How to find trends and write to them before the subject cools
People notice a subject once everybody has written about it, and their piece comes out into an already crowded feed. The problem is not writing speed: you learn about a subject from what got published, so by definition late. Below is how to watch conversations rather than posts.
People notice a subject once everybody has written about it, and their piece comes out into an already crowded feed. The problem is not writing speed: you learn about a subject from what got published, so by definition late. Below is how to watch conversations rather than posts.
---
The subject got noticed once everybody had written about it
A classic story for anybody making content. You see a subject has taken off, sit down to write — and come out with a piece when the feed is already full of the same thing.
The result is predictable: mediocre reach, weak response, a sense of perpetually catching up.
The problem is not the writing speed. The problem is that you learn about the subject from the same sources as everybody else — from material already published. Which is by definition late.
There is a way of looking not at posts but at conversations — where a subject lives before it gets shaped into posts.
How finding trends differs from reading news
News is an event already selected by an editor. By the time it is published it has passed through a filter: somebody decided it mattered, wrote it and put it out. Everybody reads the same thing because the source is one.
A trend is a change in what people are talking about. It is visible earlier and in a different place: in ordinary people's posts, in comments, in which questions started getting asked more often.
Hence a practical distinction. News round-ups give you what everybody gets. Watching conversations gives you what nobody else has yet.
And a second difference — in usability. News has to be retold, and a retelling interests nobody. A trend can be used: material gets written to it that answers a question people are already asking.
What to look for besides the subject
Three things usually missed.
The phrasings. Not only what people talk about but in what words. That is direct raw material for headlines: a headline written in the audience's words lands more precisely than any invented one.
The objections. What irritates people about the subject and what they argue with. Material answering an objection works harder than material repeating the consensus.
The gap between demand and supply. A subject widely discussed with little clear material on it is the most valuable thing you can find. That is the vacant place.
I found gaps like that by counting the shares of subjects among competitors: across 6 502 posts from eleven channels over 19 months, the subject of prices took up nine posts against mass demand for it. Nought point one per cent of the feed — and the first of three reasons people do not earn.
Where to look
Four sources giving different things.
Questions from your own audience. The most precise and the most underrated. A question people started asking more often is a trend nobody else has reached, because it is yours.
Comments under other people's material. There you can see what genuinely concerns people about a subject — and usually it is not what the material itself is about.
Feeds of professional communities. Practitioners' conversations run ahead of publications and with no editorial filter.
Search queries. The coldest and most reliable source: somebody searching has already formulated a task. A subject with rising query numbers and no clear material in the results is a direct indication of what to write.
What not to use as a source: "top news of the week" round-ups. They by definition reach you at the same time as everybody else.
How not to turn into a news channel
An important limit, or trend-finding does harm.
You cannot chase every subject. An audience that already understands a field does not respond to news — it addresses a stage they have passed. From my observations: the share of successful pieces in the news format is below a channel's average.
The rule: a trend is a prompt rather than the content. An event gives you a way into a conversation, and from there you talk about your own thing: what it means for your audience, what to do about it, how it differs from last time.
And a check before you take it on: will this piece still make sense in six months? If not, it belongs in a messenger and stories rather than on a page meant to live in search.
The second layer: ready scenarios
Separately from finding trends there is something that saves more time — ready multi-step scenarios for standard tasks.
The difference from an ordinary prompt is fundamental. An ordinary prompt is a request. A scenario is a pre-assembled sequence: the role, the input data, the order of steps, the format of the result. You fill in your data and get a finished document.
What such scenarios usually cover: a business plan, a marketing strategy, sales, brand work, product ideas, HR tasks, customer support, financial calculations, viral posts.
Why bother if you can write a prompt yourself. For the same reason you use a reference book rather than deriving a formula every time. A scenario assembled by somebody who has done it a hundred times contains questions you would not have thought of first time.
How to write quickly to what you found
A trend loses value in days, so speed matters more than polish.
A prepared frame. Do not invent a structure each time: what happened → why it concerns my reader → what to do about it → my conclusion. Four blocks, and you fill in the content.
Ready generators per format. One thought unfolds into a post, a clip and a carousel — if you have something to unfold it with. A knowledge base assembled in advance makes that a matter of minutes rather than an evening.
The one-thought rule. Material written to a trend does not have to be exhaustive. It has to be first and clear. The exhaustive one will come out in a week and somebody else will write it.
And a check before publishing: did you say something the others will not. If not, it is a retelling of news and better skipped.
Where to start
The "Grok: finding trends" express course works on two levels.
The methodology — eight modules: prompt engineering for the model's architecture, managing context and answer format, a universal template for a multi-step scenario.
The library — fifteen ready scenarios for business tasks: a business plan, marketing, sales, brand, product ideas, HR, customer support, finance, viral posts and others. Each is a multi-step scenario with a role, input data and a set format. You fill in your data and get a document.
Who it is for: entrepreneurs, marketers, product people, writers — anybody needing concrete strategies and copy rather than general advice.
The price is $9. The course is included in the Basic plan, and registration is free and opens three days of full Basic access — meaning you can simply take it having paid nothing.
Alongside it are five more foundational courses, including prompt engineering with a map of all the main models and an encyclopaedia of more than 190 vetted services, the ten-lesson sales module and chats with 140+ AI assistants across 14 categories.
Do one thing today: write down three questions your audience started asking more often over the last month. Not subjects from the news — questions from real people. That is your trend, and it is more precise than any round-up.
---
Three tools that find trends for you
An AI trend analyst — finds top-performing short-form video, analyses virality, collects links and descriptions and produces scripts. From $240 on the market.
The finder for hidden niches on YouTube — a different approach: it finds not popular videos but young channels that broke through in weeks. You enter a query and an age limit for a channel and get a spreadsheet. If a channel is twenty or thirty days old and a video has hundreds of thousands of views, the subject has potential and there is no competition there yet.
A smart digest maker — gathers a feed on a subject and updates a mini-site itself.
What the second tool gives that the first two do not. It shows not what is already trending but where there is space.
The archive — Make, N8N and Full.
---
What to read next
[Analysing a competitor by figures](/en/blog/competitor-analysis-by-numbers) — counting rather than reading a feed.
[The content factory: a nine-week plan](/en/blog/content-factory-nine-weeks) — where found subjects fit in.
[Checking a fact with AI search](/en/blog/fact-checking-with-ai) — the precise search tool.
[Turning a blog into a client source](/en/blog/blog-as-client-source) — the whole chain of eleven stages.
[AI agents: what they are and what people pay](/en/blog/ai-agents-guide) — the full guide to agents, with market rates.