Picking your AI newsletter niche:the four-question test.

If you have ever thought about starting a paid newsletter, you have probably spent at least one weekend "picking a niche". Probably more than one. The activity feels productive — there are workshops, frameworks, IkigAI diagrams, Venn-diagram exercises. The output is almost always the same: another month passes, no newsletter.

Niche selection is a four-question test, not a months-long meditation. If you cannot answer the four questions, no amount of further research will help. If you can answer them, you should commit by Tuesday and start writing Issue Zero this weekend.

The four questions are below. They are stolen, in spirit, from the niche-selection chapter of Bible 02 — but the version here works on its own.

Question 1: Who's the £8/month person?

Forget readers. Forget audience. Forget "would I read this myself". The only question that matters is: who is the person who would pay £8 a month for this newsletter, and what do they currently do instead?

Most aspiring operators answer this with a category — "founders", "marketers", "investors". Wrong granularity. The £8/month person is a specific role at a specific stage with a specific problem. "First-time founders trying to hire their first non-technical person." "B2B SaaS marketers at companies under $5M ARR who do their own content." "Solo property investors managing 3–10 doors in the UK."

The test: can you describe the £8/month person in one sentence specific enough that they would email a friend in the same role and say "have you seen this newsletter, it's exactly for us"? If not, the niche is too broad. Cut it in half and ask again.

Question 2: What's the alternative right now?

The £8/month person is not currently sitting in the dark waiting for you to arrive. They have an alternative — a Slack community, a forum, a podcast, a competing newsletter, a book they keep re-reading, or "asking ChatGPT and hoping". You need to know what it is.

This is uncomfortable, because most aspiring operators don't actually know. They have a feeling that "no one is doing this", which is almost always wrong. People are doing this. They might be doing it badly, or for a slightly different audience, or with a different format, but they exist. Find them.

Do this: open a fresh AI chat and ask:

If the AI struggles to find 8 alternatives, your niche is probably too narrow to support a paid newsletter at this stage. If the AI finds 8 strong alternatives with no obvious gaps, your niche might be saturated. The sweet spot is when there are alternatives, but each has a clear weakness — outdated, too generalist, behind a community paywall, only available in another language, abandoned in 2023.

Question 3: What can you ship every week, with no research, when you're tired?

This is the one most niche-selection frameworks skip. They focus on demand. They forget supply.

You will write this newsletter on Tuesdays at 11pm after a long day. The question is not "could I write this if I had three uninterrupted days". The question is: what can I ship in 90 minutes on a tired Tuesday?

If your only sustainable output is "deep research-based investigative pieces", you will burn out in week 11. If your output is "field notes from running my own thing in this space", you will still be writing in year three.

The test: name the format you can ship every single week, on no budget, with no research staff, when you're tired. Common formats that pass this test:

• One question from a paid subscriber, answered in detail.
• One thing you tried this week, what worked, what didn't.
• One model or framework you used, with the exact application.
• One under-discussed pattern you noticed in your own niche.
• One interview, recorded by phone, transcribed and edited (this one is harder than it looks).

If you cannot name a sustainable format, the niche is the wrong shape for you, even if the demand is real. Don't fight it. Pick a different angle.

Question 4: Why you?

The least romantic question. Most aspiring operators flinch at it. "Why does it have to be me?"

It does have to be you, because the alternative — a generic newsletter that could be written by anyone — gets nowhere. The best paid newsletters in any niche are written by someone whose specific perspective, scars, and weird obsessions are inseparable from the content.

The test isn't "am I the world's leading expert". It's: do I have a specific lens on this niche that no one else has? Lens examples that work:

• You did the thing yourself, badly, and learned the actual lessons. (Operator-as-writer.)
• You sit at a useful intersection. (UK + B2B SaaS + sub-$5M ARR is a specific lens. UK + B2B is too broad.)
• You have access. (You interview 50 people in this niche a year. No one else does.)
• You have an idiosyncratic taste. (You're allergic to the prevailing wisdom and the audience is starting to be too.)

If you cannot articulate the lens, your newsletter will read interchangeably with the alternatives in Q2. That's the death sentence — not failure, but blandness.

The 25-minute version

If you want this in one sitting, run the four questions sequentially with an AI as a thinking partner. Open a chat. Paste this:

Twenty-five minutes. By the end, you'll either have a niche worth committing to or a clear reason to pick a different one. Either outcome is faster than three months of "researching".

If all four answers are weak

Sometimes you run the four questions and every answer is fuzzy. The £8/month person is a category, not a sentence. The alternatives list is short. The format you can ship is "deep essays" (you can't, sustainably). The lens is "I'm interested in this stuff."

That is not a niche. That is a vague enthusiasm. Two paths from there.

Path A — sharpen the niche by halving it. If your niche is "B2B marketers", halve to "B2B marketers at sub-$5M ARR companies who do their own content." Run the four questions again. If the answers are still fuzzy, halve again — "UK B2B SaaS marketers at sub-$5M ARR who do their own content and have a CEO who hates spending on agencies." Keep halving until the £8/month person is unmistakable. Most aspiring operators stop halving three rounds too early.

Path B — start free first, paid later. If after halving you can answer Q1, Q3 and Q4, but Q2 (alternatives) is bleak — i.e. nobody is paying for anything in this niche — start free. Build to 1,000 free subscribers, learn whether they will reply to emails (the leading indicator of paid willingness), and only switch on the £8/month tier when you have evidence.

The pivot rule (issue 12)

Most operators bail on a niche too early — by issue 4 or 5, they decide it isn't working and switch. That is almost always wrong. The signal-to-noise ratio in newsletter feedback is brutal in the first three months; you have to ride it out to know.

The rule: do not seriously consider pivoting before issue 12. Three months of weekly publishing is the minimum data set. After issue 12, run the four questions again. If two or more answers are weaker than they were on day one, pivot. If all four are stronger, you have a real niche — keep going for a year before you change anything.

If the verdict is "commit"

Don't take a victory lap. Open a fresh document. Write the first 200 words of Issue Zero before you go to bed. Send it to two friends in the niche. Watch their reactions. Then schedule Issue One for next Sunday.

The full chapter on niche selection — including the Issue Zero prompt, the sponsor outreach kit, and the cadence that survives the first 12 months — is in Bible 02 (Digital Publishing Empire). £47 launch price (rises to £97 after the first 1,000 buyers).

The first chapter, including the niche framework above in expanded form, is free. No card, instant download.

Whichever path you take: pick the niche by Tuesday. Issue Zero by Sunday. The newsletters that actually compound are the ones that started before they were ready.