5 ChatGPT Prompts for Blog Niche Selection

Picking a niche is not about finding a topic you can write about for years. It is about finding the overlap between what you can write, what people search for, and what somebody will eventually pay to reach. Most advice skips the third part.

What’s on this page

A candidate test that scores niches you are already considering.
An audience-versus-topic check, because “tech” is not a niche.
A monetisation path map for each candidate.
A durability test for topics that survive algorithm changes.
A first-90-days validation plan.

The prompts

1. Candidate Scorer

Here are niches I am considering: [list 3-6 niches]. Score each one from 1 to 5 on: how specific the audience is, how easy it is to name ten post ideas, how likely a commercial product exists for that reader, and how much you would have to repeat yourself in year two. Show the scores in a table and state which one you would drop first and why.

What it does: Forces a comparison instead of letting you fall for the idea you are emotionally attached to.

How to use: Fill the list with real candidates, not hypothetical ones. If you cannot name ten post ideas for one, delete it before you score.

2. Audience Narrowing

My topic is [topic]. Narrow it into three genuinely distinct audiences, each described by the situation they are in rather than by demographics. For each audience, give an example of a problem only they would search for.

What it does: Converts a broad topic into a specific reader, which is what makes a blog findable.

How to use: Replace [topic] and pick the audience whose problem you can describe without researching it.

3. Monetisation Path Map

For the blog niche [niche] and audience [audience], list every realistic way this blog could earn money: advertising, affiliate products, digital products, services, sponsorship, community. For each one, describe what has to be true before it works, and roughly how much traffic or audience depth it needs to be worth launching. Do not give revenue figures.

What it does: Shows what you are building toward and which paths are compatible with your traffic reality.

How to use: Replace [niche] and [audience]. Pick one primary path; a blog chasing all six earns from none.

4. Durability Test

For the niche [niche], split typical post topics into three groups: evergreen questions that will still be asked in three years, topics tied to a specific tool or platform, and news-dependent topics. Then tell me what percentage of my content plan should come from each group.

What it does: Protects you from building a blog that dies when one platform changes its rules.

How to use: Weight your first thirty posts toward the evergreen group, and keep the news group for short pieces you can retire.

5. First-90-Days Validation

Design a 90-day validation plan for a blog about [niche] for [audience]. Break it into three 30-day blocks: what to publish, how many posts, what to measure, and the decision rule at the end of each block that tells me to continue, narrow, or stop.

What it does: Turns niche selection into an experiment with a stop condition instead of a leap of faith.

How to use: Decide the decision rules now, while you are unbiased. Changing them at day 60 is how people talk themselves into continuing a dead niche.

How to adapt these prompts

If you are building for a business rather than for yourself, add “the audience must overlap with people who could buy [product]” to prompt 2.
For a YMYL-adjacent topic such as health or finance, add “flag anything that would require a qualified professional to review” to prompt 1.
Run prompt 4 again once a year — the ratio shifts as your niche matures.

FAQ

How wide should a niche be?

Narrow enough that one sentence describes the reader, wide enough that you can list fifty post ideas. If it takes two sentences, it is probably two blogs.

Should I pick a niche with lots of competition?

Competition usually means the audience exists. The better question is whether the top results are beatable on specificity — a narrower problem is often untouched.

How long before a new blog earns anything?

Long enough that you should have a second reason to keep writing. A plan that only survives on eventual ad revenue tends to stop around month four.

Keep going

GuPrompt keeps every prompt copy-paste ready. No account needed.