Frequently Asked Questions

How many visits do I need to start seeing results?

In our experience, you should start seeing results with 1500 to 2000 unique visitors. It depends on the type of conversion event you choose: it’s usually faster to get more conversions in a newsletter form than selling a 100 USD digital product. So, the amount of needed visitors might vary, but you should feel comfortable thinking about 1500 to 2000.

When do I start paying?

We have a free plan that enables you to run your first experiment for free. After the first experiment reaches statistical significance, you will need to upgrade to keep using the product.

What is “statistical significance”?

An A/B test reaches “statistical significance” when you can say confidently that one of the variations has a higher conversion rate than the others. You can read more about this [here](https://en.wikipedia.org/wiki/Statistical_significance).

What is a Verified Conversion Rate Increase?

Once you do your first experiment our algorithm will find a variable that converts more. When we get a statistical significance of 90%, you will receive an email with the results and a congrats for achieving your first Verified Conversion Rate Increase. If you subscribe, the tool will start running a new experiment automatically, with the winning variation as the original.

Can it break the layout of my page?

No. The tool is designed to respect the size and CSS of the elements of your page. You can see a preview of each variation before going live with the A/B test, both on the mobile and desktop versions. Once you integrate our javascript snippet, you can also check out the live versions.

How do you track conversions?

You can add the URL of your “thank you” page, add a JavaScript event anywhere you want, or let our AI recognize your conversion event automatically.

Can I lose conversions while running the experiment?

We use a multi-armed bandit approach to A/B testing that will show more often the variation of your landing page that has a higher conversion rate while running the test. You can read more about this in our article Introducing ABtesting.ai.

Can I cancel anytime?

Yes.

Does it work with any language/CMS?

Yes, it doesn’t depend on how your website is built. We detect automatically everything in the HTML, so the language or CMS will never be a problem. We have an WP plugin for a faster installation on WordPress.

Do I need to code?

No! You only need to add a small JavaScript snippet on the "head" tag of your HTML and the tool will start making changes on your web page.

What does “powered by AI” means?

We use AI in multiple ways during your A/B testing implementation. We detect automatically what elements of your page you could test, and make suggestions based on your site and the data points we have already collected. Nothing will happen on your landing page if you don’t want it that way, and you always have a preview available.

Do you always show the same content to the same users?

Once a variation has been assigned to a user, we will set up a cookie for that user so they will see the same variation if they enter again to your web page.

How many generated combinations do you usually produce for a site? And are they equally randomly shown (based on a percentage) to visitors? Or are better-performing combinations (based on past results) shown more often?

We start showing each variation the same amount of times randomly to different visitors. Once the algorithm starts to understand which variation converts more, it starts showing it more. We make sure you don’t lose conversions on a bad variation, and keep going until we get a 90% confidence on the result.

How involved is the JS code installation?

Is it a single line of code you add to your site’s head tag. A short and light snippet that takes care of everything.

Can 1 account test more than one site? And what about pages?

You can test as many domains as you want from one account, and you are also able to perform multiple tests on different pages from the same domain at the same time.

Do I have a detailed overview of how each variation is performing?

Yes! We have a dashboard that shows you each variation of the landing page, with individual statistics about visitors, conversions, and conversion rates. You can check the desktop and mobile versions of each variation in real-time.

If you do have a dashboard, can you invite team members to view it?

Our Team Feature allows you to invite team members and clients, as you can adjust the type of permission and the experiments a user is allowed to see. This feature is available for Business and Enterprise plans only.

For how long are you showing the same version to the same user?

We don’t want your user to feel confused about changing titles or copy, so the same variation is shown for each user every time he/she visits your site.

Can I stop using this service after the optimal layout is found?

Yes, and as long as you have both the test running and the snippet correctly placed, the winner variation will be the one shown to all of your visitors. If you are planning to delete any of those, make sure to make the changes manually on your site before!

Will I get access to the AI's data? I want to see which combinations work best.

In the Overview you'll be able to see the performance of the test and also each of the running variations separately.

Who else uses ABtesting.ai?

Toyota, Suzuki, Claro, Movistar, Doctors Without Borders, El País, and other cool brands trust us to increase their conversions.