Almost every social media brief has a line like “Target audience: women 25–45, middle income, interested in beauty and wellness”. That description fits a dental patient, a sneaker buyer and a food blog follower equally well. You can’t write a post for it: it doesn’t tell you what to say, to whom, or why now.
The problem isn’t missing research but the wrong research. Age and gender answer “who”. Content needs different answers: in what situation someone thinks of you, what they use instead today, and what stops them from buying. Those answers live in your DMs, reviews and five conversations with customers.
Below: where to find the data in one evening, what to fill in instead of a persona, and how to turn two or three segments into content pillars. Running example: a dental practice, illustrative numbers.
Why “women 25–45” doesn’t work
Demographics help with ad targeting but say almost nothing about what to post. Two 34-year-old women from the same neighborhood see a dentist for different reasons. One brings a child who needs a first filling without tears. The other put off an implant for five years and finally decided. They need different posts, arguments and even posting times.
So use jobs to be done instead: people “hire” a product to get a job done in a specific situation. A segment is people with the same job, not the same age. Age, city and income stay on the card as detail.
Where to get data: five sources in one evening
| Source | What to look for | Time |
|---|---|---|
| Instagram DMs, WhatsApp, email from the last 3 months | Repeated questions, customers’ own words, where chats go quiet | 60 min |
| Reviews on Google Maps, Yelp, Trustpilot (yours and 3–5 competitors) | What people praise, what they complain about, what came “before” | 40 min |
| Account insights | Location, age, active hours, where followers come from | 20 min |
| Competitors’ comments and niche communities | Objections people raise in public | 30 min |
| Google autocomplete, AnswerThePublic, Reddit | How people phrase the problem themselves | 20 min |
- Conversations are the main source. Scroll through your last 100 chats and copy the questions word for word. “Does it hurt?”, “Can I pay monthly?”, “What’s the total cost?” Each is a post topic with its first line written.
- Read reviews from the middle. Three-star reviews are more honest than five-star ones. Look for “finally”, “I was afraid that”, “before this I went to…”: they reveal the situation and the alternative.
- Insights. Instagram shows follower demographics from 100 followers. YouTube Studio’s Audience tab shows other channels your viewers watch, a hint about where else your segment spends time. What each Instagram metric means is in our guide to Instagram Insights.
- Competitors aren’t there to copy: look for objections they leave unanswered. The method is in our piece on competitor analysis.
A segment card instead of a persona
One card per segment, six lines. If a line can only be filled with a guess, mark it with a question mark: interviews will check it.
| Field | Question | Example: “put off an implant” |
|---|---|---|
| Situation | When do they think about the problem? | A bridge broke, chewing hurts, a family wedding is coming |
| Job | What result are they hiring for? | Eat and smile without thinking about it, in one course of treatment |
| Alternatives | What do they do now? | Live with it, dentures, dental tourism abroad |
| Barriers | What stops them? | Fear of pain, paying it all at once, distrust of warranties |
| Criteria | How do they compare options? | Warranty in years, financing, before/after photos, the dentist’s experience |
| Words | What do they call it? | ”Replace a tooth”, “a post”, “something that won’t fall out”, not “dental implantology” |
People skip the last line most. A patient searches “replace missing tooth cost” while the profile says “premium implant solutions”. Matching words in your profile name, first line and cover text affect Instagram search and whether someone recognizes their problem instantly.
A 15-minute interview script
Five to seven conversations with people who bought in the last three months beat any survey. Usually by the fifth, answers start repeating: the segment is described. Offer something small for their time, like a free cleaning.
- “Think back to when you decided to deal with this. What happened?”
- “What did you try before us? Why didn’t it work?”
- “Who else did you consider, and why did you pick us?”
- “What almost stopped you from booking?”
- “How would you explain what we do to a friend?”
The key rule: ask about past behavior, not hypotheticals. Almost everyone says yes to “Would you buy if…”; “When was the last time you…” actually works. Write answers verbatim, slang included: that’s your Words line.
How many segments and how to choose
List every segment you found and score each from 1 to 3 on four criteria:
| Segment | Size | Average order | On social | Urgency | Total |
|---|---|---|---|---|---|
| Put off an implant | 2 | 3 | 3 | 2 | 10 |
| Parents of kids 5–12 | 3 | 1 | 3 | 3 | 10 |
| Whitening before an event | 2 | 1 | 3 | 3 | 9 |
| Out-of-town patients | 1 | 3 | 2 | 1 | 7 |
Work with the top two, three at most. The rest just don’t get their own pillars. With two or three hours a week for social media, even three is a lot: each needs at least one post a week, or it won’t notice you’re talking to it. How segments fit the bigger picture is in our one-page social media strategy.
From card to content
Each field becomes its own type of post:
- Situation → “what to do if” posts. “Broke a bridge before a big event: what can be done in a week”. It lands exactly when the problem is urgent.
- Barriers → objection-handling posts. “What goes into the price of an implant and how to spread it over 12 months”. For fear of pain, a Reel where the dentist shows the anesthesia and answers three common questions.
- Alternatives → honest comparisons. “Implant, bridge or dentures: who each one suits”. Don’t trash the alternatives; explain when each makes sense.
- Criteria → proof. Before/after with patient consent, the warranty in numbers, one case from first visit to result.
- Words → headlines and keywords. In your profile name, first lines and cover text.
A quick test for any post: which segment is it for? If the answer is “everyone”, it’s probably for no one.
How to check you got it right
Audience analysis is a hypothesis until content confirms it. For two or three weeks, post for each segment and watch intent signals, not likes:
- Saves and sends. The post gets forwarded to people with the same job.
- DMs with questions from the card. You hit a barrier.
- Story polls. “What’s holding you back: price, fear, time?” tests the Barriers line on real followers.
If a segment ignores every format for three or four weeks, either they’re not on this platform or you got the job wrong. Go back to conversations, not demographics.
When this method won’t help
- Nobody needs the product. No card sells something without demand.
- You have a few dozen clients and know each one. In niche B2B, calling five of them beats building segments.
- Under 100 followers. No demographics in Insights yet: rely on conversations and competitors’ reviews.
- No capacity to create for each segment. Pick one and write only for it rather than spreading thin.
Segment cards pay off when they’re part of every post, not sitting in a folder. In Repost21 you can keep them in Google Docs or Notion connected to your brand knowledge base: Luna drafts posts for a specific segment in its own words, and account analytics show which pillars earn saves and sends. You can try it free on your own audience.