Audience research: where real content ideas come from

The reason invented ideas underperform is not that they are badly made. It is that they answer a question nobody asked. Your buyers have already written the real questions down in six places — this is how to go and read them.

Last updated 30 August 2026

There is a specific kind of video that gets 300 views: technically fine, decently shot, clearly the product of someone sitting down and thinking hard. It fails because thinking hard is the wrong input. You cannot deduce what a stranger is confused about; you can only go and find out.

The good news is that they have already told you, at length, in public. Audience research is the unglamorous business of collecting what they said and using their words instead of yours. It is a couple of hours once, then twenty minutes a month, and it is the difference between an idea list you have to force and one you have to trim.

The rule that makes all six sources work

Collect verbatim. Not a summary of what people ask — the actual sentence, with their spelling, their word order and their hedging intact.

This matters because the summarised version is where the value evaporates. “Clients worry about cost” is not usable. “Is it worth it if I’m only going to use it a couple of times a month?” is a hook, a title, and the whole video, already written by the person who needed it. The gap between those two sentences is the entire discipline.

1. The first five minutes of a sales call

The highest-value source you already own. Before you have explained anything, the person on the call tells you what they think the problem is, what they have already tried, and what they are afraid of. Every one of those is a video.

Practically: keep a running note, and after each call write down the three questions they asked and the one objection they did not quite say out loud. Ten calls gets you thirty questions, most of them repeats — and the repeats are the important part, because a question asked by seven people out of ten is a video that seven out of ten strangers also want.

2. Three-star reviews — theirs, not yours

One-star reviews are usually about something going wrong operationally. Five-star reviews are usually adjectives. Three-star reviews are where people explain, in careful and slightly disappointed detail, exactly what they expected and what they got.

Read the three-star reviews of your five closest competitors. You are looking for the sentence that begins “I wish someone had told me” or “it’s good but”. Each one is a video that pre-empts the disappointment — and posting it makes you the person who said the thing out loud, which is worth more than any claim you could make about yourself.

3. Search autocomplete and People Also Ask

Type your category into a search box and stop before pressing enter. The suggestions are aggregated real queries, phrased the way real people phrase them, which is almost never how a business would phrase it. Do the same on Instagram, TikTok and YouTube search, because each surfaces a different population.

Two tricks that widen the yield: start the query with a question word (“why does…”, “how long…”, “is it normal…”), and run the alphabet — type your category plus each letter in turn. It is tedious for ninety seconds and produces more usable phrasing than an hour of brainstorming.

4. Forums, subreddits and the group nobody moderates

Reddit, niche forums and Facebook groups are where people ask the question they would be slightly embarrassed to ask you. Sort by top posts of the last year and read the ones phrased as questions.

What you are mining is not the answers — you know the answers. It is the shape of the confusion: which two things people mix up, which step everyone skips, which piece of standard advice keeps getting argued about. A video that resolves a live argument in your industry always outperforms one that restates the consensus.

5. Your own DMs, replies and support tickets

You have been answering these one at a time for years. Every question you have typed more than twice is a video you should have made instead, and the fact that you can answer it without thinking is the reason you have discounted it.

Search your own sent messages for a question mark and read what you replied. That archive is usually a month of content by itself, already in your voice.

6. The comments on accounts one tier above you

Not your comments — theirs. Find the five accounts in your niche that are a step ahead, open their best-performing recent videos, and read the top comments. Two patterns are worth extracting.

The first is the unanswered question: someone asking for the bit the video left out. That video is yours to make. The second is the disagreement — the comment with a lot of likes that says “this only works if…”. That is a genuine gap in the category’s conventional wisdom, and it is pre-validated, because people have already spent attention arguing about it.

Turning language into hooks

Once collected, most of the work is already done. Three moves cover nearly everything:

• Use the question as the first line. Unedited. “Is it worth it if I’m only going to use it twice a month?” needs nothing added to it.
• Answer the question the question is hiding. Someone asking “how much does it cost?” is usually asking “how do I know I won’t regret this?” Answer the second one out loud and you have a far better video.
• Name the mistake in their words. If five people described the same wrong assumption, say the assumption back to them in the first two seconds. Recognition is the cheapest form of retention there is.

What you must not do is translate. The moment you rewrite “is it normal for it to still hurt after a week” into “understanding your recovery timeline”, you have converted a hook into a title and lost the video.

Where to keep it

One document, three columns: the verbatim quote, where it came from, and which of your pillars it belongs to. That is the whole system. Do not tidy the quotes; the untidiness is the value.

Add to it whenever you notice something, and reread it whenever you sit down to film. An idea bank built this way does not go stale in the way a brainstormed list does, because it is made of things people actually said rather than things you thought might be interesting.

How often to redo it

Properly, twice a year. In between, five minutes after each sales call and a monthly skim of one forum is plenty. The questions change slowly — far more slowly than formats do, which is why competitor research is a weekly job and this one is not.

Questions people actually ask

How do I find out what my audience wants to see?

Read what they have already written rather than guessing: the questions asked in the first five minutes of sales calls, three-star reviews of your competitors, search autocomplete, niche forums and subreddits, your own DMs and support tickets, and the comment sections of accounts one tier above you. Collect the sentences verbatim.

What is the verbatim rule in audience research?

Record the exact sentence a person used, not a summary of it. 'Clients worry about cost' is unusable; 'is it worth it if I'm only going to use it a couple of times a month?' is a hook and a whole video. Summarising is where the value is lost.

How is audience research different from competitor research?

Audience research tells you which questions are worth answering; competitor research tells you which structures are working right now. Questions change slowly, so audience research is a twice-a-year job. Formats and hooks shift constantly, so competitor research is weekly.

Why do my content ideas not perform even though they are well made?

Usually because the idea answers a question nobody asked. Production quality cannot rescue a video aimed at a problem the viewer does not have. Ideas sourced from things real buyers said outperform invented ones for that reason alone.

How long does audience research take?

A couple of hours to work through the six sources the first time, then about twenty minutes a month to keep it current - five minutes of notes after each sales call and a monthly skim of one forum.