Methodology
How PainMiner validates software opportunities
PainMiner turns public complaints into validated, high-intent software opportunities through a six-stage pipeline: source, prefilter, deduplicate, classify with a two-pass LLM, score, and editorially approve. Every opportunity is scored for commercial intent, confidence, demand trend, and willingness to pay, and reviewed by a human before it is published. The pipeline runs weekly.
The pipeline
1. Source from public complaints
We monitor public, attributable sources where people describe real problems in their own words: Reddit, Hacker News, app store reviews, and public product forums. No private data and no surveys. We read what people are already saying.
2. Prefilter the noise
A lightweight keyword and heuristic pass discards anything that clearly is not a monetizable complaint (praise, news, memes, support chatter) before any expensive analysis runs. This keeps the signal high and the cost low.
3. Deduplicate
Near-identical complaints are collapsed so a single loud thread does not masquerade as broad demand. Frequency across distinct authors and sources is what makes a signal trustworthy.
4. Two-pass LLM classification
Survivors go through a two-pass language-model pipeline: the first pass confirms it is a genuine, software-shaped problem; the second extracts the opportunity, the target persona, existing workarounds, and the unmet gap.
5. Score the opportunity
Each opportunity is scored for commercial intent, confidence, demand trend, and willingness to pay (defined below). Scores are comparable across categories so you can triage quickly.
6. Editorial approval gate
Nothing is published automatically. An opportunity must clear score thresholds and pass an editorial quality review (genuine, specific, monetizable, not thin or duplicate) before Jerry Plymouth approves it for the database and the weekly Top 10.
What each score means
Commercial intent
0-100How strongly the complaint implies someone would pay for a fix. High scores mean explicit budget signals, business impact, or active searching for a paid alternative.
Confidence
0-100How sure the pipeline is that this is a real, well-formed opportunity rather than a vague gripe. Higher confidence opportunities clear the approval bar more easily.
Trend velocity
% changeWhether the volume of this complaint is rising or falling over time. Rising trends flag emerging gaps before they are obvious.
Willingness to pay
signalA flag set when the source language indicates the person already pays for an inadequate tool or explicitly wants to pay for a better one.
Frequently asked questions
- Where does PainMiner's data come from?
- From public, attributable internet sources where people describe problems in their own words: Reddit, Hacker News, app store reviews, and public product forums. PainMiner does not use private data or surveys.
- How are opportunities scored?
- Every opportunity gets a commercial-intent score (0-100), a confidence score (0-100), a trend-velocity signal, and a willingness-to-pay flag. Scores are comparable across categories so founders can triage quickly.
- What does the commercial intent score mean?
- Commercial intent (0-100) estimates how strongly a complaint implies someone would pay for a solution, based on explicit budget signals, business impact, and evidence of actively seeking a paid alternative.
- How often is the data updated?
- The pipeline runs weekly. Each cycle ingests new complaints, classifies and scores them, refreshes the Top 10, and publishes a curator briefing.
- Is every opportunity reviewed by a human?
- Yes. Nothing is published automatically. Each opportunity must clear score thresholds and an editorial quality review before Jerry Plymouth approves it.
- How is PainMiner different from a keyword research tool?
- Keyword tools tell you what people search for. PainMiner reads what people complain about, then scores whether that frustration is monetizable, validated, and trending, and packages it as an opportunity with a validation playbook.