Search depth: video that works for years

Search depth: video that works for years

YouTube is built like a search engine rather than a feed: a video gets found by a query a year after publication. Hence different subjects, different titles and different economics — one piece of material works for months. Below is what follows from that for content and how to choose subjects.

YouTube is built like a search engine rather than a feed: a video gets found by a query a year after publication. Hence different subjects, different titles and different economics — one piece of material works for months. Below is what follows from that for content and how to choose subjects.

---

The video got two hundred views and died

That is what almost everybody who tried YouTube and gave up concludes. They recorded a video, put it up, it picked up a little in a week, then silence. Conclusion: it did not work, we should have done short formats.

The conclusion is wrong, and here is why. In social media material lives for hours. On YouTube, for years.

A video filmed today can pick up most of its views eight months later, when people start finding it by a query. Judging it after a week is like judging a planted tree after a week.

That changes everything: what to film, how to measure, and why a business wants this at all.

YouTube is a search engine

The root cause of most mistakes: people treat it as a social network.

In a social network material gets shown to people who follow you or resemble them. Here the person types a query themselves — they have already formulated a problem and are looking for the answer.

Three consequences that invert the approach.

The subject is chosen by query rather than by interest. Not "what do I feel like talking about" but "what do people search for, in their words". Nobody searches for "My thoughts on the future of the profession". People search for "how much does a landing page cost" every day.

The title matters more than the content. Not because content is unimportant but because nobody reaches it if the title did not match the query.

The audience is colder than anywhere. Somebody does not know you and did not come to you — they came for an answer. Trust has to be earned in the first thirty seconds.

The misconception: "we need lots of videos". No. You need a few that land on real queries. Twenty videos about nothing lose to three that sit in the results.

What to film

Three types that work in search.

An answer to a specific question. The most reliable type. Somebody is looking for a solution and you give it in full. "What to say when a client asks the price", "How much does a landing page cost", "What to do when they ask for a test task".

An analysis of a real case. A format almost nobody uses: from first contact to result, with figures and with what went wrong. People watch to the end because they want to know how it turns out.

An analysis of a mistake. Performs above average and is almost unoccupied. Plus it is the fastest way to earn trust from a cold viewer.

What not to film: news and release round-ups. Aggregators produce those industrially and they age within a month — meaning they contradict the very nature of a platform where material should live for years.

How a video that gets watched to the end is built

The first thirty seconds decide everything. You cannot introduce yourself and talk about yourself here. Here you name the problem in the viewer's words and promise a concrete outcome: "We'll go through what to say when a client asks the price outright — I'll show you the actual conversation."

One question, one video. The temptation to pack everything in kills the view count: somebody was looking for one answer and got a lecture.

Evidence inside. A screen, a document, figures, a real example. Talking to camera with nothing shown loses to an analysis with the material on screen.

And a rule of the platform people forget. YouTube restricts monetisation and reach for coarse language. A register that passes in Telegram costs money here. No swearing at all.

What to measure

Not views in the first week. They mean nothing on a search platform.

Retention. Where people leave. A mass drop-off in the first minute means the title did not match the content. In the middle means there is padding there.

Traffic source. Whether they came from search or from recommendations. Search is what all of this is for: it does not depend on algorithmic luck and does not run dry.

Clicks onwards. How many people followed the link. That is the only metric connected to money.

The assessment horizon is six months. Judging earlier is not possible. A page enters the results in two or three months, and that is a normal timescale.

Title and thumbnail

The two things that carry most of the decision to watch.

The title is written in the words of the query. Not as it looks nice but as somebody phrases the problem. "How much does a landing page cost" rather than "Pricing in web development". The test is simple: could you type this into a search box yourself?

The thumbnail repeats the promise rather than decorating. Three or four words large, readable on a small screen. Not the whole title and not an abstract picture.

The title and the content must match. A mismatch produces the worst possible result: people arrive and leave immediately, and the platform stops showing the video at all.

And one expensive mistake: optimising the title for a video that has already been filmed. The correct order is the reverse — the query first, then the script for it.

Why a business wants this

Three reasons, and none of them is about fame.

Search traffic does not run out. A video that lands in the results brings people in for years with no further spend. That is an asset rather than a post.

Video explains what text cannot. Showing how something is done on screen beats any description. For complex products, or ones people are wary of, that is the decisive argument.

One video feeds every other platform. A recorded video turns into a reel, a carousel, a series of stories and a post. So YouTube is not additional work but a source of raw material for the pipeline.

The last point changes the economics: filming becomes cheaper than not filming.

How many videos are needed and how often

The answer is counterintuitive: not many, and regularly.

Five to seven videos landing on real queries work better than fifty random ones. A search platform does not reward volume — it rewards matching a query and being watched to the end.

Rhythm matters more than quantity. One video a week, or even a fortnight, but consistently. A burst of ten videos in a month followed by silence produces a worse result than a steady pace.

And the horizon. For the first three months you work blind: the page is not yet in the results, the views are random, there is nothing to draw conclusions from. That is the most common moment people quit — and the worst one.

What the business ends up with

By the methodology I use, YouTube sits inside the social packaging stage — 4–5 days for all the platforms. Part of the full positioning package: 3–4 weeks, 25–35 hours, around $1 900. Honestly: that is the price of the service by our methodology, not market statistics.

Do it yourself or commission it

Do it yourself. The mechanics above are the "YouTube" block from "Freelance 2.0": choosing the channel concept, planning videos, and script generators separately for the expert and the personal formats. All work from the knowledge base — the same one feeding the other platforms.

The course costs $9 and is included in Basic. Registration is free and opens three days of full Basic access — plus 140+ AI assistants across 14 categories and the ten-lesson sales module.

Commission it. The service exists. The sign it is being done wrong: if a contractor promises results within a month, they do not understand how a search platform works, and you will be paying for an expectation that will not be met.

Do one thing today: write down the five questions clients ask you most often, in the words they use. Those are ready titles for your first five videos, and they already match real queries because they came from live speech.

---

How this looks on a real project

What follows is the same work on a live client. Fomin Studio, "the Fomin Method": two doctors, Zakhar and Alexandra, husband and wife. The work was delivered as eleven sections, the seventh being YouTube as search depth.

Why this example is here. It shows that a positioning package is not an abstraction but a service with a clear composition. And it sells to any business that has an audience and a product: a clinic, a school, an agency, a shop. The mechanics are the same; the contents change.

An analysis of that deal from the first message to the payment is included in the sales module on the Basic plan.

The platform's job

Nobody scrolls here — people search for an answer and are willing to watch at length. Somebody searches "why does my jaw click", finds a long explanation and goes into the bot.

The main principle: editing, not filming

One filming session — one day, all 29 subjects live: solo, as a pair, in the clinic.

From there it unfolds into two formats: a long cut of 3–10 minutes with chapters, timestamps and a search-driven title, and Shorts of up to 60 seconds — the same hook clip pointing at the long version.

The cost of an adaptation = editing the cut + the Short + the title + the thumbnail. No reshoots.

Four differences the other platforms do not have

A search title is a query rather than a provocative hook. "Why does my jaw click in the morning", not "you won't believe what happened".

Chapters and timestamps — the viewer jumps to the part they need and retention rises.

Thumbnails with faces — the doctors' faces plus three or four words. For case studies the result is shown.

The description leads into the bot — a hook, the chapters, the code word and a link to the same bot. No separate funnel needed.

Five playlists from the same pillars

The content plan's pillars renamed as a shop window for the viewer: "The cause everybody missed" — the flagship · "Genuinely not frightening" · "Two doctors behind the scenes" · "Plain truth" · "Honestly about aesthetics".

The rhythm: around two long videos a week plus two or three Shorts. Of the 29 subjects, 21 go out as long form, 3 as long form with a teaser, 1 vertical only, and 2 as a written guide with no video.

---

What to read next

[Finding trends](/en/blog/finding-trends) — how to find subjects that match queries.

[Editing expert video](/en/blog/expert-video-editing-pipeline) — how to produce volume without a studio.

[The content factory: a nine-week plan](/en/blog/content-factory-nine-weeks) — where the substance comes from.

[Turning a blog into a client source](/en/blog/blog-as-client-source) — the whole chain of eleven stages.

[AI video generation in 2026](/en/blog/ai-video-generation-guide) — the full guide to video models.