Platform shopping has cooled: CMS category search is down a third from its 2025 peak

The market has stopped shopping for platforms and started moving to them. There was a stretch, somewhere around early 2025, when I could not open a vendor email without the word "composable" sitting in the subject line, usually next to an exclamation mark. Then, quietly, it thinned out. I wanted to know if I was imagining it, so we tracked fifteen of the phrases a buyer types while circling the thing they eventually buy: every flavor of headless, composable, DXP, and enterprise CMS wording, combined into one US monthly volume. I was expecting a gentle plateau. What came back was a full arc. Interest crested in March 2025 at about 22,600 searches a month and has slid roughly a third since, down to about 15,200 in July 2026.

The wave, measured

From an August 2024 baseline of about 13,700 monthly searches, category interest climbed 66 percent in seven months to the March 2025 peak: the crest of the composable and headless marketing wave. It has declined almost every month since. By July 2026 the category was 33 percent below peak and within sight of where it started, 11 percent above the 2024 baseline. Year over year, July is down 15 percent. This is not a seasonal dip; it is a round trip.

What is in the basket

The fifteen terms, chosen to cover the destination-platform vocabulary buyers actually use:

• api-first CMS
• cloud CMS
• composable CMS
• decoupled CMS
• digital experience platform
• DXP platform
• enterprise CMS
• enterprise content management system
• enterprise website platform
• headless CMS
• hybrid headless CMS
• SaaS CMS
• web content management system
• web experience platform
• website content management platform

Three readings

1. The hype cycle is normalizing. Composable and headless were the loudest words in the CMS market from 2023 through early 2025. Vendor messaging has since cooled, analyst coverage has moved on, and the search curve looks exactly like what follows a peak of inflated expectations.

2. Buyers moved from research to execution. A team that spent 2024 asking "what is a composable DXP" spends 2026 asking how to get eleven years of content out of the old system without breaking what still works. Category research is a phase; the migration that follows it is a project. The demand did not disappear: it moved downstream, to the move itself.

3. AI assistants are absorbing the definitional queries. "What is a headless CMS" is precisely the kind of question people now ask an AI assistant instead of a search engine. We cannot measure that substitution directly from this dataset, but the timing is consistent with it, and it matches what we see across our other research: discovery is shifting to answer engines. Which is its own reason to care whether machines can read your site.

If you are mid-decision

The practical takeaway is the one we build the product around: the platform category you land on matters less than what survives the trip. Whichever way the headless-versus-traditional decision goes, the duplicate content, orphaned files, embedded scripts, and findability gaps travel with you, or quietly die in transit. If you have a destination shortlisted, grade your current site in sixty seconds, browse the migration guides for your pair, or run a full pre-migration scan before the statement of work is signed.

The data

Method and limitations

So what do I actually think is happening? Not the hype-cycle story, or at least not only that. My honest read is that two of the three readings are doing the real work together: the definitional questions have wandered off to AI assistants, and the people who used to ask "what is a composable DXP" are now heads-down on the far messier job of actually moving. The demand didn't evaporate. It changed shape and went downstream, which a search graph is simply too blunt an instrument to see. I'll flag the honest limit: I can't prove the AI substitution from this dataset, so treat that as the way I lean, not a verdict I'd stake the quarter on. And here's the part I'd bet on regardless of which platform word wins the year. When "composable" fades from your inbox, that's not a signal to relax. It means the buyers who once emailed you the word have stopped talking about the destination and started worrying about the trip, and the duplicate content, orphaned files, and quietly dying redirects don't care what the category is called this season. Grade the site before the statement of work is signed, not after the traffic graph tells you.

Analysis by Replatform Radar. Take it, quote it, build on it. Just point a link back here so people can check my working (I'd want to check mine too).