To pick a content niche, stop asking "what topic should I cover" and start asking "which audience do I help with which problem." A niche is a specific group of people plus a problem you can help them with on camera; the topic is just the delivery vehicle. And if you've posted anything before, even 20 messy posts, your analytics already contain a vote on which niche is yours, because shares and saves cluster around the content people needed rather than the content you enjoyed making.
The reason this question paralyzes capable people is that they're capable of too much. If you're good at one thing, the choice is made for you. If you're good at four things, every option feels like abandoning three identities. So let me take some pressure off before we get tactical: you are choosing a starting lane, and you get more than one.
Why "pick one topic" is bad advice for multi-skilled people
The standard advice says niche down to one topic and repeat it forever, and there's a version of that advice worth keeping. Audiences do need to predict what they'll get from you; an account that posts about sourdough on Monday and Series A fundraising on Thursday builds no expectation and therefore no following.
But topic-niching misreads why the predictability works. People don't return for a topic. They return because you reliably help someone like them with something they're dealing with. A career coach can post about salary negotiation, burnout, and difficult bosses, three different topics, and lose nobody, because the audience is the same person the whole time: the mid-career professional who feels stuck. The account with one topic and no defined audience is more scattered than the account with five topics and one audience.
So the unit of focus is the audience-problem pair. Get that pair specific and you can roam across every topic that serves it.
How to pick a content niche with the niche equation
The equation is: niche = who + what's wrong. Here's the worksheet version, and it works even if you've never posted.
- List your raw material. Write down the last 20 to 30 pieces of content you've made, or if you're starting from zero, the last 20 questions people asked you for free: at work, in group chats, at family dinners.
- Mark the evidence. For existing content, flag which posts earned shares and saves, and ignore likes. For questions, flag the ones people came back to follow up on. Both signal need rather than politeness.
- Name who's asking. For each flagged item, write down the person: their job, their stage of life, what they were trying to fix. Look for the repeating face.
- Write the pair. Fill in one sentence: "I help [that person] with [that problem]." If the sentence could describe a thousand creators, sharpen the person, since the problem side usually can't get much sharper.
- Choose three lanes inside it. Pick three pillars, meaning three recurring angles on the pair, and give them a rough 50/30/20 split of your posts.
- Run it for 30 days, then let the numbers renegotiate. Eight or so posts is enough for the data to start talking back.
Your old content already voted
This step deserves its own section, because your instincts about your best content are usually wrong. The posts you're proud of and the posts your audience needed are different lists, and only the analytics know which is which.
Take a nutrition coach I'll call Simone, torn between three directions: meal prep content, hormone health, and building her coaching practice in public. All three felt like her. Rather than choose by vibes, she exported her TikTok analytics, which covers roughly the last 60 days, and looked at shares and saves per video instead of views. The results surprised her: her five most-saved videos were all about eating for energy in your 40s, a theme she'd posted casually and never considered her lane. Meal prep pulled views and almost no saves; snackable, forgettable. Practice-building content earned engagement from other coaches, an audience that would never hire her.
Her pair wrote itself: women over 40 who are exhausted and confused about why. Her pillars: eating for energy at 50 percent, hormone-aware nutrition at 30, and client stories and myths at 20. The meal prep skill didn't die; it became the vehicle, meal prep for the energy-crashed 45-year-old, aimed at the audience the data chose. Sorting your own export the same way takes an evening by hand, or Batchly does it on upload, scoring every video with shares and saves weighted over likes and clustering the winners into named pillars.
Your audience has been voting on your niche the whole time; the ballots are in your analytics.
Why three pillars beat one lane
Three pillars solve the multi-talented person's dilemma without wrecking predictability. Each pillar is a recurring angle on the same audience-problem pair, so the audience always knows who the content is for, while you get enough variety to stay interested for a year. Creator burnout is usually niche claustrophobia wearing a different outfit.
Pillars also de-risk the choice. You're never betting everything on one guess; you're running three small experiments with a built-in scoreboard, and every 30 days the weakest pillar can be replaced. There's a full breakdown of how pillars work in content pillars explained, and if the fear is running dry, three lanes is also the practical fix covered in how to never run out of content ideas.
One warning: three is the ceiling. Four pillars is a variety show, and five is the sourdough-and-Series-A account wearing a system as a disguise.
What if you're starting from zero?
No analytics, no problem, but you do lose the shortcut, so borrow evidence from your offline life. The questions people repeatedly ask you are analytics; so is the topic where friends say "you should charge for this." Write the pair from step 4 using that, pick your three pillars as hypotheses, and commit to eight posts before judging anything. Your first month of posting exists to generate the data you don't have yet, which is why the first 90 days plan treats month one as evidence-gathering rather than performance.
The trap to avoid at zero is niche-by-aspiration: choosing the audience you wish you served instead of the one whose problems you can improvise answers to at a red light. Pick the lane where you have opinions you'd defend, because you'll need 50 videos of them.
Do this next
Tonight, write the sentence: "I help [person] with [problem]," using the repeating face from your DMs, comments, or coworker questions as evidence. Then name three pillars underneath it and write one post for each this week. If you have posting history, upload your export to Batchly first; it finds the pillars your audience already voted for and writes the opening scripts, so your niche decision starts from receipts instead of a guess.