Download websites that use Convert

About Convert

Not every script on a page exists to make the page work. Some are there to find out whether the page works. Convert belongs to the second group. The tool holds back a share of arriving traffic, shows it a different headline, price or layout, and reports which version produced more of whatever the operator counts as success.

A domain running an experimentation tool is describing the organization behind it, not only the site. Testing needs enough traffic for a result to mean anything, somebody whose job includes designing the test, and a management culture willing to change the page because a measurement said so. Webatla files Convert under A/B Testing, and in practice that category behaves less like a technology preference and more like a marker of commercial maturity. Sites that carry it are usually past the stage of arguing about a logo and into the stage of arguing about revenue per visitor.

Convert Insights Inc. has sold the platform since 2009 and keeps it Proprietary, positioning it on how carefully it handles visitor data rather than on breadth of features. The positioning shapes the customer base. Teams that pick it are often ones with a compliance constraint or a stated reluctance to load third-party cookies, which means the install base is not a random sample of companies that run experiments.

Detection has a specific shape here. The product is delivered as a JavaScript snippet that has to load early, before the visitor sees the original version of the page, so when it is present it is usually visible in the markup rather than hidden. The complication is targeting. Experiments run on chosen templates, so a shop may carry the snippet on product pages and not on its home page, and a team that injects it through a tag manager can leave a thinner trace. As of October 2026 we record Convert on 1,338 sites, and the true number of companies with an account is higher than the number of domains where a crawler happens to land on a tested page.

The value of having every one of those domains in a file is that experimentation is otherwise invisible from outside. No registry lists which companies run a testing programme. A competing vendor can size its addressable market and find accounts to approach, an agency can prospect for teams that already believe in testing rather than teams it would have to convert, and an analyst can ask which sectors, platforms and traffic ranges correlate with running experiments at all. Those are aggregate questions, and a lookup answers them one domain at a time, which is to say not at all.