Download websites that use Google Optimize

About Google Optimize

The interesting thing about Google Optimize is that Google switched it off and the code is still on the web. The service was a free A/B testing and personalization layer wired into Google Analytics, and when the vendor retired the product the experiments stopped running. The snippets did not disappear from the pages that carried them.

A detection in a current export therefore means something different from a detection of a live product. Finding Google Optimize on a site today says that somebody once ran experiments there, that a marketing team or an agency installed the tag, and that nobody has been through the tag configuration since the shutdown. The presence is evidence of history and of maintenance habits, not of an active testing program, and any reading of the data that treats it as current usage will be wrong.

The narrowness of that reading is what makes the list actionable. Experimentation vendors are the obvious audience, because every domain here belongs to an organization that adopted testing, budgeted for it, built the internal habit and then lost its tool, which is a better qualified prospect than a company that has never run a test at all. Agencies auditing client sites want the same list for the opposite reason, since a dead tag is a measurable piece of tag debt and a reason to open a conversation about what else is still loading and slowing pages down for no return.

As of September 2026 we detect Google Optimize on 15,751 sites, from a product Google first launched in 2017 on a Freemium model, and the count is really a measure of how slowly tags get removed. Watching the figure fall across successive exports is a proxy for how quickly the web cleans up after a discontinued service, and the rate of decay is a genuine finding about site maintenance in general rather than a fact about this tool alone.

Detection reads the JavaScript snippet and its container reference, which is dependable when the code sits in the page and less so when a tag manager injects it at runtime, so the surviving remnant is larger than what we publish. We hold no version data here either, since the product was served entirely by the vendor and there was never a per site copy to number. What the file answers, and a per domain check cannot, is the population question. Which industries and country markets still carry the tag, what analytics and testing tools sit beside it now, and which sites finally cleaned it up between one export and the next.