The classic mistake when picking a niche for a tool site is to look at search volume alone. Lots of searches does not mean anyone will pay. We judge every niche with three rulers and only trust them together — any single one will mislead you.
Three rulers for telling whether a niche pays
Search pond: volume × CPC
We sum volume × CPC over every keyword in a niche that passed review. CPC is what advertisers pay per click, and advertisers only bid high where buyers are — so volume × price says more about how much a pond is worth than volume alone. Data: Similarweb estimates, global, last 28 days.
TrustMRR: is anyone actually earning?
Founders publish their MRR on TrustMRR. We group those sites by niche and count how many report revenue and which band the median falls in (<$1k, $1k–10k, $10k+). Self-reported, not audited — but it shows someone has turned the niche into a business.
TopCheckouts: is there traffic on checkout pages?
TopCheckouts estimates checkout-page visits per site. If people reach the checkout, users are getting to the point of paying. We only show a tier — none, weak (<10k/month), medium (10k–50k), strong (50k+) — and link back to the source page.
Putting it together: a real pond has a big search pond and at least one of the two payment rulers lit. Today’s data for a few big ponds:
| Niche | Monthly searches | CPC | Checkout traffic | Sites reporting revenue | Median MRR |
|---|---|---|---|---|---|
| AI video generation | 1,300,700 | $2.35 | Strong | 10 | $1k–10k |
| AI voice, TTS & voice changer | 518,200 | $2.44 | Strong | 7 | <$1k |
| LLMO / GEO / AEO | 73,000 | $4.05 | Weak | 9 | $1k–10k |
The free-pond trap: transcription, translation, PDF
Some of the biggest-volume niches are free ponds: transcription, translation, PDF and logo tools get millions of searches, yet founders report little revenue. Free substitutes are everywhere — built-in captions, browser translation, free tiers from big vendors — so users do not expect to pay.
| Niche | Monthly searches | CPC | Checkout traffic | Sites reporting revenue | Median MRR |
|---|---|---|---|---|---|
| Transcription & subtitles | 912,300 | $2.55 | Strong | 9 | <$1k |
| Translation | 1,124,200 | $2.63 | Medium | 6 | <$1k |
| PDF tools Free pond | 2,264,100 | $2.14 | Medium | 4 | <$1k |
We flag these as Free pond when search value ranks in the top 10 while reported revenue is below the median of all niches. Sorting by volume alone walks you straight into them.
How the niche score works
A 0–100 percentile score: every day we rank all niches on search value (weight 0.4), total revenue of paying sites (0.3), checkout traffic (0.2) and new kept keywords in the last 7 days (0.1), then combine and scale to 100. It compares niches with each other; it is not a revenue figure.
How each keyword is reviewed
- Rules remove what cannot be built: brand downloads and logins, brand pricing pages, account resellers, NSFW, MCP tool names and similar.
- Every remaining keyword gets an AI review (Claude / Grok): verdict (keep / drop / copyright risk), page type, whether the SERP is a red ocean, audience, and a one-line reason.
- Synonym variants are grouped into one keyword group — one need, one page.
- Niches are rule-based groupings and contain noise. Trust the order of magnitude, not the exact rank.