Methodology
How we turn a complaint into a scoped business idea
The pipeline
Six stages, in order
- 01
Sourcing
We search a fixed set of operator communities for posts matching pain-signal phrases — things like “I wish there was”, “is there a tool for”, “spending hours doing this manually”, “we tried X and it does not work”. Only public posts are read, and only posts from the last few days, so the signal stays current.
- 02
Deduplication
The same pain resurfaces constantly. Every source post is fingerprinted by content and by the problem it describes, so a repeated complaint does not become five near-identical ideas. Recurrence is recorded as evidence instead.
- 03
Hard filters
Before any model runs, candidates are dropped for: being an ad or promotion, being a simple question with no described problem, having no operational detail, being outside business software and services, or relying on a single unverifiable claim. This is what stops the archive filling with noise.
- 04
Stage one — is it real pain?
A first pass asks whether this is a genuine, repeatable problem with a plausible payer, and extracts the problem statement in plain language. Candidates that fail here are discarded without a second pass.
- 05
Stage two — scoping the idea
Survivors get a structured analysis: the vertical, the exact target buyer, the product form, the core value, the current workaround, the competitor gap, itemised pain points, an urgency rating, a build-difficulty rating, a monetization angle and a five-day validation checklist.
- 06
Scoring and human review
Each idea is scored out of 100 across pain acuity, recurrence, willingness to pay, reachability and competition gap. An editor then reviews the batch, verifies sourcing, adjusts scores where the model over-reached, and publishes.
The score
What 0–100 actually means
The score is an editorial judgement, not a forecast. It answers one question: if you had to pick something to build for a small team with limited time, how attractive is this compared to the rest of the archive?
| Band | Reading | Access |
|---|---|---|
| 85–100 | Acute, recurring pain with an identifiable payer and an obvious gap. The rare ideas worth starting this week. | Free in full |
| 70–84 | Solid opportunity with one open question — usually reachability or willingness to pay at the price band. | Pro |
| 55–69 | Real pain, but the buyer is hard to reach or the workaround is good enough for most people. Needs a sharper wedge. | Pro |
| Below 55 | Published only when the pain is unusually interesting. Treat as research input, not a lead. | Pro |
Filters
What gets thrown away
- Promotional posts and affiliate pitches
- "Recommend me a tool" with no described problem
- Venting with no operational detail
- Consumer, political and health-support topics
- Anything depending on one unverifiable claim
- Duplicate problems already covered
- Content-farm and AI-generated source posts
- Posts too old to reflect current pain
Limits
Where this breaks down
Reddit skews technical, English-speaking and younger than the average small-business owner, so the archive under-represents offline trades and non-English markets. Public complaint volume is also a proxy for pain, not for budget — some of the loudest problems belong to people who will never pay to fix them.
Every idea therefore ships with a five-day validation checklist instead of a projection. If you cannot get a deposit in five days, the idea is not validated, whatever its score says.
Corrections are welcome: email editor@painradar.io.
Questions
Methodology FAQ
How does PainRadar decide what counts as a pain point?
A post qualifies when the author describes a specific, recurring operational problem they are actively trying to solve, and names either the workaround they use or the money it costs them. Complaint posts with no operational detail, venting with no attempt at a solution, and pure questions asking for recommendations are filtered out.
What is the opportunity score out of 100?
It is our editorial rating of how attractive the idea is to build. It blends five sub-signals: how acute the pain is, how often it recurs, how likely the affected person is to pay, how reachable that person is, and how large the gap in existing tools appears to be. Higher is better, 85 or above means we consider it strong enough to give away free.
Which communities do you source from?
Operator communities where business owners talk shop: small business, SaaS, ecommerce, agency, trades, healthcare practice management, legal operations, manufacturing, and niche verticals where practitioners discuss their own workflows. We do not source from political, personal or health-support communities.
Do you publish every idea the pipeline produces?
No. The pipeline scores many more candidates than it publishes. Anything that fails the hard filters, scores below our publishing floor, or cannot be given a plausible buyer and price band is discarded. A human editor also reviews each batch before it goes live.
How often are ideas published?
Once a day, on a morning UTC schedule. Each batch is capped so the archive stays readable rather than flooding subscribers with low-signal entries.
Can I see the raw data?
The public metadata of every idea — title, vertical, opportunity type, score, urgency, build difficulty, source community and a short summary — is available openly via /feed.json, /feed.xml and /llms-full.txt. The full analysis for each idea is the paid product.
See also: About · The archive · Machine-readable corpus