Somewhere in late 2025 the WordPress product world agreed on a story: sales are falling, AI is eating us, the sky is on its way down. A December survey of 33 plugin companies went around showing 80% with sales flat or lower than the year before. The number is real. The story attached to it deserves a second look, because a company selling less and a market breaking are two different claims, and only one of them is supported.
- The survey says
- 80% flat or lower
- The part left out
- hard-to-replace plugins grew
- The real question
- were those sales ever earned?
The plugin gold rush, five years ago
Rewind to how the plugin library used to work. The first author to ship a plugin that disabled comments collected money for years. Not for being good. For being early. The pattern is still standing in the directory today: a plugin whose entire function is duplicating a page sits at over three million active installations, with a Pro version that charges you for choosing which post types it applies to.
Was that innovation?
No. Nothing about a checkbox behind a paywall improved a process, opened a new angle on a problem, or moved the ecosystem anywhere. It was arbitrage on a distribution system that handed out visibility in a deeply abnormal way and then never corrected itself.
The system that made it possible
The official directory ranks search results with active installations as a heavy input. The count enters the score logarithmically, and in the analysis of one developer who took the algorithm apart in 2026, install count outweighs quality signals roughly nine to one at the million-install threshold. An independent developer “cannot manufacture a million installs,” as that analysis puts it. An incumbent does not have to. Whoever got there first, stays.
And that is the root of it: years without meaningful moderation, without built-in licensing or commerce tooling, without product thinking from the platform, and with a cohort of authors perfectly comfortable with the arrangement. Nobody was building forward. Not WordPress as a company, and not the part of the community that was cashing the checks. I keep a whole wishlist of what a marketplace should have built instead, and most of it is fifteen years overdue.
What it produced
Single-purpose tools, often badly written, alive for a decade on the strength of a search ranking. Millions of installations, minimal maintenance, a paywall on the second checkbox. As a business, brilliant. As software, filler. And filler has one specific weakness: it is only worth money while making it yourself is harder than paying.
The moment it stopped working
That moment arrived with the free tier of every coding assistant. The founders in the survey said it themselves: customers are building their own solutions, relying on AI for support, and skipping paid tools altogether. One practitioner put it more bluntly: the job of a simple utility plugin can now be done by pasting a prompt into any coding assistant. The plugin review team even accepts AI-generated code into the directory now.
The moat was never the code. The moat was that nobody could be bothered, and now everybody can.
Where the money actually went
Here is the part of that survey that did not make it into the panic. Two thirds of the plugins described as hard to replace were growing. The steepest declines, the drops of 30% and more, clustered among plugins described as easy to replace. Same year, same market, same AI.
| Company | Self-reported result | Year |
|---|---|---|
| Barn2 Plugins | new sales down 17.8% | 2025 |
| WP Fusion | revenue down 3%, new customers down 25% | 2024 |
| WP Umbrella | revenue up 67%, $1.3M ARR | 2025 |
| Studio Wombat | revenue up 8% on 28% less traffic | 2025 |
Barn2, the company whose CEO ran that survey, reported new sales down 17.8% in the same year its newer Setary app grew its active subscribers 48%. WP Fusion, one of the most respected products in the space, logged its first down year in 2024 and pointed the finger at Google search changes rather than AI; they cut its organic traffic by a third. Meanwhile WP Umbrella grew 67% to $1.3 million ARR, and Studio Wombat grew revenue 8% in a year its traffic fell 28%, noting that AI assistants now send it more visitors than social media does.
A dying market looks different. This looks like money changing addresses: away from tollbooths built on a badly designed road, toward products that survive being compared to an alternative. WordPress is not losing this revenue. The people who mistook a distribution anomaly for a business are.
The moat was never the code. The moat was that nobody could be bothered.
So did sales drop?
At four out of five companies, yes, the chart points down, and some of the loudest voices in the mourning chorus are the same ones that mourned exclusivity. But ask a different question and the year reads differently. For the first time since the directory existed, those sales had to be earned. Some products noticed no difference. The rest are calling it a crisis.
Last updated
