The short answer before anyone panics
WordPress is not dying. It has been shedding share every quarter since January 2025, and the pace picked up in the summer of 2026, but on 21 October 2026 W3Techs still has it powering 39.6 percent of all websites on the internet and 58.3 percent of the sites that use any recognized CMS at all. The nearest rival, Shopify, sits at 5.3 percent. That means WordPress is still more than seven times larger than whatever is next in line. The headline about a sustained decline is true. The conclusion that the platform is winding down is not.
I run a WordPress agency and I have helped organize the European side of this ecosystem for years. I am not writing this to defend my own stack. I am writing it because clients have started forwarding these charts asking whether they should flee to a no-code builder. The short answer is: look at who is actually gaining in this data, because it is not a competing CMS.
WordPress market share in 2026 according to W3Techs
Start with the numbers, not the feelings. W3Techs is the survey behind most of the industry headlines, and its reports are updated daily. On 21 October 2026, WordPress sits at 39.6 percent of all websites. On 1 January 2025 it was 43.6 percent, its highest reading, and on 1 December 2025 it was still 43.2 percent. Search Engine Journal (Roger Montti, 28 May 2026) counted six consecutive monthly declines from December 2025 to 27 May 2026, after five quarters of smaller drops through 2025.
| Date (W3Techs) | WordPress | No CMS (None) | Shopify | Wix |
|---|---|---|---|---|
| 1 Apr 2025 | 43.5% | 29.1% | 4.8% | 3.6% |
| 1 Dec 2025 | 43.2% | 28.6% | 4.9% | 4.1% |
| 1 Mar 2026 | 42.7% | 28.8% | 5.1% | 4.2% |
| 1 May 2026 | 42.2% | 29.2% | 5.2% | 4.3% |
| 1 Jul 2026 | 41.5% | 29.9% | 5.2% | 4.3% |
| 1 Aug 2026 | 41.2% | 30.4% | 5.3% | 4.3% |
| 1 Sep 2026 | 40.7% | 30.9% | 5.3% | 4.2% |
| 1 Oct 2026 | 40.2% | 31.5% | 5.4% | 4.2% |
| 14 Oct 2026 | 40.0% | 31.6% | 5.4% | 4.2% |
| 21 Oct 2026 | 39.6% | 32.1% | 5.3% | 4.2% |
Source: W3Techs monthly and quarterly usage trends, read on 14 and 21 October 2026.
Update, 21 October 2026: the W3Techs yearly trend has not shown a reading this low since 1 January 2021 (39.5 percent). In the week since 14 October, None rose from 31.6 to 32.1 percent and Shopify slipped from 5.4 to 5.3 percent.
Global market significance of WordPress (2025/2026)
Market data clearly shows why WordPress developer skills are so highly demanded globally.
Three things stand out once you set aside the WordPress row itself. First, the decline is no longer slow and even. WordPress lost 2.0 points in the eight months from December 2025 to August 2026, and then another 1.6 points in just under twelve weeks to 21 October. Second, the classic CMS competitors moved much less: Shopify gained four tenths of a point over that whole stretch and Wix a tenth of a point. Third, the largest change is None, W3Techs’ label for sites where it detects no monitored CMS, which gained three and a half points in the same period.
How W3Techs measures WordPress market share
Before reading any of these numbers as a verdict, it helps to know what the instrument counts. The W3Techs methodology page says the survey covers what it calls the relevant web: sites with meaningful content or functionality, picked using content analysis and the linking structure between sites. The sample is “well over 20 million sites”. Default server pages and near-duplicate sites are excluded, and so are domains that only redirect elsewhere.
Two rules matter more than they look. W3Techs investigates websites, not individual pages, and if a technology shows up on any page, the whole site counts as using it. Subdomains are not separate websites, so every subdomain of wordpress.com or wix.com counts as one site. That is why hosted platforms with millions of free subdomains do not inflate their share in this survey.
W3Techs also publishes two percentages for every CMS. The first is the share of all websites in the sample. The second is the share of the CMS market, meaning only the sites where a known CMS was detected. When the None category grows, the first number falls for every CMS at once, even if no site switched platforms. The second number filters that out and shows how WordPress is doing against other CMSs specifically. Most headlines quote the first.
WordPress market share history since 2015
Zoom out and the panic loses its shape. The W3Techs yearly trend starts on 1 January 2015, when WordPress ran 23.3 percent of all websites. It climbed every year after that, crossed 30 percent during 2018 (29.2 percent on 1 January 2018, 32.7 percent a year later) and reached 43.2 percent on 1 January 2022. From 2022 to January 2026 the line stayed between 43.0 and 43.6 percent. What we are looking at now is not a plunge from a peak, it is the first sustained dip after a long plateau. A platform that spends four years flat and then gives back about four points has stopped growing. It has not started dying. Those are different diagnoses, and the ten-year chart makes the difference obvious in a way a six-month window never will.
The CMS-market column tells a quieter version of the same story. According to the W3Techs market share trend, WordPress held 60.7 percent of the CMS market on 1 January 2015, peaked at 65.2 percent on 1 January 2022, and stands at 58.3 percent on 21 October 2026. Over the same years Shopify went from 0.7 to 7.9 percent of the CMS market and Wix from 0.3 to 6.2 percent. So the named builders did take ground from WordPress among CMS sites, slowly, over a decade.
It also helps to see WordPress next to the rest of the field rather than in isolation. The W3Techs overview for 21 October 2026:
| CMS | Share of all websites | Share of the CMS market |
|---|---|---|
| WordPress | 39.6% | 58.3% |
| Shopify | 5.3% | 7.9% |
| Wix | 4.2% | 6.2% |
| Squarespace | 2.4% | 3.6% |
| Joomla | 1.1% | 1.6% |
| Drupal | 0.6% | 0.9% |
Read the right-hand column slowly. Among sites that chose a CMS, nearly six in ten chose WordPress, and the entire chasing pack of Shopify, Wix, Squarespace, Joomla and Drupal combined adds up to about a third of WordPress’s slice. BuiltWith, which counts detected deployments rather than surveying a sample, also ranks WordPress as the most widely deployed CMS. Two independent measurement approaches, one conclusion about who leads. A category leader does not lose its lead by shedding the cheapest tail of its own userbase.
Where is WordPress market share going
None in the W3Techs vocabulary does not mean nothing. It means sites where no known content management system could be detected. That can include hand-coded HTML, output from frontend frameworks like Next.js, static-site generators, headless delivery, and sites assembled by AI tools that produce finished output with no detectable CMS. The category grew from 28.6 percent on 1 December 2025 to 32.1 percent on 21 October 2026.
This is the whole point the headlines miss. The named competitors did not grow fast enough to mirror the WordPress decline, while the no-detectable-CMS category grew more strongly. That is correlation, not migration evidence: W3Techs cannot tell us which WordPress sites moved where, or why. It does show that the market change is broader than one CMS replacing another. Why WordPress stopped being the automatic first pick for new projects, even among people who still build on it, is a separate question I answered in WordPress stopped being the default.
Break it down by intent and the picture sharpens further. The bottom of the market, the simplest sites that historically landed on a free WordPress theme, are now just as easily produced in a browser by an AI website builder in fifteen minutes. Those sites were never high-margin projects. Their departure lowers the share bar but barely touches the value the WordPress ecosystem actually creates.
Which AI website builders W3Techs detects
If AI builders were swallowing WordPress by name, it would show up in the CMS table, because W3Techs already tracks several of them. On 21 October 2026 the W3Techs overview lists Lovable, which it describes as “an AI-powered platform for building websites”, at 0.2 percent of all websites and 0.4 percent of the CMS market. Vercel’s v0 sits at 0.1 percent. Framer is at 0.2 percent. Astro, a static site generator, is at 0.2 percent of all websites and counts sites such as cloudflare.com among its users.
Add them up and the named AI and static tools are well under one percent of the web. The 3.6 points WordPress lost since December 2025 did not land in any of those rows. They landed in None. That fits two explanations at once: much AI-generated output ships as plain HTML that no detector can attribute to a builder, and many detectors only learn a new tool’s fingerprint after it has spread. W3Techs data cannot tell those apart. What it rules out is the simple story that one AI product is taking WordPress’s place.
Why WordPress market share is declining
Four plausible forces are worth testing rather than presenting as settled causes. AI-built sites can ship static output with no detectable CMS. SaaS builders continue to grow slowly. Headless architecture can keep WordPress underneath while hiding its fingerprint from a frontend scan. Ordinary churn can also remove abandoned WordPress domains from the sample. W3Techs alone cannot rank those explanations.
Notice that three of those four forces are not a story about WordPress being weak. They are a story about the simplest site no longer requiring any CMS, and about a measurement method that structurally undercounts headless. WordPress still wins where a site is meant to live: to be edited, extended, integrated with payments and logistics, and maintained for years. It loses where a site is meant to be built once and never change, which was never its strongest or most profitable territory.
Signs that WordPress is not dying
There is a number the “dying” story needs but these market-share charts do not provide: a comparable time series of active WordPress installations. Share and absolute count are different measures. WordPress can lose share while gaining sites, or lose both at once. W3Techs’ percentage alone cannot decide between those scenarios, and BuiltWith’s deployment totals should not be substituted without comparing the same population over time.
What W3Techs does show is how the installed base behaves. On 21 October 2026, version 7 runs on 66.6 percent of WordPress sites in its sample and version 6 on another 26.5 percent, per the W3Techs WordPress page. Owners of an abandoned platform do not keep updating it. The same page shows WooCommerce on 19.8 percent of WordPress sites, which is the catalogue-and-checkout work no prompt-to-site tool carries, and lists microsoft.com, mit.edu and walmart.com (on subdomains) among the sites using WordPress.
The project is not sitting out the AI wave that supposedly threatens it either. WordPress has a dedicated AI team that, in its own words, is “exploring and coordinating artificial intelligence projects across the WordPress ecosystem”, with regular meetings in the #core-ai Slack channel. Gutenberg and block themes have already moved the design flexibility that once pushed people toward page builders into core.
There are moves in the other direction too. On 19 October 2026 Wil Reynolds of the agency Seer Interactive described on X moving the company’s entire HubSpot blog to a new WordPress site with Claude Code, which worked unattended for six hours. He chose the platform by the quality of its MCP server: “WordPress beat webflow for my use case.” Asked in the replies why, he wrote: “I gave Claude my specs, it recommended WordPress”. Two replies told him to use Astro instead. One case is not a trend, and the migrated blog was still on a staging site when he posted.
None of that erases the decline. It reframes it. A dying platform stops being updated and stops investing in what comes next. WordPress is doing the opposite on both counts while giving up the one segment, the disposable single-page site, that it was never especially good at monetizing in the first place.
Should a business still choose WordPress in 2026
The useful question is not “WordPress or not”, it is “what kind of project is this”. If you need a one-off brochure page with no plans to grow, then yes, a static generator or an AI builder will close the case and there is no shame in that. But if your site is meant to be alive, regularly updated, extended with new sections, wired into a store, a CRM or a booking system, and maintained for several years, then WordPress’s advantage has not changed over these few points of share. The scale of the ecosystem, plugin availability, depth of talent, and absence of vendor lock-in are exactly where they were.
Here is the practical test I give clients. Write down how many times a year someone on your team will want to change content, add a page, or wire in a new integration themselves. If the answer is near zero, a builder is fine. If it is a two-digit number, every hour saved today on a cheaper tool comes back as a cost the moment you want to do something the tool did not anticipate. That is why our WordPress project pricing stays individual: it depends on whether we are building a living site or a tombstone.
Is WordPress development still a good career
For people who make a living on this platform, the decline headline sounds scarier than it is. Demand is not shrinking, it is shifting. The simplest, cheapest sites are leaving, and those are precisely the low-margin jobs that were first to lose the price war with freelancers anyway. What remains, and relatively grows, is the work an AI builder will not touch: architecture decisions, performance budgets, security posture, payment and logistics integrations, migrations, maintenance, and fixing what AI generated and then collapsed under its first real traffic.
Losing the bottom of the market raises the relative value of senior work. This is exactly the piece I wrote about in the article on who a WordPress developer is and what they really do: value shifts toward judgment, and judgment has no “generate” button. The most exposed person is not the WordPress developer, it is the seller of the cheapest templates who competed on price alone. When AI shows up in hiring, real code still decides the hire: WordPress AI hiring and real code, October 2026. For the hosted Agent write path versus self-hosted abilities, see WordPress.com Agent and MCP write.
Should you move a WooCommerce store to a site builder
It shows up most concretely in a real question I got recently. A client running a WooCommerce store had just read one of these charts and wrote that, “since WordPress is dying”, maybe he should move to something more modern. We sat down with the numbers. His store had tens of thousands of SKUs, an integration with a warehouse system, and monthly campaigns that needed on-site changes. Migrating to a closed builder meant rewriting every integration from scratch, losing control of his data, and a monthly fee that scaled with revenue. WordPress “dying” had no operational meaning here at all. What mattered was that no no-CMS tool would carry his catalogue and his process.
The second observation comes from June 2026. A few days after this article first appeared, WordCamp Europe opened in Kraków, the largest European conference for this ecosystem, and the rooms were full of agencies and companies building real businesses on WordPress. An ecosystem that is losing relevance does not draw several thousand people to its own conference. That, too, is data, it just does not fit in a W3Techs table.
Is WordPress dead in 2026
Put the two readings side by side. The bear case is a sustained decline after a long plateau, and one that sped up in the third quarter of 2026. The bull case is a still-dominant CMS-market share of 58.3 percent (21 October 2026), a user base that keeps updating to the current major version, and active investment in editing and AI. Both descriptions fit the measured data. “Maturing” is a more useful working diagnosis than “dying”, but it remains an interpretation that should be tested against active-site counts, project investment, and demand rather than claimed from one percentage alone.
Does falling market share make WordPress a bad choice
The drop from 43.6 to 39.6 percent between 1 January 2025 and 21 October 2026 is real and worth watching, especially after the faster slide since July. But it measures one thing that nobody should confuse with another. It measures what fraction of W3Techs’ surveyed websites can be attributed to WordPress. It does not establish the cause, the absolute active-site trend, or whether WordPress is a good choice for the project in front of you. Those are separate questions.
The most honest summary is this: WordPress is handing the bottom, cheapest segment of the market to no-CMS tools, while keeping the projects that have to be edited, integrated and maintained. For a business with that kind of project, and for a developer who can run one, this data is good news rather than a warning. If you are wondering which side of that line your idea falls on, get in touch and let’s scope it before you decide on the strength of a headline alone.
Last updated: 21 October 2026.
