How we verify

01

▸ What we track

We read the public record behind two catalogs: plugin and app directories, developer forums and reviews, and the release notes each platform ships itself. Nothing here is scraped once and left to rot.

That’s 67,000+ WordPress plugins, 15,600+ Shopify apps, and millions of individual data points behind them: exact download counts (opens in new tab), ratings, release dates, and the threads asking for what still doesn’t exist. Every number on this site is read live from the record, never hand-typed.

02

▸ The recurrence gate

An unmet need doesn’t get a page here because one person complained once. It gets a page because at least 5 distinct sources, different people in different threads asking independently, raised the same gap and nobody built it. That number is the floor, not an average: every published need on this site cleared it before it shipped.

“Distinct” is counted precisely: once per person, per thread. A single post with a hundred upvotes counts as one source, not a hundred. An upvote isn’t a voice; a separate person asking is.

Clearing the gate makes a need eligible for a page. It doesn’t publish one: a person reads every brief before it publishes.

As of the last pass: 6 cases are cleared and build-ready. The 50 others are logged and on the watchlist. See the board →

03

▸ Where each number came from

Every chart and every table on this site tags its own data by where it came from. Four labels cover everything we publish:

Exact count

Pulled directly from WordPress.org's own stats API: an exact daily download count, not an estimate.

Catalog snapshot

The value on file as of our most recent read of the source.

Archive reconstruction

Reconstructed from a page as the Internet Archive saw it, at quarterly granularity. Used for Shopify's adoption history, since Shopify doesn't publish a public download API the way WordPress.org does.

Live read

Read from the platform itself, not backfilled from an archive.

A chart built from more than one of these carries a caption saying so: a reconstructed stretch is never presented as an exact one.

04

▸ The publish gate

A plugin or app doesn’t get a public page just because it exists in a catalog. It has to clear a bar first:

• WordPress: not closed on WordPress.org · at least 25 ratings · at least a 90-day span of download history on file · at least one tracked signal on record (a download row, a rank snapshot, a Reddit mention, or a matched GitHub repo).
• Shopify: at least 5 Internet Archive snapshots showing its adoption history over time.

A page that clears the bar keeps its URL even if a later refresh dips it back below. We don’t pull a plugin’s page out from under an inbound link over one bad week. A page is retired only when the plugin closes on WordPress.org or the app is delisted from the Shopify App Store.

As of the last refresh: 3,212 WordPress plugins and 2,007 Shopify apps meet the bar. Browse WordPress plugins → · Browse Shopify apps →

05

▸ What we never do

The rules below aren’t aspirational copy. They’re enforced in the same code that renders every page.

• No fabricated zeros. A missing number reads "not tracked yet": never a zero, never a blank cell, never silently dropped mid-sentence.
• No composite scores. No 0–100 "heat" index, no star-rating composite, no color-coded rank or direction. A page states the individual, sourced facts and lets you judge.
• No risk words. A plugin eighteen months without an update is described in months and versions, never called "dead" or "abandoned."
• No republished Reddit content. Evidence is a subreddit, a date, a link, and a count. Never a post body, never a username, never more than a short fragment.
• No blur-and-paywall. A module is either real, complete content, or it isn't on the page at all. Nothing here is greyed out and stamped with a lock over content that doesn't exist.
• No manufactured urgency. Scarcity is stated plainly when it's real (a founding-rate window, a seat count), and never invented: no countdown timers, no fake viewer counts.