it. Demand is often existing, which means context is strong.

Audience and pains

Segments: by urgency (urgent / planning), by budget, by experience (“got burned by contractors before”). Pains: “they’ll do it badly/expensively/miss the deadline,” “I don’t want to figure it out myself,” “how do I choose and not get burned.” Job: solve the problem reliably and calmly.

Offer and line

A free audit / quote / consultation (lead magnet) → a trial service or a diagnostic (tripwire) → the core package (flagship) → add-on services, cross-sell (maximizer) → a retainer / ongoing support (return). Amplifiers: a KPI/deadline guarantee, fixed price, cases and reviews, “we work under contract.”

Funnel and channels

TOFU: expert content, before/after, reviews. MOFU: a free audit/quote, case studies. BOFU: a proposal, a contract, guarantees. Traffic channels: search (hot demand), geo-services (maps for local businesses), review sites, referrals. Critical: speed of response to an inquiry.

Here the product and the seller are one person. The audience buys access to the personality, the expertise and the community. Monetization comes after trust, and trust is built through regular content and honesty.

Audience and pains

Segments: “I want the same” (followers), “I need a result/answers” (clients), “along for the ride/inspiration” (observers). Pains: “I don’t know whom to believe,” “I want to grow in the author’s field,” “I need a mentor/community.” Job: inspiration, answers, belonging, transformation.

Offer and line

Free content (lead magnet) → an inexpensive product: a guide, workbook, mini-course (tripwire) → a course/intensive (flagship) → mentorship (maximizer) → a mastermind/subscription club (return). Amplifiers: a limited number of spots, personal reviews, a closed community.

Funnel and channels

A warming content funnel: reach (Reels/collabs) → usefulness/expertise → personal/values → offer. Move the audience into a closed bot/channel for launches. Channels: Reels/Shorts, collaborations, seeding, podcasts; sales, a bot, a landing, personal selling.

An expensive, rare, emotionally significant “for years” purchase. A long decision cycle, a high cost of error for the client. What sells is reliability, quality and peace of mind. Demand is often existing, so context works, but warming and trust decide it.

Audience and pains

Segments: building/moving (a planned purchase), replacing the old (a replacement), fitting out a country house/sauna. Pains: “I’ll overpay,” “they’ll do it badly/late,” “it won’t fit the size/style,” “it’ll fall apart within a year.” Job: reliability, comfort and savings for years, beauty, status.

Offer and line

A free measurement / 3D design / quote (lead magnet) → an inexpensive accessory or a standard item (tripwire) → a made-to-order core product (flagship) → components, installation, extra sections, cross-sell (maximizer) → service, warranty support, new items (return). Amplifiers: an extended warranty (years), free delivery/installation, “doesn’t fit, we’ll replace it,” installment plans.

Funnel and channels

TOFU: “how it’s made” content, stress tests, reviews, interest-based targeting (sauna, construction, renovation). MOFU: measurement/quote, catalog, case studies, consultation; warming via messenger/retargeting (the cycle is long!). BOFU: a contract, a deposit, guarantees. Traffic channels: search on hot demand (“buy a sauna stove”), marketplaces, specialized sites/forums, geo. Sales: site+manager, a showroom, dealers, a marketplace.

A niche of high distrust, complexity and strict advertising limits. Here the winner is whoever explains simply, behaves transparently and honestly warns about the risks. Trust and security, the main currency.

Audience and pains

Segments: beginners (“I want to understand it and try”), active users, investors/traders, developers. Pains: “it’s hard and scary,” “afraid to lose it/get scammed,” “don’t want to miss out,” “whom to trust in this chaos.” Job: earn/preserve, understand, not lose, not fall behind.

Offer and line

Free educational content, an airdrop (lead magnet) → inexpensive access/an analytics subscription (tripwire) → the core product: a service/platform (flagship) → a premium tier, extra tools (maximizer) → retention through new features and community (return). Amplifiers: transparency, a security audit, reputation, a trial period. No promises of returns.

Funnel and channels

TOFU: educational/news content, X (Twitter), Telegram, influencers. MOFU: a Telegram channel/bot, analytics, community, an airdrop. BOFU: registration/first transaction (minimize friction and fear as much as possible). Sales: the platform/app itself. Important: account for platforms’ strict limits on crypto advertising and your jurisdiction’s regulation.

In crypto, “guaranteed returns” guarantees exactly one thing: returns for the author of the scheme.

There are two types of research, and confusing them is expensive. Qualitative answers “why” and gives depth. Quantitative answers “how much” and gives scale. Strong decisions come from using both, in the right order.

What’s the difference.

The right order: “why” first, then “how much.”

You start with qualitative: interviews and observation give hypotheses, pains and the customer’s exact phrasing. Then quantitative checks how widespread it is and whether it’s worth investing in. The reverse order (jumping straight to a survey) often produces useless numbers: you measure what you made up yourself, not what matters to people.

The most underrated source: your sales team

Before commissioning research, talk to those who already speak with customers every day: your salespeople. They know exactly what the product is loved for, what it’s criticized for, what words people use to ask about it, and where deals fall apart. An hour with two salespeople often yields more than a week of desk analytics. Four questions for that conversation: which three objections do you hear most often; what do buyers praise us for; describe the best client, the one who’s easy and profitable; which customer words do you remember verbatim. The last one is gold: it’s ready-made language for offers and headlines (ch. 4, 16).

“Everyone” is not an audience. Segmentation splits the market into groups with different needs, and the awareness ladder shows how “ripe” a person is to buy. Addressing a cold newcomer and a ready-to-buy person with the same words means losing both.

Segmentation criteria

Sherrington’s 5W method

Five questions to describe a segment: What (what we sell / the product type), Who (who buys), Why (why, what motivation/job), When (when the need arises), Where (where they decide and buy). Different answers = different segments.

The awareness ladder (Ben Hunt).

The idea is simple: a person doesn’t arrive “ready to buy,” they climb the rungs from “no problem” to “buying.” At each rung they need their own content and message. Head-on advertising at the bottom rung doesn’t work, the person hasn’t recognized the need yet.

The awareness ladder: the higher the rung, the “hotter” the person and the more direct the offer should be.

The 5 stages of awareness (Eugene Schwartz).

A similar scale for copywriting: unaware → problem-aware → solution-aware → product-aware → most aware. The higher the awareness, the shorter the path to the offer. The cold need a story and a problem; the hot, the price and terms.

5W: segmentation in five questions

When there are too many segmentation criteria, the rough and fast 5W method helps. Five questions: What (what we sell: which exact product and variant), Who (who buys: the segment), Why (why they buy: motive and pain), When (when they buy: occasion, season, moment), Where (where they buy and where they search: the touchpoint). Run every product in your line through the five questions, and you have a rough segment map in half an hour, which you can then deepen with avatars and JTBD (ch. 4, 40).

Acquiring a customer is expensive. So the ability to sell them more and sell them again is often the main profit lever. Three upselling techniques, upsell, downsell and cross-sell, work at the moment of purchase and after it, raising the average check and LTV without spending on new traffic.

The three techniques, what’s the difference

Where this lives in the product matrix

Upselling is the “profit maximizer” from chapter 7, built into the funnel. The matrix walks a person up the rungs of value: from a free introduction to large, repeat purchases. Upsell and cross-sell raise the check at the flagship stage; the downsell saves the deal; the return path brings them back.

The product matrix as a funnel: the volume narrows, value and check grow. The figures on the right, approximate conversion. Rules of healthy upselling.

An upsell should reinforce the main purchase, not look like a “hard sell.”

The best moment for an upsell/cross-sell, right after the decision to buy (trust is at its peak).

A downsell isn’t a “fire sale” but a way not to lose the customer and close them later.

Down the ladder there are fewer people, up the ladder the price and value grow.

No more than 1–2 additional offers at a time, or you get choice paralysis.

The line “Want fries with that?” has earned more than all your stories combined. Make your peace with it and implement it.

In chapter 9 we separated traffic channels from sales channels. Now let’s file the traffic itself neatly: paid, quasi-free and partner, and break down each tool so you understand what to switch on, when and why.

A map of traffic channels. Quasi-free ones require not money but time and content; partner ones, pay for results. Three groups of channels.

Key tools and when to use them

Traffic channels

B2C, selling to people (business to consumer). B2B, selling to businesses (business to business). The logic of funnels and content is similar, but the motives, the deal cycle and the decision participants differ so much that a head-on strategy from one world fails in the other.

Key differences

If you’re a marketer or an entrepreneur getting into this, it’s important to understand not only “what to do” but “in what order.” Below, the professional work sequence: from first contact with a client (or with your own project) to optimization by the numbers. It turns chaos into a manageable process.

The workflow: a linear path from introduction to execution, plus a constant “analytics → optimization” loop. The steps in order.

Defending the strategy: five slides, but really two

A strategy that isn’t approved is worth zero, no matter how much research is in it. And it’s almost always approved by someone else: an owner, a director, a client. So the strategy has to be sold, and what sells it isn’t the thickness of the document but five slides: problems and goals; the target audience briefly (so the listener sees you understand their customer); what exactly we’ll do; how much it costs; how and when it pays back. Everything else, a hundred pages of audits and tables, lives in the appendix for questions.

“Want to earn $10,000: invest $5,000.” There’s a healthy grain in this: to earn, you must invest, and you should start with a test slice. But as a formula it’s dangerous: it ignores margin, acquisition cost and conversion. Let’s figure out how to budget like a grown-up while keeping it simple.

Why “invest half” doesn’t work

How much you’ll earn from a dollar invested depends not on your wishes but on unit economics: the margin, the cost to acquire a customer (CAC), the conversion, how much a customer brings over their whole lifetime (LTV). If marketing is set up badly, you can pour in even 100% of revenue with no result. If it’s set up well, $5,000 can bring $10,000, or $50,000. “Half” is not a calculation but a blind bet.

Four ways to calculate a budget

1. Percent of revenue (a quick reference).

2. From the business goal (the objective-and-task method).

3. From unit economics (the CAC ceiling).

Margin sets the “ceiling” on acquisition cost. A healthy reference: LTV / CAC ≥ 3, meaning the maximum allowable CAC ≈ LTV ÷ 3. The higher the margin and LTV, the more you can spend on acquisition and the more aggressively you can scale. Low margin = tight CAC control.

4. Competitive parity.

Gauging “how much competitors spend” as a lower-bound reference so you don’t lose on share of voice. Useful as a check, not as a basis, a competitor may have different economics.

The marketer: a head chef, not the whole restaurant

A special case of the budget mistake: hiring one marketer and expecting them to “build our entire internet marketing.” A marketer is a head chef: they know the recipe and the ingredients, but the kitchen cooks the dish. Around the tasks stand separate professions: an analyst, a traffic specialist, a copywriter, a designer, an editor, a developer, and they aren’t interchangeable. One person can cover two or three roles at a decent level, but not eight. The practical budget takeaway: either budget for a team or contractors around the marketer (ch. 66, 68), or honestly cut expectations down to the two or three roles they can actually handle. Expecting an orchestra from one musician is the cheapest option on paper and the most expensive in reality.

A “spend whatever we can spare” budget usually ends at a number you’ll very much regret sparing.

A launch is when you don’t sell “a little all the time” but concentrate the audience’s attention into a narrow window and make a strong offer at the peak of interest. A well-run launch can deliver more in a week than months of sluggish sales. Below, the universal anatomy and a day-by-day example.

Three phases of a launch (Jeff Walker’s formula).

The psychology of warming up

Warming leads the audience along an emotional arc: first we show the problem (the person recognizes themselves), then the possibility (“it can be different”), then the proof (cases, results), and only at the end, the offer. Decision accelerators at the finale: scarcity (limited spots), urgency (a deadline), early-bird bonuses, the FOMO effect (“everyone else is already in”).

Example of a 10-day launch: warming, selling, final scarcity. The days are illustrative and adapt to the niche. A day-by-day launch example.

Launch budget and energy: don’t spread it evenly

A common mistake is to split the launch budget and effort evenly across all the days of the window. The audience’s attention and the platforms’ algorithms work differently: they respond to spikes. A launch wins when resources are concentrated at three points, the announcement, the sales opening and the finale, with a supporting rhythm in between.

Two techniques that amplify the launch window and cost almost nothing. First, a retrospective of past launches: results, reviews and breakdowns of the previous cohort or event are published during the warming days; this is both proof and a reminder to the old audience. Second, the finale ladder: the frequency of deadline reminders escalates as it approaches, a week out, a day out, a few hours out, an hour out. It’s important that the deadline is real: “there are still reserve spots” after “no spots left” works once and burns trust forever (ch. 70, rule 19 in ch. 75).

Most blown budgets and disappointments come not from “a bad market” but from a few typical misconceptions at the start. Let’s break down the main ones, beginning with the most expensive, and at the end critically assess a couple of “facts” that get passed from article to article.

Myth #1: “Why pay for ads when I’ll write 10 posts myself and something will take off?”

This is perhaps the most common and expensive beginner mistake. Let’s see why betting “on organic and luck alone” is unreliable: Survivorship bias. We see the videos that “took off” and don’t see the millions that got nothing. One viral post from an acquaintance is not a system but a survivor out of a thousand.

Algorithm dependence. You don’t control organic reach: the platform decides whom to show it to. Today it “took off,” tomorrow, silence, and a business needs predictable sales.

Reach ≠ sales. Even a viral video to the wrong audience brings no money. Advertising lets you show the offer to exactly those who’ll buy.

Your time isn’t free. “10 posts” is hours of work. If you count the cost of your time, “free” organic is often more expensive than ads.

Organic is hard to scale and predict. Advertising gives leverage: invest more, get more (as long as the economics add up). Organic can’t be steered that way.

Other common misconceptions

Critical audit: careful with “facts”

A number of striking claims circulate in marketing literature that deserve a critical eye. We deliberately soften them in this guide too.

“Word of mouth will keep us fed,” said the business you’ve never heard of.

Marketing is extracting profit from satisfying a need. Its job: to attract new customers by promising superior value, and to keep existing ones by delivering that value. Everything else is technique.

The five stages of the marketing process

Notice: money appears only at the fifth step. The first four are an investment in value for the customer.

Need, want, demand

A need, a basic lack (food, safety, connection). It can’t be created.

A want: a need shaped by culture and personality. A hungry American thinks of a burger; a Japanese person, of rice.

A demand, a want backed by money.

Value and satisfaction

The 4 Ps: the marketing mix

The customer and value first, then the 4 Ps, not the other way around. If the market doesn’t need the product, no tuning of price and advertising will save it.

Five marketing management concepts

Before launching ads, it’s worth answering: why the business exists, which directions feed it and which eat it, and what’s happening around us. This is a level above funnels, but it’s exactly what determines where to point them.

Mission: market-based, not production-based

A weak formulation: “we make furniture.” A strong one, through the customer’s need. Nike doesn’t “make sneakers,” it brings inspiration to every athlete; Disney doesn’t “build parks,” it creates a world of fantasy. Test yourself with four questions: what is our business? who is our customer? what do they value? what should our business become?

The BCG matrix: where to invest

The Ansoff grid: growth directions

SWOT, internal (strengths/weaknesses) and external (opportunities/threats). PEST helps fill the external block: politics, economics, social factors, technology.

The management cycle

Analysis → Planning → Execution → Control. Control means: set a goal → measure the actual → find the cause of the gap → correct. Without the last step the cycle turns into reporting for reporting’s sake.

The business model canvas: the whole company on one sheet

Before a SWOT, a simpler and faster tool helps: the business model canvas, nine blocks that let you break down any company, your own or a client’s, in an hour. Take a sheet and answer:

People rarely buy rationally and almost never for the reasons they state out loud. And no product sells forever. Understanding these two things saves budgets.

What influences the choice

Four types of buying behavior

Diffusion of innovation: who buys first

A novelty is adopted in waves: innovators → early adopters → early majority → late majority → laggards. At the start it’s pointless to hit the mass; work with the innovators and early adopters, they shape everyone else’s opinion.

The speed of adoption depends on five properties: relative advantage (how much better), compatibility (does it fit existing habits), complexity (is it easy to grasp), divisibility (can you try it cheaply), observability (is the result visible to others). Improve any of them and you speed up sales.

The product life cycle

A brand is a promise. A strong brand is a kept promise. And customers aren’t a uniform mass: with some you build relationships over years, from others it’s enough to take the maximum here and now, and from a third group you’re better off walking away.

What makes a name a brand

• Hints at the product’s benefit or quality
• Is easy to pronounce, recognize and remember
• Is distinctive, doesn’t blur into competitors
• Translates into other languages without losing meaning
• Is legally protectable (can be registered)

Four brand-development strategies

Customer equity: four types of customers

A customer for life: count the stream, not the check

A classic retail calculation: a car buyer isn’t a single deal but decades of purchases, service and referrals; their lifetime value is measured in the hundreds of thousands of dollars. Once you see this stream in a person, the economics of decisions flip: fixing a minor problem for free, one that formally “isn’t a warranty case,” stops being generosity and becomes cheap retention of an asset. Three rules from the same school. Promise less, do more: a quoted price slightly above the expected final one creates a pleasant surprise instead of an unpleasant one. Systems, not smiles: politeness doesn’t save you from recurring failures, service is fixed with processes, not with friendliness training. Ask customers rather than guess: three questions at the checkout give more than expensive research (ch. 27). All of this is the practical substance of Retention from ch. 41: a retained customer costs many times less than an acquired one, but only if someone designed the retention.

“Our audience is men 25–45”, a phrase that sounds professional and means almost nothing. Demographics describe who a person is. But a purchase is determined not by who they are, but by what job they’re solving right now.

Why demographics break down

The same person walks into a coffee shop in the morning for speed (“a pick-me-up on the way”) and in the evening for atmosphere (“to hold a meeting”). Those are different jobs, different choice criteria and different competitors, even though it’s the same person of the same age. Segmenting by passport, you average different situations into a meaningless mean.

Substitutes: who your real competitor is

A competitor isn’t only whoever sells the same thing. A competitor is any way to get the same job done, and any expense category a person pulls the money from.

Context decides more than the product

A model that breaks the whole business into five stages and instantly shows where exactly growth is breaking. It’s called “pirate” because of the acronym: AARRR, acquisition, activation, retention, referral, revenue.

AARRR: five stages and each one’s metrics. Growth is capped by the narrowest link.

1. Acquisition

The most common point of failure. People run a blog with 200 followers for years and wonder why there are no sales. If few people enter at the top, a lot of money physically can’t come out at the bottom, arithmetic, not mysticism. You can’t dig a pond with the same spade you dug a small hole with.

2. Activation: the “aha moment”

The moment a person understands why they need this. Before it, they’re nobody’s customer.

The bottleneck rule

However many people entered at the top, that’s the ceiling at the bottom. Fix the weakest stage, not your favorite one.

3. Retention

4. Referral and 5. Revenue

People recommend not “the good” but what’s interesting to talk about: the unusual, the beyond-expectations, the thing that gives the teller status. “10% for a friend” referral programs usually motivate no one: the small have nothing to offer, and for the large it’s not worth it.

At the revenue stage we count: average check, revenue per customer, revenue per seat/area, who the decision-maker is (who really decides on the purchase), and how to raise the check through upselling and bundling (chapter 42).

Three quick levers: understand what you’re “the -est” at, choose a magnet product, and learn to raise the check with a single line at the register.

The EST model: what you’re “the -est” at

English superlatives end in -est. The idea: in the customer’s mind, whoever is the “most” at something wins. Choose your category and defend it.

The loss-leader

The product people come for isn’t necessarily the one you earn on. Consoles sell nearly at cost; the money is made on games and subscriptions. A coffee shop earns on desserts, an electronics store, on accessories and service.

Bundling and upsell scripts

The fastest way to raise the average check is not to find new people but to ask the right question of those already paying.

Most “competitor analyses” are a list of links and a sigh, “theirs looks nicer.” A proper analysis answers two questions: what customers love them for, and where the gap is that we’ll walk into.

The main source: reviews

The comparison table

Reconnaissance of others’ traffic: what to look with

Everything above in this chapter can be done by eye. But competitors have numbers you can’t see by eye: where their site traffic comes from, which videos carry the channel, which creatives they’re running in ads right now. For each type of reconnaissance there’s a dedicated tool, and almost every one has a free tier that’s enough for a small business.

Peeking at competitors isn’t shameful. Copying without looking is.

An expert’s most common pain, “I don’t know what to film about.” It’s solved with a tool from TRIZ (the theory of inventive problem-solving) by Genrikh Altshuller, adapted for content. Take one topic, get nine different content units, each with a different job.

How the matrix works

Down the vertical, three system levels; across the horizontal, three time frames. Supersystem, what larger phenomenon your topic is part of. Gives reach. System, the topic itself, the object as it is. Gives follows. Subsystem: what the topic is made of, the details and stages. Gives sales.

One topic × three levels × three time frames = nine ready ideas with different jobs.

Content gets views but brings no customers, the most common complaint. The cause is usually one: the content units aren’t linked into a funnel. A lead magnet sells nothing on its own if the person hasn’t “ripened” to it and if nothing happens after it.

The five stages of the content funnel

The strong-lead-magnet formula

Variability: a funnel, not a straight line

A linear funnel (“saw it → read it → left an inquiry → bought”) is a fantasy. In reality a person drops off at every step. A good funnel anticipates in advance that they WON’T take the next step.

Speak the customer’s language

Experts often write to make themselves feel smart: terminology, precision, professional rigor. But content wins when it makes the reader feel smart. Gather the “raw material” (real phrasings of the pain from forums, reviews and comments) and speak in them. Colleagues may scold you for oversimplifying. Customers will understand and come.

There are many myths around algorithms: “the magic posting time,” “the right hashtags,” “secret sounds.” The real mechanics are simpler and harsher, and understanding them changes your approach to content forever.

The platform’s single goal

Keep the person in the app as long as possible: more time (more ads) more money. So the algorithm is a matchmaker: it pairs content to a person. Want to be shown, help it make a good match.

Three algorithms on one platform

Saying “the algorithm” in the singular is already a mistake. Within one social network, formats are ranked by different systems, and the head of Instagram has publicly confirmed this: the feed, short video, stories and recommendations each have separate ranking, and the main signals are watch time, likes-per-reach and shares-per-reach.

What happens after publishing

The promotion mechanics: digital fingerprint → audience prediction → test sample → waves of impressions. Two metaphors instead of a hundred instructions.

Rooms. Picture a building where each room holds an audience united by an interest. There’s a “social media marketing” room (a million people). There’s a “rare hobby” room (three and a half, and they drop by on Tuesdays). Choosing your topic, you choose your room. Two mistakes cost equally dear: too narrow a room (you’ve shrunk your own audience to the size of an elevator and complain that “the niche is small”) and an unclear topic, when the algorithm can’t tell which room to take you to, it solves the simpler problem: it takes you nowhere.

Rows. You’ve stepped onto the stage, but only the first row sees you (a screen hides the rest). The first row watched to the end, reacted, shared, the screen lifts before the second. The second liked it, the third opens. Reached the seventh, and the seventh got bored? The screen lifts no further. Ever. These are the waves of impressions from the scheme above, but now the main thing is clear: each next row is opened by the viewers, not by you. Your job ends at giving the first row something to vote for.

The main principle: audience fit

Four properties of content that “flies”: the topic is relevant to your viewer and solves their real problem; the information is non-obvious, they haven’t heard it yet; it’s digestible, explained so it’s clear the first time; it can be applied quickly, a short distance from viewing to action. The mechanics: 1, the fingerprint (video + sound + text); 2, a fit score (who will this land with?); 3, a test on ~200 people, mostly non-followers; 4, the reaction, good data? good → 2,000 → 20,000 → 200,000 and beyond; bad → impressions stop. What counts as “good data”: average watch time and % watched through, engagement (likes + comments + shares ÷ views), the share of your content in the viewer’s session. These same four properties turn a viewer into a buyer, not by chance.

How to get comments

Take a clear stance: people comment when they agree or (more often) disagree. Pick the side opposite the conventional one, if you have the arguments. Sharpen the claim (not “you could do it this way too,” but “don’t do it this way, and here’s why”). Build topics around concepts people already have an opinion about. Hit an emotion, indifference doesn’t comment.

Shares: the currency of reach

Most social media users don’t create content. Their main action is forwarding a clip to a friend. So a share is worth many times more to the algorithm than a like: one “paper airplane” weighs more than several hearts, it’s a signal that the content carries the person beyond the app and brings two back.

The market doesn’t grade, the market votes

The final and most important thing. There’s no teacher who’ll give you an A for a proper hook and clean editing. The viewer scrolls the feed in the evening and doesn’t think “the hook’s a bit weak, the font’s debatable, a B-minus,” they were either hooked or not. They vote with their finger: watch-through, like, share. Money is the same: the buyer doesn’t score you on criteria, they simply pay for what got to them.

When the algorithm will “pick up” a channel: on the first clip or the twenty-fifth

A beginner’s most common question sounds like this: “I posted eight clips, reach is 200 views each, does the platform not like me?” The answer is unpleasant in its honesty: the platform doesn’t know you yet. And the moment it does know you comes off no schedule. For some, a channel “takes off” on the first clip; for others, on the twenty-fifth; for a third group, never, though the content is no worse. This isn’t mysticism or a conspiracy; it’s a direct consequence of how audience matching works.

Every recommendation system solves one task: whom to show a clip so the person watches it through. About a new channel it knows nothing, not the topic, not the audience, not the quality. So it does the only reasonable thing: shows the clip to a small test sample and watches the reaction. If the signals are good (they watch through, save, share), the sample is widened. If not, the test is closed and it moves to the next candidate. The clip doesn’t “die,” it just doesn’t pass selection.

Over time the noise decreases, but not because you were “liked,” rather because statistics accumulate. The system gradually builds a picture of the channel: which topics you’re associated with, which people watch you through, whom you resemble. Until that picture exists, each clip is tested almost from scratch. When it appears, a new clip starts already among those with a high probability of watching it through, and the result stops being a lottery.

How many clips does it take for the picture to form? As many as the system needs to find a repeating pattern. If you make the same type of content for the same audience, the pattern is found fast, sometimes from the first clips. If every clip is about something new, shot differently and addressed to different people, you start over each time, and “the twenty-fifth clip” turns into twenty-five first ones.

The format window: platforms pay extra for what they promote

Recommendation systems have a pattern rarely spoken aloud: when a platform is developing a new format, it temporarily hands it reach more generously than usual. At various times live streams and short vertical videos “took off” this way, and now on some platforms photo carousels outpace video in organic reach. It’s not format magic but signal mechanics: each swipe through carousel slides is an active “I’m interested” action that weighs more than a passive view; text slides get reread, and contact time grows; and list slides get saved (“save it, it’ll come in handy”), and a save falls among the heaviest signals.

The algorithm isn’t punishing you. You’re just not important enough to it yet.

“Give me a working funnel” is the same kind of request as “give me a working tool.” A screwdriver is no worse than a saw; they’re just for different jobs. Below, a catalog, not a recipe. Choose by the task, not by the beauty of the diagram.

When you break down one funnel, you see a funnel. When you break down a hundred, you see the patterns. Below, the lessons that recur among those who earn and are missing among those who “make content.”

1. Simple funnels win

The typical scheme of a successful project looks boring: video → a simple landing → a short video with an offer → book a call. No twelve-step branching diagrams. Content does the main part of the sale; the landing merely accepts a decision already made. People polish landing pages for months instead of filming the thing that will convince.

2. Long content, the printing press of trust

3. Guarantees no longer work. Proof does

4. Don’t sell in short videos

A selling clip in the feed cuts reach (the algorithm sees a weak reaction, see chapter 46), and thirty seconds isn’t enough for trust. A short video has three legitimate goals: entertain, teach, inspire. The goal is follows and recognition, not payment.

5. Give the best away for free, but in parts

The fear “I’ll tell everything and they won’t buy” shatters against practice: information has long been free, people pay for the system, the guidance, the speed and the removal of risk. Give away methods in full, but in parts, and don’t fear that someone will assemble it all themselves. They won’t anyway: people buy not the knowledge but its implementation.

6. People buy you, not the product

The same product sells differently depending on who’s selling it. The audience is drawn by what you broadcast: a display of luxury attracts dreamers, breakdowns of mistakes and step-by-step instructions attract practitioners and beginners. Want different customers, change not the offer but the message.

7. If a person recognized themselves, half the sale is done

The main job of organic content is resonance. Not “to show expertise” but to trigger the reaction “this is literally me.” A person who saw their present self (or their future self) in you is already half a customer, all that’s left is not to get in the way.

Postscript: the combo that launched the AI startups

Breakdowns of the public stories of three services that grew from zero to hundreds of millions in revenue in a year or year-and-a-half (app and deck builders, 2024–2026, figures from the founders’ own open accounts) reveal one and the same recurring combo. It deserves a place here because it contradicts half of the “obvious” advice.

Marketing brings a person to the door. Sales open it. You can build a flawless system and lose it all in a chat where the manager replies six hours later with “hello, how can I help?”

Sales begin where there’s a desire, a problem or a pain. If a person has none of the three, it’s not selling but coaxing. Work with those already in one of the three categories; move the rest into content and warm them up (chapter 45). Cold sales aren’t “bad”: they’re just many times more expensive, and manager burnout is baked into the model.

Techniques that actually change conversion

Objections: what’s really behind them

What people are actually buying

A list worth keeping in front of you when writing any offer. A person buys: simplicity, speed, safety, status, relief from pain, pleasure, belonging to a group, contribution to a common cause, meaning. If your offer hits none of these points, it’s not an offer but a product description.

Lead qualification: how to end the marketing-vs-sales war

In any company where marketing and sales are different people, sooner or later the same argument starts. Sales: “The leads are garbage, there’s no one to sell to.” Marketing: “The leads are great, you’re blowing them.” The argument is eternal because both sides appeal to feelings, and feelings can’t be compared. It’s closed only one way, by introducing an objective dividing line: a qualifying action.

An objections phrasebook: what to say word for word

An objection isn’t a refusal but a request for missing information. But answering it with improvisation is a luxury: there are only five typical objections, and each has a worked-out mechanic. First the general rule, then a breakdown line by line.

“I’ll think about it.” Worst replies: “okay, I’ll wait” (the initiative is dead), “what’s there to think about?” (an argument). The working sequence: a falseness check, “What information is missing to decide? How do you like it overall?”; if the objection is true, give time but fix the next step with a choice without a choice: “Shall we talk tomorrow at 10 or at 4?”; you can reserve a spot and send a payment link “just in case”, a noticeable share of payments come in the evening and at night, impulsively. If the person “ghosts” for weeks, an honest question: “Tell me straight, is this a polite no? I won’t take offense.” The “no” you get frees up resources; eternal “waiters” in the base bring nothing but the illusion of a full funnel.

“Too expensive.” Two roots: they genuinely have no money (cured by installments, staged payment, a pared-down package, a different tier) or there’s no value (cured by returning to needs, above). Five working moves: comparison, “Compared to what?” (expensive against what backdrop?); breaking it down, price ÷ usage period = “a cup of coffee a day”; a budget question, “What amount did you set aside?” and reassembling the offer to fit; the cost of inaction, “How much do you lose over six months if nothing changes?”; a flexible format, 50/50 with features limited until full payment, reserving the next cohort. What not to do: give a discount at the first “too expensive”, often it’s a probe, and a gifted discount becomes the new “normal price” (the math is in ch. 76).

“Give me a discount.” Instead of lowering the price, amplify the offer: “Same price, but I’ll add X.” The customer’s goal isn’t to save but to solve a problem; remind them of it. A discount with no reciprocal condition (volume, prepayment, a review) teaches the customer that your prices are a draft.

“Competitors are cheaper.” Break the packages down point by point: what’s included there and what’s included here; calculate the cost of ownership, not of purchase (ch. 76); and a seed-question: “Did they tell you what they cut to get there, or will you find out along the way?”

“I need to consult someone” (another person decides). The broken-telephone problem: the intermediary will relay your arguments with losses and won’t answer the decision-maker’s questions. Two moves: reach the decider directly (a three-way call) or prepare short talking points the intermediary can forward without distortion, just by copying. And always a date for the next contact.

“I’ve been cheated before.” Join them (“I understand, contractors vary”), clarify what exactly happened last time, and don’t step on the same rake in your presentation; remove risk with facts: a contract, a legal entity, staged payment, an escrow, verifiable reviews.

“The customer will think about it”? They will. About something else. At a competitor.

There’s a popular religion: “good content will find its viewer, and advertising is a crutch for those with a weak product.” This religion is preached by people who either sell you a content course or forgot to mention they already have an audience. Let’s take the dogmas one by one.

Dogma one: “organic is free.”

Dogma two: “my followers will see my content.”

They won’t. Platforms are commercial enterprises, not charitable foundations, and their business model is the direct opposite of handing out reach for free.

Dogma three: “advertising is for those with a bad product.”

This inverts cause and effect. Advertising for a bad product really does burn money: it amplifies what’s there, it doesn’t fix. But the opposite conclusion follows: advertising is effective precisely when the product is good. A bad product is an argument not against advertising, but against advertising now.

Dogma four: “I run on word of mouth.”

Word of mouth is wonderful. It has one flaw: it doesn’t scale and can’t be steered. You can’t decide on Tuesday “I need 40% more referrals by Friday.” Word of mouth is a consequence, not a channel. Paid traffic doesn’t replace it but gives what word of mouth lacks: predictability and leverage.

What advertising actually buys

Advertising used to be won by whoever set up their targeting best. Today the machine does the settings for you, and the winner is whoever has the better creative. This isn’t a marketers’ slogan, it’s a consequence of how the auction is built.

How the impression price is formed

You don’t buy impressions off a price list. You take part in an auction where the platform picks the winner not by the highest bid but by roughly this logic: bid × probability of the desired action × quality/relevance. Hence a non-obvious but fundamental consequence:

Modular creative: a construction kit instead of a masterpiece

Professionals don’t “shoot a clip.” They assemble it from interchangeable blocks:

Test with components, not clips

The question “is advertising expensive” is meaningless without a second question: “how much does a customer bring you.” A liter of gasoline is “expensive” or “cheap” depending on how far you’ll drive on it and what you’re hauling.

Calculating the allowable cost of a customer

The learning phase: why you can’t “just try $100.”

Ad systems learn on conversions. While there are few, the algorithm shows ads almost at random. The reference for major platforms is on the order of 50 target actions per week per ad group to exit the learning phase.

Where the money actually leaks

Attribution: why “I don’t see sales from ads” is often a measurement error

The practical takeaway: the ROAS figure in the ad dashboard isn’t truth but a hypothesis. It’s tested only by switching off: pause the channel in some regions or for a period and look at total revenue. It’s uncomfortable and “scary,” and that’s exactly why almost no one does it.

When paid traffic really does burn money

Advertising doesn’t work for those for whom everything after the click doesn’t work.

Between “pay the platform” and “wait for organic” there’s a third path: pay people who already have an audience. It’s a distinct type of paid traffic, with its own economics, its own rakes and, arguably, the best trust-to-price ratio.

Bloggers: why micro is often more profitable than macro

Partners and ambassadors: leverage without a budget

The funnel is an honest model with one fatal property: it ends. A person fell to the bottom, paid, and that’s it, the system waits for you to pour something in the top again. As long as you pay, it flows. Stop, it dries up. This isn’t a growth machine. It’s a hose.

What a growth loop is

A loop is a closed system in which the output of one cycle becomes the input of the next. Not “acquired → sold → the end,” but “acquired → something happened → that something attracts the next people.” The difference isn’t cosmetic: a funnel gives linear growth (invest double, get double), a loop, compounding (each turn makes the next one cheaper).

A funnel spends what’s invested. A loop reinvests what’s produced.

Why this isn’t “just another trendy scheme.” Because it has an unpleasant corollary. If growth comes in loops, then your product must be built into a loop, not the other way around. The classic mistake: build the product first, then look for a channel. The opposite rule works: products adapt to channels, channels don’t adapt to products. The algorithm won’t love you for trying hard.

“I don’t know what to film” isn’t a creative crisis. It’s the absence of a production system. An artist waits for inspiration. A factory ships output on schedule. Guess which one will outpace you on reach.

The engine’s core: pillar → atomization → distribution

The mechanic all big content machines stand on has three phases. You produce one large unit (the pillar), break it into dozens of derivatives (atoms), send them out across platforms, and gather signals that become the raw material for the next pillar.

The proportions that work

Balance: reach, expertise, sale

The engine must ship three types of content, and the proportion matters more than the quality of any single unit. Classic working splits: 80/20 (value/promo), the rule of thirds (value / personal / sale), 50-30-20 for consumer niches. The specific numbers aren’t sacred, discipline is: if you have 100% selling content, you don’t have an engine, you have a teleshopping channel, and the algorithm will notice before you do.

Adapt to the platform, don’t copy-paste. What lands becomes the topic of the next pillar: the loop closed. The signals (what got watched through, what got asked, what got bought) become the topic of the next pillar, the audience writes your content plan in the comments; you just read it.

The planning horizon: why a month doesn’t work

The classic advice “make a content plan for a month” breaks in practice against one thing: topics age faster than the plan runs out. Teams that live content professionally arrive at a three-level horizon: strategy and meaning for a month, topics and scripts for two weeks, filming for one. By the end of a monthly plan, topics “burn out” and perform noticeably worse: the news hooks cooled, the audience’s mood changed, competitors already worked the same trend. And the more unstable the times, the shorter the horizon should be: posts and clips recorded a month ahead are canned goods, and the audience senses canned goods from the first second.

The muse visits those who have a content plan. Everyone else just gets a visit from the deadline.

If you test ads as whole clips, you’re not testing, you’re guessing. The clip lost. Why? A bad hook? A boring body? A lousy CTA? The wrong offer? Congratulations: you spent money and learned nothing.

The creative breaks down into parts

Testing in layers

Why the creative became the targeting

Manual targeting died not because marketers got dumber but because platforms took the work for themselves (chapter 51). The algorithm itself decides whom to show. On what does it decide? On the creative’s contents: what’s in frame, what’s in the sound, what’s in the text, who reacted. Your creative is your targeting. A bad creative now doesn’t just “fail to convert,” it wrongly defines your audience.

The creative bank: your main asset

Keep a library of “what already worked.” Separately: winning hooks, winning bodies, winning CTAs, visual techniques, formats. In six months this stops being a spreadsheet and becomes a weapon: a new product launches not from scratch but by reassembling proven modules. A competitor who “thinks up something” every time is playing chess against you having just seen the board for the first time.

Cross-posting is when you do it equally badly everywhere at once and call it omnichannel. The distribution cascade. The order in which content moves across platforms isn’t random. The logic is simple: where trust is built, the long stuff; where attention is caught, the short; where the sale happens, the personal.

One shoot day, three endings

Multi-platform seems expensive as long as it’s done head-on: separate content per platform. The working mechanic is many times cheaper. All the clips of a two-week cycle are shot in bulk in one shoot day: one location, one light, one outfit, a dozen scripts in a row. And the key technique: at the scripting stage, two or three different endings are written for each clip (and, where needed, different opening lines) for different platforms, and the expert records them right there, without getting out of the chair. The editor then assembles versions: one with a follow call to action, another leading into the funnel, a third with a native platform tie-in. One shoot day covers two weeks of presence everywhere, and the cost of adaptation drops to almost zero. It’s the same modular thinking as in ch. 56, only applied before the shoot, not after.

You have a choice: use AI as a machine tool and produce ten times more, or use it as a replacement for thinking and produce ten times more of what no one will want to watch. The second path is more popular. It’s easier.

Where AI really adds value

A human in the loop (and exactly where)

A fully autonomous content conveyor is a way to automate reputational damage. Control is placed by risk level.

The main danger: greyness at industrial scale

While you were finishing this book, some people stopped googling. They ask a machine and get an answer, not ten blue links. And in that answer, either you’re there or you’re not.

From search to answer

The task used to be: land in the top of the results. Now a second one appears: land among the cited sources of the answer. The model doesn’t show a list: it synthesizes a conclusion and mentions a few sources. Being on the second page of results was bad. Not being in the answer means not existing at all.

What will age and what will remain

By the time a person has read your headline, color has already told them everything: whether this is expensive or cheap, whether you’re one of them or an outsider, whether you’re shouting or speaking. Color is a brand’s fastest message. And the most illiterately used.

First, an honest frame

What works then? Three things stronger than “meanings”: contrast, category code and consistency. Contrast matters more than hue.

The classic argument “which button converts better, red or green” is an argument about nothing. In well-known tests, the winner wasn’t the “psychologically correct” color but the one that stood out with more contrast against the rest of the page. The button should be the only hot spot in its zone: then it works, be it orange or purple.

Category code: follow it or break it

Every niche has color conventions: finance and tech drown in blue (“trust and reliability”), eco-products in green, luxury in black-and-gold, children’s in pastel. It’s useful to know these codes, because you always have exactly two meaningful moves:

Consistency: color as a signature

Recognizability is built not by the “correct” color but by the same color. An accent repeated across a hundred content units turns into a signature: the viewer recognizes you in the feed before reading the name. So a brand palette is a discipline of four decisions: one dominant, one accent, two neutrals. That’s all. Every “one more pretty color” weakens the signature.

A person considers themselves a rational loner right up until they see a queue. Then a program a hundred thousand years old kicks in: where everyone is, it’s safe; where no one is, there’s a catch. Marketing that ignores this program plays against the species’ firmware.

The crowd in your pocket

Back in the late 19th century, Gustave Le Bon described what happens to a person in a crowd: personal responsibility dissolves, critical thinking weakens, emotions infect faster than arguments, and actions become imitative. Le Bon wrote about town squares. Social media put that same crowd in everyone’s pocket: the feed is a square, the comments are a chant, a hype wave, panic or euphoria, spreading by the same laws of emotional contagion. Hence two practical consequences: high-energy emotions (delight, outrage, astonishment) spread better than calm ones, which is exactly why the awe-inspiring and the outrageous go viral, not the “balanced”; and the first visible reactions set the tone for all the rest, the crowd watches the crowd.

Social proof: the main safety label

When a person is unsure, they look at what others do, especially others like them. This is social proof, the most exploited mechanism in commerce.

Six levers of influence

Robert Cialdini systematized the mechanisms all persuasion practice stands on. Here they are in one table, with an honest note on where each lever’s influence ends and manipulation begins:

Cascades and herding: the anatomy of hype

An information cascade is when each next participant orients not on facts but on the actions of the previous ones. The first bought → the second saw the rise and bought “because it’s rising” → the third bought because the second did. No one in the chain checked the basis. This is how crypto pumps inflate, “last spots” sell out and trends take off; and this is how they collapse: a cascade works both ways, and panic is more contagious than euphoria.

The line: influence or manipulation

Most businesses start social media on autopilot: “we’ll make a channel → we’ll post → people will follow → they’ll buy.” Every arrow in that chain is an unproven hypothesis. This chapter replaces the autopilot with a model that works boringly but brings money.

Test one: is a subscription needed to buy from you

Ask yourself one question: must a person be following your channel to be your customer? To fill your car with gas, you don’t need to read the gas station’s channel. To order on a marketplace, the marketplace’s channel. For most routine purchases, a subscription isn’t technically needed at all.

A subscription becomes natural when it benefits the customer to hear about you before others: a clothing brand (a new collection sells out by size), a restaurant with reservations, a studio with limited slots, a product with updates and releases. If there’s no such reason, you can create one, promo codes, early access, closed terms, but it costs real money. A discount, a chocolate with your coffee and a giveaway aren’t “content” but a marketing budget, and it must be counted like any other budget.

The representation model: social media as an office, not a show

The working baseline strategy for a business is to run a channel as a representation: an additional service office, a showcase of goods and services. What’s in it:

• Official information. Who you are, where you are, hours, current goods and services, prices or price ranges. Ideally a full catalog or shop right on the platform, at minimum a catalog with a “want to order, DM us” link.
• Cases and the work as it is. For services: “for this client we did this, here’s the process, here’s the result.” Plenty, in detail, no filler. Writing a real case via an outside agency is nearly impossible, you have to be inside the business; it’s a job for your own person.
• A duty person in the comments. The most expensive and the most important. A person can write under any post “I want a table at 3pm” or “is it in stock?” and get an answer in minutes, not “we’ve forwarded your question.” This is customer service, and it sells better than any reel.
• A sign of life. The last post no older than two weeks. A channel with perfect design and a two-year-old last update tells the customer one thing: “we’re gone.” Freshness is trust, and trust is the multiplier from the chapter 1 formula.

This model has three “drawbacks” that make almost no one choose it: it’s unsexy (no splashes of emotion), it’s hard to delegate to an agency “turnkey,” and, most annoyingly, it doesn’t need followers as a metric. Representation serves those who came, not collects those who walked by. That’s exactly why it works: all three “drawbacks” are signs of a tool, not an attraction.

What a representation should publish

The range and changes to it, cases and the work process, answers to real customer questions (the best content plan is the log of questions from DMs and calls), social proof (ch. 61), terms, promos with honest mechanics, news that affects the customer (a move, a schedule, price increases announced in advance). All of it is boring right up until a person with money is looking for whom to give it to.

You can’t hold a department to “sales,” so you hold it to “reach.” The department optimizes reach and hits it. The business is neither better nor worse for it, but everyone has pretty dashboards. This chapter is about how to stop looking for keys under the streetlight.

How the trap is born

In a large company, tasks cascade: the marketing director has a budget to spend “usefully.” You can’t tie the SMM unit’s usefulness directly to revenue, revenue is affected by product, prices, sales, season. So the unit is given something synthetic and measurable: reach, followers, engagement, amount of content. Then Goodhart’s law kicks in (ch. 3): a metric that becomes a target stops measuring anything. The team honestly grows the numbers, buys traffic for giveaways, shoots viral clips with no product, in the worst case tops up with bots. The numbers grow, sales aren’t personally acquainted with those numbers.

How to measure a representation

The trap question: “how will we measure the effectiveness of social media?” The counter-question: how do you measure the effectiveness of an office? You don’t, it’s a condition of doing business, a fixed cost, hygiene. A social media representation (ch. 62) is the same category. It’s not measured by reach. What can and should be measured:

• Speed and completeness of response. How many minutes a person waits in comments and DMs. It’s painful and costly to measure, which is precisely why it’s the real quality metric.
• Inquiries and leads from the channel. Tagged links (UTM), channel promo codes, the “how did you hear about us” question in the CRM. Not for “post payback” but to see that the showcase is alive.
• Freshness and completeness of the showcase. A checklist: is the catalog current, are the top-20 questions answered, is the content no older than two weeks.

But a media project (ch. 64) and paid traffic (ch. 50–52) are performance tools, and they must be measured in money: CPM, CPL, CAC, payback. The trap is the swap: measuring representation as performance, and a media project as “image.” Do the opposite.

Every content unit has a goal, and exactly one

The theater of metrics starts where a publication is judged “overall”: liked or not. A working system is stricter: before it goes out, every content unit is assigned one goal of four (reach, follow, funnel entry, inquiry), and it’s judged only by its own goal. A reach clip with zero inquiries is a success if it gathered reach. A selling post with small reach but five inquiries is a double success.

The dashboard’s green, the till’s empty. Welcome to the theater of metrics.

People come to social media to have fun, zone out and treat themselves. This isn’t an insult to the audience but the main law of the platforms, from which everything follows: what sells easily there, what doesn’t sell at all, and why a brand can almost never be a good media outlet.

What an entertainment feed sells

Vertical videos brilliantly sell goods in the “treat yourself” category: inexpensive, emotional, bought without engaging the brain, cosmetics, clothing, accessories, gadget-toys, food. The threshold formula: a price below the amount at which a person starts to think, the purchase happens on the same emotion as the viewing. Plus the product must be one click away (marketplace, cart in the bio).

Then the imitation mistake begins: a serious business looks at the sellers of meme T-shirts and decides to sell machine tools, industry permits and corporate rollouts “exactly the same way.” A beautiful clip is shot with an actress by the machine. The clip gathers views (beautiful people are pleasant to watch), no sales: the viewer came to be entertained, the deal cycle is six months and a tender. Estimate the economics before the shoot, not after.

When entertaining content is genuinely needed by a serious business

Three honest cases. First: you actually have a “treat yourself” product in the line (ch. 7), merch, a cheap entry, a tripwire. Second: the content entertains your professional audience, “insider” humor in a B2B niche works as a “we’re the same blood” marker and warms recognition in a narrow segment, but it’s a long game with no direct sales. Third: you’re building a media project as a separate business, that’s chapter 65. In all other cases, “let’s shoot trends” is a fashion tax paid from the marketing budget.

“Let’s start our own media project, we’ll have free traffic.” This dream has a price tag: years of investment, an ad-sales department and a complete surrender of brand policing. This chapter is about the divide between a brand and a media outlet, and why renting others’ audiences is almost always cheaper than building your own.

Why a brand is a bad publisher

Media lives on attention, and attention lives on edge: conflicts, honesty, humor, uncomfortable topics. A brand lives on safety: lawyers, “we can’t be associated with this,” seven rounds of sign-off. Every brand-safety limit makes the content a bit more sterile, and the sterile doesn’t get watched. Successful brand media exist, but they all have a hard divide: the editorial team can write what’s useful to the reader even if it’s inconvenient for the product (“don’t take out a loan,” “in this situation you don’t need our service”). Paradoxically, that’s exactly what sells: trust converts better than loyalty by decree.

The economics of a media project: a slow plane

A media project reaches operational payback slowly: audience and views grow as a trend, not from “one genius clip.” All that time you have to pay production, and not let the CFO cut “non-core costs” mid-year, right before the season. And the main thing: a live media outlet pays off through the advertising of many advertisers, one product’s revenue can rarely feed a media outlet of noticeable size. So you’re effectively opening a second business: with an ad-sales department, a sales kit and a CPM price list.

Renting audiences: a house or a hotel

The alternative to building is renting. There’s a whole industry of people who’ve already sacrificed themselves to the feed: young, charismatic, with an engaged audience. You can bring them money and a product, and they’ll make advertising better than you would about yourself, because your brand policing doesn’t affect them and their audience trusts them. This is chapter 53 in action, taken to a strategy: your own channels are the showcase (ch. 62), and the main driver of reach and leads is others’ media: bloggers, industry publications, topical channels, podcasts.

It’s counted simply and boringly. Ad sellers have an industry standard, CPM. Write it into a table: blogger A, blogger B, an industry channel, targeted ads, search, and their forecast CPM/CPL. Run a test on a small budget, measure, reallocate. No magic, just a column in a spreadsheet, and immunity to “fetish placements” with the idols of your youth, which warm the ego, not the funnel.

“We hired a copywriter, where are the sales?” Nowhere, and it’s not the copywriter’s fault. You can only demand responsibility for results together with the tools of influence over results. This chapter is about an honest contract with a content team, and about when you shouldn’t run a personal channel at all.

A copywriter and a content marketer are different professions

A copywriter executes: give them a topic, a channel and a format, and they write the text. The result (leads, sales) is affected by the topic, the channel, distribution, the offer, the moment, all things outside the copywriter’s control. The topic works more than the execution. So hanging a leads KPI on a copywriter is asking a carpenter to answer for the architecture.

A content marketer is a different role, with a different zone of responsibility: find the channels where the target audience lives; devise content that will land in those channels; organize production and publishing; set up analytics to track the effect. You can hold such a person to results, if you gave them the levers:

• the right to choose topics and formats themselves (otherwise they’re an executor, not an owner);
• the right to choose channels, including others’ platforms for money (ch. 65);
• a promotion budget and sales tools, promo codes, discounts, special offers;
• access to analytics: metrics, UTM, CRM, otherwise they’re blind;
• real autonomy in decisions within the budget.

How to assess an executor without fake KPIs

If your content function is representation (ch. 62) and you set the strategy yourself, then an honest assessment of an executor is subjective, and that’s fine. The main question: is it getting easier or harder to work with this person over time? A normal specialist learns: by the second month they know the product better, hit the style more precisely, need fewer edits. If the arrow points the other way, replace them without waiting for “objective” numbers. Additionally: adherence to showcase hygiene (freshness, completeness, response speed) is already measurable and 100% within the executor’s control.

A personal channel: a marathon, not a sprint

A separate case, “an entrepreneur starts a personal blog because everyone has one.” Before you start, answer the uncomfortable question: could you do this for ten years? Better, twenty. A personal channel is a media project (ch. 65) in miniature, with all its slow-plane economics. It pays off over a long distance and only if the production pace is natural for you. You read a book a week, a channel about books is possible; a book every six months, no. An extrovert finds sharing easy, an introvert will burn out after a month of squeezing out posts by willpower: the first excuse to quit will be used. Successful media personalities are people whose temperament, energy and life stage matched the format. If that’s not you, no problem: the whole audience-rental industry works for you, and let your channel be a showcase.

How involved the expert is: an honest delegation scale

The expert’s main fear, “content will devour my whole life,” is cured not by heroism but by a distribution of roles. In a mature editorial setup the expert is involved at specific points: signs off topics once a week, shoots one shoot day every two weeks (ch. 57), the team carries the rest. Stories are assembled by the team from the expert’s points and material; the photos, videos and screenshots for them accumulate in a shared folder as life goes on. Comments are answered by an assistant trained in the tone of voice (ch. 19), contentious threads are escalated to the expert. And a normal, not shameful, practice: the expert doesn’t read the script word for word but freely retells it in their own words, keeping the hook and structure; this isn’t a process defect but the last mile of authorship (ch. 79 on points instead of a teleprompter).

“Which social network should I choose?” is a question posed upside down. The right order: audience → the format you can reliably produce → the platforms where that format and that audience meet. This chapter is a map for that choice.

Three platform types by attention mechanic

Recommendation-based (TikTok, Reels, Shorts, partly YouTube): the algorithm hands content to a cold audience, these are reach machines, the top of the funnel. Following is secondary, every clip competes from scratch (ch. 46). Subscription-based (Telegram, email, largely VK communities): reach rests on the base, these are retention and sales machines, the middle and bottom of the funnel. Search-based (YouTube search, Pinterest, maps, partly Reddit): content lives for years and catches existing demand, an evergreen asset. A healthy system (ch. 18, 54) combines all three types: recommendation-based catch, subscription-based warm and sell, search-based work in the background.

The platform map

* Instagram belongs to Meta, which is designated an extremist organization and banned in Russia; account for the platform’s legal status in your jurisdiction.

Language and geography: the axis forgotten when choosing platforms

The platform map above silently assumes they’re all equally accessible to your audience. For a Russian-language author that isn’t so, and planning without this correction means building distribution on sand. The reality: some global platforms are restricted or throttled in Russia, uploading content from Russian accounts is unavailable on some, and a noticeable share of the audience goes there via VPN. And that share is uneven: the technical and young audience lives on VPN almost entirely, the mass older audience barely goes at all.

And one clarification about short and long formats, important when allocating effort: a follower who came from a short vertical clip and a follower from a long video are different people in depth. Short gives “broad” followers: many, cheap, shallow. Long gives few, but deep: authority and readiness to buy. So the combo is always two-way: short catches and must lead to long, long converts (ch. 45, 55).

Running five platforms by hand is a courier’s job: noble, but scales badly. Software won’t invent your strategy, but it removes 60–80% of the manual labor: planning, publishing, analytics and production routine. Let’s break down the classes of tools and how to stitch them into a conveyor.

Classes of tools

Three tools up close

Metricool is a “scheduler + analytics” combine for small teams. In practice: you connect accounts (Instagram*, TikTok, YouTube, Facebook*, LinkedIn, X, Pinterest, Twitch, Google Business and others), get a single calendar, auto-posting, a heat map of the best publishing times, consolidated analytics across all networks and competitors, an ad planner. There’s a free tier with limits (one brand connection, a monthly post cap), enough to tell whether it’s your tool. Scenario: you run 3–5 platforms by hand and want one panel instead of five tabs and a weekly report with no copy-paste.

Postiz is an open-source scheduler. The key difference: you can deploy it on your own server (self-hosted) and pay no subscription at all, you pay with the server and your time to set it up (Docker, and self-deploying also requires registering API apps with the platforms). It supports dozens of networks (X, Telegram, Instagram*, TikTok, LinkedIn, YouTube, VK, Mastodon, Bluesky, Discord and others), has a public API, integrations with n8n/Make/Zapier, team roles and post approval, AI generation features. Scenario: an agency or a technically savvy team that cares about data control, many accounts and no “per-seat” fee. There’s also a paid cloud version if you’d rather not fuss with a server.

Upload-Post is a “postman” for automations: a service that hides publishing to a dozen platforms (TikTok, Instagram*, YouTube, LinkedIn, Facebook*, X, Threads, Pinterest, Reddit, Telegram and others) behind a single API call. You don’t build your own integrations with each network or pass their API moderation, you just send the video and captions in one request, and the service converts the file to each platform’s requirements. There are official nodes for n8n and Make, an MCP server for AI agents, a free tier of a few uploads a month. Scenario: you’re assembling a content conveyor (ch. 55, 58), e.g. “long video → auto-cut → auto-publish shorts everywhere”, and you need a programmatic “output,” not a calendar with buttons.

How to choose: a checklist

• Does the tool support your platforms (for RU networks, VK, Telegram, Dzen, check separately: Western services often can’t handle them, here SMMplanner, TGStat and the platforms’ native tools help out)?
• Does it work through the platforms’ official APIs (see “Careful” below)?
• What about analytics: only “likes-followers” or can you export the data and tie it to UTM/CRM (ch. 63)?
• Do the economics add up: the plan price against the hours saved?
• Is there export/API, won’t you end up locked into the tool?

A scheduler won’t make content better. It’ll make it on time, and that’s already half of it.

Donation platforms aren’t a “tip button” but a separate layer of the business model: subscription-based monetization of the audience core. Built in right, they give what any business from chapter 3 dreams of, predictable recurring revenue (MRR) from the most loyal. Built in wrong, they turn into an empty page with zero subscribers and a feeling of awkwardness.

A map of core-monetization platforms

The setup that sells: tiers and promises

The mechanic on the Boosty example (the others differ cosmetically): you register, pass verification, set up the page (cover, description, this is the offer, apply chapter 16), create subscription tiers. The classic three-tier ladder:

• Entry ($2–5): early access to content, posts without platform censorship, a “thank-you” tier. The task, a minimal threshold: to convert gratitude into the habit of paying.
• Core ($7–17): the exclusive: closed materials, source files, breakdowns, access to a closed chat. The main revenue lives here.
• Top ($25+): closeness and participation: content votes, personal reviews, a credits mention, early access to products. Few buy it, but the tier raises the average check and makes the core tier look “reasonable” by contrast (price psychology from ch. 17).

The promise rule: promise a format of access, not a volume (“closed posts as they come out,” not “10 videos a month”). An unmet schedule is the main source of unsubscribes and refunds. And duplicate the donation-page link everywhere: in video descriptions, profile bios, pinned posts, people don’t look for a way to pay, it’s handed to them.

Where to build it into the strategy

The donation layer is a rung of the product ladder (ch. 7) and a Revenue/Retention mechanic (ch. 3, 41): free content builds reach and trust → the core gets a “closer and sooner for a subscription” offer → the closed community raises retention and gives you a direct channel for launches (ch. 34) without algorithm intermediaries. For whom it’s the main model: authors, streamers, artists, analysts, niche media. For whom it’s auxiliary: experts (a club as a tripwire before the flagship), game projects (early access, devlogs). For whom almost never: classic product and B2B business, your core needs not “exclusive” but the product; build a loyalty program, not a Boosty.

One post with someone else’s track can cost you a channel you built for five years. Legal hygiene isn’t about fear but about keeping an expensive asset from having a single point of failure. An honesty disclaimer: this chapter is a reference and safety technique, not legal advice; laws differ by country and change, for specific situations go to a lawyer.

Copyright: the basics in three paragraphs

Any content, text, photo, video, music, code, illustration, is protected by copyright from the moment of creation, automatically, with no registration or © marks. “Found it online” doesn’t mean “can use it”: by default you can use nothing except cases directly permitted by law or license.

Permissions come in three kinds. A license from the rights holder: purchased stock, a subscription music library, direct written consent. Free licenses: Creative Commons (read the letters carefully: BY, credit the author; NC, no commercial use; ND, no modifications; SA, derivatives under the same license), the public domain, the platforms’ own free libraries. Legal exceptions: in the US, the fair use doctrine (criticism, parody, education, assessed by a court on four factors, not by a checklist); in Russia and many other countries, “free use,” primarily quotation: in an amount justified by the purpose, with mandatory attribution of author and source, for informational, scientific, educational or cultural purposes.

Strikes, Content ID and how the penalty works

Platforms have two protection circuits. Automatic (Content ID on YouTube and the like): the system finds audio and video matches itself; the rights holder can, at their choice, take your monetization, restrict the display, or block the clip. It’s not a strike, but it’s a signal that you’re building a channel on someone else’s foundation. Legal (a rights holder’s complaint, in the US, under the DMCA): that’s already a strike. On YouTube, three copyright strikes mean channel deletion; strikes expire over time (about 90 days) and are lifted via a counter-notice if you’re right and ready to defend it. On Twitch, repeated DMCA complaints lead to a permanent ban, which is why stream music must be licensed only. A separate circuit is the community guidelines: violence, disinformation, dangerous acts, spam; their own strikes, their own penalty ladder up to deletion. And shadow limits: for clickbait patterns, mass actions and “borderline” content, platforms quietly cut reach with no notice.

Advertising and disclosure: where “just a post” ends

If you were paid (in money, goods, barter) for a mention, it’s advertising, and in almost every jurisdiction it must be labeled. In Russia there’s mandatory labeling of online advertising: obtaining a token (erid) via ad-data operators, an “Advertising” label naming the advertiser, reporting to the state registry; fines apply to both advertiser and distributor, including bloggers. In the US the FTC requires clear disclosure (#ad, “paid partnership”), in the EU and UK similar transparency rules. Use the platforms’ built-in tools (“paid partnership,” “includes paid promotion”), but check whether they suffice under your country’s law. Certain categories (finance, medicine, alcohol, gambling, crypto) are regulated more strictly, up to a total ban: check before the deal with an advertiser, not after.

People on camera and personal data

A person’s image is a distinct object of protection: commercial use of recognizable people requires consent (in Russia, the image-rights statute, with exceptions for public places and public interest; in the EU, GDPR is added). In practice: filming people for content, get consent (for shoots with participants, a written release); collecting leads, publish a data-processing policy and get consent for newsletters; store the subscriber base in accordance with local personal-data requirements.

The account is intact, the posts are in place, you can write and comment, and reach fell tenfold overnight and won’t come back. This isn’t “the algorithm in a bad mood.” It’s a quiet restriction of distribution: the platform doesn’t delete the content but stops showing it to those not following you. No platform calls it a “shadowban,” and every one does it.

The key to understanding: all big platforms have two different distribution mechanisms. First, delivery to followers: those who follow you can see your content. Second, recommendations: showing it to strangers in the “For You” feed, in search, in “similar,” in roundups. The rules differ, and so does the bar. To avoid deletion, it’s enough not to break the general community guidelines. To get recommended, you must clear stricter recommendation rules. Between these bars lives the grey zone: content is allowed but not fit for promotion. Meta calls it directly “not recommendable,” TikTok “not suitable for the For You feed,” X “visibility filtering,” YouTube “limited display.” Different words, one meaning: you weren’t banned, you stopped being distributed.

The common mechanic: what gets smothered everywhere

The triggers differ in details between platforms, but 90% of cases come down to five families, and they’re universal.

• Automation and “grey” services. Faking followers, bots for mass likes/follows, auto-DMs, services that require your password, click emulators. Even a brief connection of such a service leaves a trace and keeps the account under suspicion for weeks.
• A robot-like rhythm. A hundred follows an hour, identical comments under others’ posts, identical texts in DMs, copy-pasting the same caption under every publication. A live person doesn’t act like that, classifiers see it by speed and uniformity.
• Unoriginality. Re-uploading someone’s video, clips with another platform’s watermark, compilations with no contribution of your own, auto-generation with no value. All platforms tightened their unoriginal-content rules over 2024–2026: such content isn’t deleted but isn’t let into recommendations either.
• Borderline topics. Weapons, drugs and supplements, gambling, crypto earnings, “medical advice,” diets and weight loss, soft erotica, conflicts and politics, shocking footage. Formally allowed, not recommendable.
• Complaints and reports. Mass complaints, from competitors, trolls or just irritated viewers, trigger an automatic downgrade even before a human looks at the material. It’s the most common trigger, and one you’re objectively not party to.

What happens on each platform

The recovery protocol

A universal sequence that works on all platforms. It’s boring and takes patience, but there are no alternatives, no “unban button” exists.

• Find the trigger, don’t fight the symptom. Open the account status, see which publications are flagged. Often it’s one specific one, delete it or appeal.
• Turn off all automation. Revoke access from all third-party services, especially those you gave a password. If you once bought followers, that can keep the account in the shadows indefinitely.
• Take a 48–72-hour pause. Don’t publish, don’t mass-follow, don’t send mailings. Let the speed counters reset.
• File a review request. The button is right in the account status; also in the help section. Phrase it dryly: since such-and-such date, publications stopped appearing in search and recommendations, there are no violations, please check.
• Come back gradually. First one safe publication, a day later the second. No borderline topics, watermarks or links in the first posts, minimal hashtags and only narrow, safe ones.
• Wait. A light restriction usually lifts in a few days, a typical one in 2–4 weeks, a heavy one (repeat violations, faked numbers) up to several months. The recovery metric is one: the share of reach from non-followers returned to its former level.

The good news: you weren’t banned. The bad news: you’re just not being shown to anyone.

A person didn’t “fail to notice” your banner. They saw it, classified it as an ad in a fraction of a second, and erased it from attention before becoming aware of it. This isn’t audience laziness or bad design, it’s a filtering skill trained over twenty years, and it works faster than consciousness.

The term “banner blindness” appeared back in nineties web-usability studies: users asked to find information on a page couldn’t recall large bright blocks they’d looked straight at. The brain conserves resources: it builds a model of the page (“header, navigation, text, ads on the edges”) and preemptively spends no attention on zones where experience says nothing useful ever is. Then comes the nastiest part: blindness is learnable and transferable. As soon as a format becomes recognizable, a rectangle on the edge, a pop-up, “Ad” in the corner of the feed, an integration at the fifth minute of a clip, the filter generalizes it and starts cutting everything similar, including your honest and useful message.

How blindness looks today

Banners on sites are yesterday’s problem. The same mechanism moved into all modern environments and mutated.

• Scroll blindness in feeds. The finger scrolls faster than awareness works. The skip decision is made on the first frames: the recognizable “talking head + logo” cut, “today I’ll tell you,” and any hint of a sale are a command to scroll. This is blindness not to the format but to the intonation.
• Blindness to native. The audience learned to recognize an integration by small things: a change of angle, suddenly even lighting, an enthusiastic tone, the word “partner.” An ad pretending not to be an ad is caught by the same filter, and punished harder, because a sense of deception is added.
• Blindness to the interface. Inside your own product, users stop noticing banners, badges, “important announcements” and even buttons placed where promo usually lives. Hence the paradox: the brighter and more “ad-like” you made an important message, the less often it’s read.
• Creative burnout. The same clip stops working after 3–5 contacts not because it’s bad but because it became familiar: CTR crawls down, cost per click up (ch. 55). This is blindness accumulated personally to you.

What actually breaks through the filter

• A pattern break. Attention wakes on a mismatch with expectation: an unexpected first frame, an unusual angle, an honest phrase instead of an enthusiastic one. Not shock for shock’s sake, shock is filtered just as fast, but a break of the advertising form (see hooks, ch. 11–13).
• Usefulness before payment. Material that’s valuable even if you buy nothing doesn’t read as an ad, because structurally it isn’t one: value first, offer after (ch. 21, 57).
• The form of the environment, not the form of an ad. Content shot and edited like an ordinary platform post passes the filter. Note: this is about form, not about hiding the fact of advertising, no one canceled labeling (ch. 70).
• Frequency management. Rotate creatives and cap frequency per person: refresh before CTR falls, not after. One offer, several different angles (ch. 58 on modular creative).
• A place where people expect information, not ads. Blindness is tied to zones: where a person seeks an answer (search, comparison, reviews, an instruction), the same offer is seen, because it matches the intent.

Chapter 61 was about the customer’s biases, the ones persuasion stands on. This chapter is about the other side of the table: the systematic errors of the one who sells, plans and calculates. They’re more dangerous than the customer’s, because the customer loses money on one purchase from their bias, and you lose a business direction from yours.

It’s important to grasp the nature of the phenomenon: a cognitive bias isn’t stupidity or a result of fatigue. It’s a predictable deviation of thinking, built into the human design and working the same in a novice and a veteran. That’s exactly why it can’t be beaten by willpower (“I’ll be objective”), only by procedure: a rule set in advance that fires at the moment the bias is interfering with sober thinking.

Everyone around you is subject to cognitive biases. Except you, of course.

A niche is chosen out of love or hearsay, and paid for over years. Yet a niche can be evaluated in advance, not perfectly, but well enough not to spend a year on a topic that structurally doesn’t pay off.

Let’s agree up front on what a “bad niche” means. Not “you can’t succeed in it”, you can in almost any. It means: for the same effort you’ll get less money than in the one next door, because what’s in your way isn’t your talent but the market’s design. Games gather millions of views and pay pennies per thousand impressions, because the audience is young and advertisers don’t fight for it. Finance pays ten times more at a tenth of the reach, because for one client a bank will give a sum a mobile-game developer never will. This isn’t about content quality. It’s about who stands behind the viewer with a wallet.

The six criteria

Score the niche from 1 to 5 on each. The sum ranges from 6 to 30.

Scores for twenty niches

Below is a working table. The money scores rest on the spread of per-thousand-impression rates in the 2020s: the gap between top and bottom is roughly tenfold. The other criteria are an expert assessment of production reality. The figures aren’t dogma: recalculate them for your country, language and monetization method.

Four good niches and why

• Marketing, B2B, software (22). A rare combination: an expensive audience and cheap production (screen, talk, breakdown). Plus your own products sell directly, consulting, courses, services. The weak spot is low virality: you’ll grow slowly and through value, not hype.
• Renovation, construction, home (23). An underrated niche: advertisers pay decently (tools, materials, equipment), the topic is eternal, only partly seasonal, run with your hands without a face, easily translated to other languages. Competition, meanwhile, is noticeably lower than in gadgets and beauty.
• Craft, survival, workshop (24). Cheap views, but there are very many and they’re international: the content barely needs language. Full anonymity, high virality, moderate competition. The monetization logic here is different: not a per-impression rate but volume, goods and affiliates.
• Science and “how it works” (23). A universal topic, high shareability (the main algorithm signal, ch. 46), a voiceover instead of a face, easily turned into sponsored integrations with educational services.

Four hard niches and why

• Travel (15). Romance versus economics: production is expensive (flights, filming, time), rates are mid, competition enormous, and anonymity impossible, you are the product. It only works with your own monetization (tours, courses, hotel partnerships).
• Beauty and skincare (16). Rates are decent, but the niche is packed to the limit: to be noticed you need either a budget or a very strong personality. Plus strict platform rules around “medical” promises.
• Psychology and self-development (16). Easy to produce, so everyone produces. The audience is huge, but advertisers pay mid for it and trust is won slowly. A niche where it’s easy to start and nearly impossible to stand out.
• Games (18 with paltry “money”). The classic trap: the easiest niche to enter (screen recording, no face), enormous reach, and some of the lowest rates on the market, because the audience is young and not wealthy. People live here not on platform ads but on donations, subscriptions and affiliates (ch. 69).

Everything in this book comes down to a few rules. Half are about what to always do. Half are about what to never do, even when you badly want to and “it works for everyone.”

Ten rules that always work

• Audience first, content second. Not “what I enjoy filming” but “for whom and why they need it.” Until there’s an answer to the second question, the first doesn’t matter (ch. 4).
• One channel, one promise. The viewer must be able to say in one phrase what you’re about. If they can’t, neither can the algorithm, and it’s the one handing out reach (ch. 46, 62).
• The first two seconds matter more than everything else. Up to half of viewers leave in the first three seconds. The hook isn’t decoration but a condition of the clip’s existence (ch. 11–13).
• Rhythm beats inspiration. Regularity gives the algorithm data and you statistics. One clip a week for years beats twenty clips in a month and silence after (ch. 55).
• Judge by the series, not the clip. Below 15–30 publications there are no conclusions: it’s noise, not signal (ch. 45, 46, 73).
• Change one parameter at a time. Otherwise you won’t know what worked, and won’t be able to repeat it.
• Build your own base. Email, messenger, community, the only asset a ban or an algorithm change can’t take from you (ch. 65, 71).
• Answer people. A duty person in comments and DMs brings more money than another post. It’s the most underrated sales channel (ch. 62).
• Count money, not reach. Every metric should have a safety metric: reach with revenue, leads with the share of qualified (ch. 49, 63).
• Multiply what landed. A successful clip isn’t cause to celebrate but a topic for the next five: the platform just showed you who your audience is.

Ten things you must not do

• Don’t buy followers or activity. It’s a direct road to a shadow restriction, and a “dead” audience kills engagement and reach forever (ch. 71).
• Don’t connect services that require your password. Mass follows, auto-mailings, faked numbers, the trace stays even after you turn them off.
• Don’t re-upload others’ content or leave others’ watermarks. Unoriginal content isn’t deleted, it’s just not recommended (ch. 70, 71).
• Don’t sell head-on to a cold audience. A marriage proposal on the first date; attention and trust first, offer after (ch. 8).
• Don’t chase every trend. One experiment at a time while the core works, or resources spread thin and the channel loses its promise (ch. 73).
• Don’t change everything at once after a failure. That’s not searching, it’s wandering: neither you nor the algorithm can make sense of anything (ch. 46).
• Don’t delete old publications. A clip from six months ago can “fire”: recommendation feeds aren’t tied to the date.
• Don’t argue with the platform on its own field. The rules change without you; build so that a ban on one platform doesn’t kill the business (ch. 71).
• Don’t publish advertising without labeling. Fines and blocks cost more than any integration (ch. 70).
• Don’t measure success by others’ numbers. A million views on a meme channel and ten thousand on a B2B channel can be worth the same, or the reverse (ch. 74).

The rules are boring, yes. But breaking them is expensive and keeping them is profitable. Pick your emotion.

Price is the only element of marketing that brings money; all the others spend it. Yet price is handled most carelessly of all: the most common business mistake is cutting the price for volume instead of proving the value. Meanwhile, in a mature business, raising the price by just 1% can lift profit by tens of percent, because that percent falls straight into the margin.

Chapter 17 covered the psychology of price, anchors, scarcity, framing the offer. Here is the foundation under it: how price is structured at all, where its floor and ceiling come from, and what strategies exist beyond “looked at competitors and set it a bit lower.”

The break-even point: the figure you can’t set a price without

Before arguing about “expensive/cheap,” calculate how many units you must sell to break even. Fixed costs (rent, salaries, subscriptions) don’t depend on volume; variable ones (unit cost, fees, shipping) grow with each sale.

Elasticity: are you allowed to raise the price

Demand elasticity is how much demand reacts to a price change. You raised the price 2%, demand fell 10%, demand is elastic, the increase will kill revenue. It fell 1%, inelastic: raising the price is worthwhile. Demand is insensitive to price when: the product is unique or prestigious; the spend on it is small compared to the benefit; the one paying isn’t the one choosing (corporate purchasing); switching to a substitute is expensive or troublesome. For prestige goods the curve is partly inverted: a higher price means higher perceived quality and demand, up to a point.

Two strategies for a new product

A third, underrated path is target costing: first choose the price the market will buy en masse, then design the product to that price. Not “we made it and calculated what it costs” but “we decided what it should cost and figured out how to fit.”

Price isn’t one: line architecture

• Assortment steps. Prices within a line must reflect differences visible to the customer: if the gap between the basic and advanced version feels smaller than the price gap, the whole line seems overpriced (echoes the product ladder, ch. 16, 29).
• “Razor and blades.” The main product cheap, the mandatory consumables at a high markup: printers and cartridges, consoles and games, coffee machines and capsules. Works only with a real tie of the consumable to the product.
• A two-part tariff in services. A fixed entry fee plus a variable one for consumption: a base fee + minutes, park entry + rides, a subscription + a surcharge for volume.
• Bundles and passes. A package cheaper than the sum of parts, people take more and upfront; by-products of production (what used to be thrown out) can subsidize the price of the main item.

Seven adjustments to a set price

How to raise your price and how to answer someone else’s

You can raise it quietly, cancel discounts, drop cheap items, raise the free-shipping threshold, or openly. Openly is better: explain the reason and warn in advance, letting people buy at the old price; this turns the increase into a sales occasion. A separate fork is shrinking the product instead of the price (less volume, a simpler recipe, “shrinkflation”): economically the same, but when the customer notices it themselves, trust is lost forever, it’s more honest to say it straight.

When a competitor changes the price, don’t answer by reflex. Five questions before reacting: why did they do it (market share? capacity load? the start of a price war?); is it temporary or permanent; what happens to our share if we don’t respond; will the others respond; what will they do in response to our response. Often the right answer isn’t to lower the price but to raise the service level or strengthen the offer (the same logic as in the “expensive” objection, ch. 49).

The copywriting formulas (ch. 16) answer “what to say.” This chapter is about “how to say it”: the rules of cleaning text so it gets read to the end. They’re duller than formulas and more important: a brilliant structure in murky text doesn’t work, while a simple one in clean text sells.

The order of work: meaning → visual → text

Any content unit is assembled top-down. First the meaning: what it’s for, for whom, what the person should do or feel. Then the visual solution: what to show it with (the reader latches onto the headline and image before the text). And only then the words. The same order applies when checking someone’s or an AI’s text: first “does it solve the reader’s problem,” then “is the right thing shown,” and only at the end the commas. Four questions before you start: what value does it give? why will people read it? whom does it interest? and “I’m writing this, to what end?”

The word level

• Bureaucratese, under the knife. Not “a water-supply interruption may be observed” but “the water’s off on the 5th, fill up in advance.” The reader is busy and reads with half an eye; every extra word is a tax on their attention.
• Verbs enliven. Nominalizations (“effectuation of delivery”) become verbs (“we deliver”); two verbs in a row compress into one (“helps to maintain” → “maintains”). By the way, a glut of nouns is the main smell of AI text: when editing an LLM’s output, start there.
• Specifics instead of judgments. “Convenient,” “quality,” “profitable” mean something different to everyone. Instead of “a convenient mug,” “doesn’t burn your hand when there’s boiling water inside”: a usage scenario paints a picture, a judgment paints nothing.
• One sentence, one thought. A guide: no more than two grammatical clauses. The important part, at the start of the sentence and paragraph: people scan, not read.
• Parentheses and asides are enemies of rhythm. A parenthetical can almost always be opened into an ordinary sentence; “however,” “it’s worth noting”, delete with no loss of meaning.

The paragraph and structure level

One paragraph, one complete thought; a new one appeared, a new paragraph (a solid slab doesn’t get read, whatever’s written in it). One text, one thought: “by the way…” digressions destroy comprehension, turn them into separate posts. The skeleton of any text: a short intro → body → a conclusion with an action (a question, a takeaway, a call); a half-page intro is a body pretending to be an intro. A list is a sentence with homogeneous parts: a summarizing phrase with a colon before it, items in the same grammatical case, an optimum of 5–7, details in a paragraph after the list, not inside the items.

Bureaucratese effectuates a negative impact on read-through. See? It hurts.

Chapter 19 gave the personal brand a voice (tone of voice), chapter 24 the economics. A third layer is missing, the deepest: the role. Why is one expert listened to open-mouthed while another, with the same knowledge, is scrolled past? Often the difference isn’t in the content but in the fact that the first has a readable archetype, while the second sounds “like all normal people,” that is, like no one.

An archetype is a stable role the audience recognizes instantly and without explanation, because it knows it from thousands of stories: rebel, sage, jester, caregiver. The idea goes back to Jungian psychology and has long been a working branding tool: a buyer trusts faster someone whose role they already understand. For a personal brand this matters doubly, the person is the product (ch. 24), and the audience subscribes not to “useful information” but to a character with a personality, a stance and a predictable manner.

Twelve archetypes: a map of roles

An archetype is a filter, not a costume

A working archetype changes decisions, not just the “packaging.” It determines: which topics you take and which you ignore on principle; what vocabulary you use (a rebel can cut with words where a caregiver can’t); how covers and previews look (ch. 60); how you answer in comments; what you never advertise. Test any decision with one question: “Would my character do this?” If you have to think hard to answer, the archetype isn’t chosen yet, it’s assigned.

“Just be yourself”: the most harmful advice for a personal brand

If a blog is a source of warm words for you, be yourself and enjoy it. But if the blog is an earning tool, “just be yourself” turns into self-sabotage. You’re different: today tired, tomorrow anxious, the day after inspired. Publishing that whole amplitude means having no face at all. A person is many-sided, they have a dozen social roles: parent, entrepreneur, friend, someone’s child. A personal brand, though, is built on one chosen role with measured touches of the rest: “an entrepreneur whose family occasionally shows,” not everything at once in equal parts.

The whole book up to this point answered “what to do.” This chapter is about what you need to be able to do. A blog is a craft of half a dozen skills, and the good news is that none of them is innate: on-camera charisma, a “trained voice” and a “sense of editing” are hours of practice, not a genetic lottery.

An author’s competency map

The levels are honest: “minimum” is the threshold below which the channel stalls; “strongly” is the level that becomes a competitive advantage. You don’t need to level everything up at once, see the priority formula below.

Talking to camera: a craft, not a talent

The difference between a “stiff” and a “charismatic” author is 90% trainable things. Here they are, in descending order of importance.

• Energy ×1.5. The camera eats a third of your live energy: a delivery that feels slightly excessive in person looks normal on screen. Speak the way you’d tell a story to a friend across a noisy table, don’t “perform,” but tell it, only louder and brighter than usual.
• Pace and pauses. The main marker of uncertainty isn’t a tremor in the voice but fillers: “uhh,” “well,” “like,” “actually.” Cured by one technique: a pause instead of a filler. A half-second silence on the recording looks like confidence; “uhh” of the same length, helplessness. Pauses are also easy to cut in editing, fillers aren’t.
• Looking into the lens. Look into the camera, not at your own screen: the viewer feels a direct gaze as a personal address. A trick: a dot sticker beside the lens or a shrunken teleprompter window right under it.
• Points instead of a teleprompter. Reading off a teleprompter is always audible, the eyes dart, the intonation is flat. The working scheme: a plan of 3–5 points in large type beside the camera; each point is spoken freely. Slips aren’t scary, dead smoothness is.
• The structure of a spoken thought. Each meaning block: point → example → takeaway, in one breath. Whoever gets tangled usually doesn’t “speak badly” but hasn’t decided in advance what they want to say, that’s fixed at the script stage (ch. 14), not by diction.
• Diction and voice. Five minutes a day: tongue-twisters before a shoot as a warm-up; reading a paragraph aloud with exaggerated articulation; recording and listening back to yourself (the first times hurt for everyone, that’s normal and passes).

The first thirty takes are awful for everyone. You’re not the exception, you’re the statistic.

Four consolidated checklists for the key moments of the work. Keep them handy: every item is covered in its own chapter, here it’s just the control.

This guide synthesizes widely accepted marketing frameworks and current industry practice of the 2020s. Below: important caveats on how to treat the data given.

What parts XIV–XV rest on

The sections on paid traffic and the content engine rest on four classes of sources: annual surveys of marketing directors on budget structure; peer-reviewed studies of advertising incrementality (including the well-known experiment that switched off brand advertising); aggregated industry benchmarks for impression and click costs; and platforms’ public ad libraries, where others’ campaigns are visible. Exact figures in these areas go stale within months, so throughout the text they’re given as orders of magnitude with a year, and the conclusions are built on mechanics that outlive a change of figures.

Types of sources used

In preparing this, categories of open materials were used: industry blogs and help pages of ad platforms (Google, Yandex, VK, Telegram and others); publications on product marketing and unit economics; materials on CJM/UJM and customer experience; reviews of digital-marketing trends of the 2020s (short video, social commerce, AEO/GEO, AI, first-party data, closed communities, hyperlocal advertising); widely accepted copywriting and offer frameworks (AIDA, PAS, BAB, 4U, the value-equation offer). All statements are reworked and stated in the author’s own words; specific brands are given as examples of tools, not as advertising.

How to keep your knowledge current

Marketing changes faster than any book can be printed. Keep a finger on the pulse: follow the platforms’ official help pages, test hypotheses on your own data, keep a “library” of working combinations, and once a quarter review your strategy against new formats and algorithms.

The practical guide “Unswipable: how to become the one they don’t swipe past” · 2nd edition, the 2020s · 79 chapters · From the basics to a self-sustaining system. Compiled as a working tool: apply it, test it, adapt it to your niche.

Answer aloud or in writing. “Well, I sort of got it” is not an answer. Next to each question is the chapter to return to if you have no answer. It’s not shameful to “not know,” it’s shameful to think you know.

Block 1. Foundation and value

1. How does a need differ from a want, and a want from a demand? Why does this distinction cost money? (ch. 36)
2. Why is “promise less, give more” not about kindness but about the math of satisfaction? (ch. 36)
3. Your bonus to customers became the norm and stopped delighting. What did you do wrong at the very start? (ch. 36, 41)
4. Why is “a 35-year-old man” not an audience, while “needs to eat fast at the wheel” is? (ch. 40)
5. Name three of your competitors who don’t sell what you sell. (ch. 40)

Block 2. System and metrics

6. The five stages of AARRR. At which one is your bottleneck, and how do you know? (ch. 41)
7. What’s the “aha moment” in your product? If you can’t name it, what does that mean? (ch. 41)
8. A customer is unhappy. Why is it too early to panic at 20% negativity, but at a steady 15% it’s already time? (ch. 41)
9. Calculate your customer’s CLV. Now say: how much are you entitled to pay to acquire them? (ch. 39, 52)
10. What are you the “-est” at, by the EST model? The answer “at everything” counts as “at nothing.” (ch. 42)

Block 3. Content and algorithms

11. Why can a clip that landed on “the wrong” audience harm the next five? (ch. 46)
12. What does the algorithm do with your video in the first hours, step by step? (ch. 46)
13. Super-system, system, sub-system: which gives reach, which gives the sale, and why exactly so? (ch. 44)
14. Your lead magnet ended with the words “you’ve got something to think about.” What did you just lose? (ch. 45)
15. A person didn’t open the material. Describe your follow-up chain by timers. (ch. 45)
16. How does atomization differ from “cutting up whatever’s left”? When is it planned? (ch. 55)

Block 4. Money and traffic

17. How much does your “free” post cost? Count honestly: time, salaries, production. (ch. 50)
18. Why does a good creative lower your cost per impression? What do the platform’s interests have to do with it? (ch. 51)
19. An ad shows a great ROAS. Why does that still prove nothing? (ch. 52)
20. Name three signs that it’s still too early for you to launch paid traffic. (ch. 52)
21. Hook rate is low. Hold rate is low. CPA is bad. What do you fix first, and why exactly that? (ch. 56)

Block 5. A system that works without you

22. Funnel or loop? Describe your main loop by the “what, who, why” formula. (ch. 54)
23. What in your business accumulates, and what burns up the moment you stop paying? (ch. 54, 59)
24. Where exactly in your AI conveyor does a human stand, and why there? (ch. 58)
25. What in your last piece could only have come from your personal experience? If nothing, why does it exist? (ch. 58, 59)

Block 6. The psychology of influence

26. The 60-30-10 rule: what was the accent in your last creative spent on, and wasn’t it spent on everything at once? (ch. 60)
27. Name the Cialdini lever you use, and prove there’s truth behind it, not decoration. (ch. 61)

If you’re here, the cartridge is in: you’ve gone from the attention economy to archetypes, you can tell a system from busywork, and you know why you won’t be swiped past now. Thanks for your time, the scarcest resource, and the very one this book is about.

One request instead of an epilogue. This book is alive, and it improves through your feedback. Spotted an error, disagree with a claim, have a case that would make a chapter stronger, open the “Comments, edits and questions” panel and write. The feed is open to every reader; the author flags the valuable notes and pins them near the top, and readers’ likes push a note higher still. Nothing is added automatically, though: what actually enters the text is decided by the author together with AI, after the point is weighed and checked for accuracy. Contributions that make the cut are credited to your handle, that’s how this book grows edition by edition.

And one last thing. The knowledge in this book is worth nothing until it turns into a published post, a launched test, a fixed funnel. The one-question rule from chapter 75 works here too: close the book and do one thing today.