How StatWP estimates installs and ranks

StatWP is an independent analytics product for WordPress.org plugins and themes. We are not affiliated with the WordPress Foundation. This page explains where numbers come from so you (and AI tools) can cite them accurately.

1. Data Sources

• Public WordPress.org plugin and theme info APIs (installs, downloads, ratings, versions, tags, and related fields)
• Public WordPress.org stats endpoints (daily downloads, version share)
• StatWP’s own daily catalog snapshots for growth charts and popular-browse rank history

2. Active Installs (Reported vs Estimated)

WordPress.org publishes rounded install buckets (for example 10,000+), not an exact active-install census. StatWP always shows that official bucket. When we have enough history, we also show an estimated count inside the band and a High / Medium / Low confidence label. The likely range sits in the info tooltip on the estimate, not as a second number on the tile.

A listing can gain users every day and keep the same reported number for weeks. Charting that official step as a smooth line invents growth. Our estimated-installs chart is the inside-band estimate, not a redraw of the bucket staircase.

What We Use (and Why)

• Bucket crossings. When a listing moves from one WP.org bucket to the next, the true count crossed that boundary on that day. That is a measurement, not a model. We project from the last crossing using the observed rate between crossings. Best for larger bands, where rank among ties is noisy.
• Rank in bucket. Many listings share the same reported bucket. Popular-browse rank among those ties places a listing nearer the bottom or top of the band. Rank can rise and fall daily even with no crossing, which crossings alone cannot do. Stronger on smaller bands, where the pack is tighter.
• Download ladder (weak prior). Relative all-time downloads among listings tied at the same bucket. Used when rank is missing, and as a light third vote when it exists. Never treated as High confidence on its own.

We combine those signals instead of picking a single winner. Weights shift with band size: rank matters more below about 10,000 installs; crossings matter more toward 1 million and above. The point estimate is a weighted average, then rounded to a step that matches the band (ones, tens, hundreds, thousands, or ten-thousands). The likely range is the overlap of the signal intervals when they agree, or the union when they disagree (and confidence drops to Low).

How Much to Trust

Most confidence: the day of a WordPress.org bucket flip, especially on listings under about 10,000 installs (and even more under 2,000). That day the true count hit the boundary. It is a measurement. Rank among many listings in a tight band then keeps the estimate useful until the next flip. Confidence can be High there when signals agree.

• Highest. Recent flip, small band (under ~10,000, especially under 2,000), rank and crossings agree.
• Still useful. 10,000 to 100,000 when signals agree. Read it as which side of the band, not a headcount.
• Use the official bucket first. 100,000 to 1 million. The estimate only says closer to the floor or the ceiling.
• Least confidence. 1 million and above (including 10M+). Those steps can be a million to ten million wide. Flips are rare. Rank among a few giants is weak. We never label High. Trust 1M+ / 10M+. Do not compare two large listings by a small gap in the estimate. Use rank and download trend instead.

What We Do Not Use

• We do not invert active-version percentages to recover N. Version share is shown as a chart only. Treating those percentages as a population size is unstable and we do not treat it as an install count.
• The tile shows a grouped point estimate (for example ≈ 57,600), not a compact 57.6K. That figure is still rounded to a step that matches the band, so it is not a census. The likely range is in the tooltip.
• We never label confidence High on bands at 1 million installs and above. Those WordPress.org steps are too wide.

Estimates are informational. They are not official WordPress.org statistics and can revise as new snapshots arrive.

3. Rank

“Rank” on StatWP typically means position in WordPress.org’s popular browse ordering among plugins or themes we track, refreshed with our catalog sync. Day-to-day rank charts use our stored history, not a live scrape on every page view.

4. Downloads, Ratings, and Versions

Daily downloads and active-version distribution come from WordPress.org stats APIs. Ratings use WP.org’s rating fields (often shown as stars out of five derived from a 0–100 score).

The download counter includes updates. Shipping a new version makes most of the existing install base download again, so the series moves with release cadence as much as with new users. All-time download totals are not a people count. Chart spikes after a release are expected; they are not the same as new installs.

5. Paid Tools vs Free Stats

Core browse and detail stats stay available without a paid plan. Features such as SEO keyword ranks, higher niche-research limits, and some exports require a paid plan: see Pricing.

6. Review Gap Finder (Critical Reviews)

The Review Gap Finder reads critical WordPress.org reviews: ratings of 1 to 3 stars. Praise (4 to 5 stars) is not scored as a complaint.

A niche (tag) analysis looks at the largest plugins in that tag by active installs, up to the depth your plan allows (for example 5 or 20). The line Only 3 of 20 plugins had critical reviews means: of those top plugins, only three had enough 1 to 3 star reviews with usable complaint labels to include in the market read. The others may have had no critical reviews, only praise, or nothing we could analyse yet.

Problems many plugins share are complaint themes that show up across enough of the included plugins to look like a market opening. Only on one pluginmeans the same kind of complaint, but just on one or a few listings: a rival's weak spot, not a niche-wide gap.

7. Contact

Questions about methodology: support@statwp.com.