The right replacement for Convert depends on who runs your tests. Marketers changing pages without developers need a light visual tool, product teams shipping features need experiments tied to feature flags, and online shops may want testing inside a wider growth or behaviour tool. Below are eight options, in no particular order, grouped by that job.
What are you trying to test? A new headline on a landing page is a very different job from rolling out a checkout change to a slice of app users, and the tools here are usually built for one of those far more than the other. Convert has long been a familiar name for website A/B testing, and its own homepage still leads with split testing, personalisation, advanced goals and support. If you're leaving it, start with the kind of test you run most often, not with the longest feature list.
Two things to check before you move a single experiment
Most testing tools look alike on a feature page. Where they differ is in two places you only notice once tests are live. The first is how an experiment gets built and read. Some tools give marketers a visual editor and a results screen; others expect an engineer to set up flags in code and an analyst to read results from the data warehouse. Neither approach is wrong, but picking the one that doesn't match your team means tests sit in someone's queue for weeks.
The second is page speed. Any tool that changes a page in the browser adds a script, and the way that script loads can affect how quickly visitors see your content and whether they notice a flicker as the variation appears. Ask each vendor how its script loads, then check a test page yourself with your usual speed tools before and after installing it. Run the same real experiment through each shortlisted tool, ideally one you've already run, so you can compare the set-up time and the results side by side.
Three kinds of testing tool, sorted by the job
Visual website testing for marketers
Mida and Webtrends Optimize are pitched at teams that change web pages and want to see which version converts. The first stresses a light footprint and no developer dependency, which matters if your marketing team works on its own. This group is the closest like-for-like swap for most people coming from Convert.
Experiments built around feature flags
GrowthBook, Statsig, Split and Kameleoon lean towards product and engineering teams. Tests sit next to feature flags, so a change can go to a small group of users, be measured and then be rolled out or rolled back. The first of these describes itself as warehouse-native and open source, which suits teams that already have their own analysts and data. If your experiments live mostly in the product rather than on marketing pages, this is the group to look at first.
Testing inside wider growth and behaviour tools
Crazy Egg and Omniconvert put testing alongside other jobs. One pairs tests with heatmaps, recordings and surveys, so you can see where visitors get stuck before you decide what to change. The other is aimed squarely at Shopify stores and pitches AI agents that hunt for growth opportunities. These suit smaller teams that want one tool to suggest what to test as well as run the test.
Side by side: what each testing tool is for
| Tool | Kind of tool | Good for |
|---|---|---|
| Crazy Egg | Behaviour insight plus testing | Seeing what visitors do before you test |
| GrowthBook | Warehouse-native experimentation | Product teams testing on data they already store |
| Kameleoon | Experimentation platform | Building, testing and releasing experiments in one place |
| Mida | Lightweight website testing | Marketers running tests without developer help |
| Omniconvert | Ecommerce growth for Shopify | Online stores focused on profit rather than ad metrics |
| Split | Feature flags with impact data | Engineering teams releasing changes safely |
| Statsig | Product development platform | Product teams measuring what they release |
| Webtrends Optimize | Testing and personalisation | A/B tests and personalisation from one platform |
The eight tools, alphabetically
Listed alphabetically and in no particular order of merit. Each line describes the tool the way its vendor does.
Crazy Egg
Best for: Site owners who want heatmaps, recordings and surveys next to A/B tests, so they can see where visitors struggle before choosing what to change, per the vendor.
GrowthBook
Best for: Data-minded product teams; the vendor calls it warehouse-native and open source, with feature flags and product analytics in the same platform.
Kameleoon
Best for: Anyone wanting to take an experiment from the first test to a full release inside one platform, which the vendor describes as AI-driven.
Mida
Best for: Marketers who want a light tool for website tests, personalised pages and conversion tracking without a developer, according to the vendor; a free plan is named.
Omniconvert
Best for: Shopify merchants; the vendor pitches AI agents that look for growth opportunities, build campaigns and measure profit rather than return on ad spend.
Split
Best for: Engineering teams who ship behind feature flags and want alerts on whether a change helps or hurts, as the vendor describes its feature management.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Statsig
Best for: Product teams wanting to measure the impact of what they ship; the vendor frames it as a product development platform rather than a pure testing tool.
Webtrends Optimize
Best for: Teams that want A/B testing and personalisation together, on a platform the vendor pitches as having no restrictions on how you use it.
Advice for teams leaving Convert
Don't move every experiment at once. Let the tests that are running finish, write down what you've learned, and start the new tool on a fresh test you'd have run anyway. That way you're comparing like with like, and you won't lose results halfway through.
Be honest about who'll run the tool day to day. If it's a marketer with no developer time, a flag-based platform will gather dust, however capable it may be. If it's an engineering team, a visual editor bolted onto the front end can feel limiting. Testing also touches personal data through cookies and tracking, so check what consent you need where you trade; requirements vary by country, so check with your regulator. For a worked example of reading test results, see this comparison of how two offer types convert, and if you're tidying your wider stack, my guide to cloud collaboration tools for teams may help.
A two-week pilot plan
- Pick one page or feature with steady traffic and a clear goal, such as sign-ups or completed checkouts.
- Install your two shortlisted tools on a staging copy and note how long each takes to get a first variation live.
- Check page speed and any visible flicker before and after each script goes in.
- Run the same simple test in each tool for the same period, with the same goal.
- Compare how easy the results were to read and share, and keep the tool your team would use again without being asked.
Do you make A/B testing software? Tool makers can ask to be considered for a featured place on this list.
Want your tool on this list? Featured spots and reviews here are paid partnerships. Here's how to partner with me.
Frequently asked questions
What should I look for in an A/B testing tool?
Start with who builds the tests and who reads the results. Then check how the script affects page speed, how goals are set up and whether the tool fits your consent set-up.
Do I need a developer to run A/B tests?
Not always. Visual testing tools let marketers change pages without code, while experimentation platforms built on feature flags usually need an engineer for set-up.
Can A/B testing slow down my website?
It can, because the tool adds a script that changes the page in the browser. Test a page's speed before and after installing any tool, and ask the vendor how its script loads.
What's the difference between A/B testing and feature flags?
A/B testing compares versions to see which performs better. Feature flags switch a change on or off for chosen users, and many product teams run experiments on top of them.
How long should an A/B test run?
Long enough to cover normal swings in your traffic, such as weekdays and weekends, and to reach the sample size your tool suggests. Stopping early on a lucky day gives misleading results.
Is it hard to move from one testing tool to another?
Moving live tests is awkward, so most teams let running tests finish and start fresh in the new tool. Keep notes on past results so the learning isn't lost.
