The Gap List
— Methodology —

How we find gaps.

Every issue in The Gap List follows the same process: research, score, write, verify. No shortcuts. Here is the full method, open for inspection.

The Gap Test

Every gap must pass all six criteria before it ships. Miss one, it goes back to research or gets killed.

1
Pain documented
Linkable evidence — Reddit threads, G2 reviews, forum posts, tweets. Not vibes. Not assumptions. Dated, screenshotted, linked.
2
Demand has a number
Search volume, subreddit subscriber count, traffic estimates, GitHub stars. At least one quantified signal that people are looking for this.
3
Buildable solo in 2–6 weeks
A single developer can ship an MVP. If it requires a team, a data moat, or regulatory approval, it fails this test.
4
The wedge is real
Not 'cheaper version of X.' A real structural advantage — a different input, a different audience, a different output format.
5
Path to first dollar
A concrete revenue model. Not 'we'll figure it out after launch.' The monetization plan is in the issue, not deferred.
6
Gap is unfilled
We audit the top 50 search results, check Product Hunt, scan Indie Hackers launches. If someone already built it well, it's not a gap.

Evidence standards

Every claim in every issue must have a receipt — a numbered, dated, linkable source. We do not publish unsourced assertions. If a source goes dead, we mark it and note the date we last verified it.

Data sources include public job boards (Upwork, Toptal, public LinkedIn), keyword tools (ahrefs, Google Trends), community platforms (Reddit, Indie Hackers, Hacker News), and review aggregators (G2, Capterra). We do not fabricate data points. Every number in the starter dataset is real, sourced, and dated.

Each receipt in an issue carries a freshness badge: green means fresh (<120 days), amber means aging (120–180 days), red means stale (>180 days).

Production model

Issues are AI-drafted and human-edited. Claude generates the initial research synthesis, analysis, and build kit. A human editor verifies every source, checks every link, stress-tests the build tier estimate, and rewrites for voice. No issue ships without human review.

We disclose this openly because we believe the value is in the research rigor and the build kit — not in whether a human or an AI typed the first draft.

The build kit

Every issue ships with a complete build kit: PRD, design spec, tech spec, 10 ordered Claude Code prompts, starter dataset, and brand-aligned assets. Five rules govern every kit:

  1. No fabricated data. Starter data is real, sourced, dated.
  2. No surprise dependencies. Every dep is justified in TECH.md.
  3. Ships in a weekend. For weekend-tier issues, the kit completes in ~3 days.
  4. Composable with itself. Files cross-reference consistently.
  5. Brand-aligned visuals. Charts use the palette, mockups use the typography.

Refresh cadence

Market data goes stale. Our commitment: every issue published in the current quarter has data from within that quarter. Older issues carry a "data as of" date on every quantified claim. When a gap gets filled by a shipped product, we update the issue with a note and a link.

Corrections

If a source is wrong, a claim is unsupported, or a receipt link dies, we correct it in-place with a visible "[Corrected]" tag and the date of correction. The original text is preserved in the git history.

Contact

Found a broken source? Disagree with a gap score? Built something from a kit and want to share? Reach us at hello@thegaplist.com.