Grow
A/B testing
A/B testing shows different versions of your landing page to different visitors, so you can find out which tagline, description or theme sends more people to the store. It is available on the Studio plan.
What you can test
Each variant can override these parts of the page:
- Tagline
- Description
- Theme: Clean, Midnight or Vivid
- Accent colour
Everything else, including the name, icon, screenshots, features and store links, stays the same in every variant. A field you leave empty in a variant uses the app's regular value.
Who takes part
Only visitors who see the landing page take part:
- people who open the smart link and are not redirected to a store, such as desktop visitors or platforms whose rule is set to Landing page,
- visitors of
mapps.one/habitly/page, - visitors of your custom domain.
Phone visitors who are redirected straight to a store never see the page, so they are not part of the experiment.
mapps.one/habitly/page in that campaign instead of the smart link. Every visitor then sees the page and gets a variant.Create an experiment
- Open the app's Experiments tab and create an experiment. Give it a name you will recognise later, such as "Tagline test, September".
- Set up variant A and variant B. Give each one a name, a weight and the overrides you want to test.
- Choose Start.
One experiment runs on an app at a time, and it needs at least two variants. Weights are relative: 50 and 50 splits traffic evenly, 80 and 20 sends four times as many new visitors to A. A variant with a weight of 0 gets no new visitors.
How visitors are assigned
On their first visit, each visitor gets a variant at random, according to the weights. A first-party cookie named mapps_variant_{experiment ID} stores the variant number for 30 days, so returning visitors see the same version. The store badges on the page carry the variant, so store clicks are credited to the version the visitor saw.
While an experiment runs, the page is not served from a shared cache, so each visitor gets their own variant. Visitors who block cookies may see a different variant on a later visit.
Read the results
| Column | Meaning |
|---|---|
| Views | Landing page views, including smart link visits that showed the page |
| Store clicks | Store badge taps. These are the conversions. |
| Visitors | Unique visitors, counted per day |
| Conversion rate | Store clicks ÷ views |
Results are counted from the moment the experiment starts.
Finish an experiment
Choose Stop experiment when you have a clear result. Assignment stops and every visitor sees the same page again. You can use version B for everyone, which saves its changes to your page; keep your current page; mark B as the winner and apply the changes yourself later; or stop without a winner.
If the workspace moves off the Studio plan, running experiments stop affecting the page and visitors see the regular version.
Privacy
The variant cookie is only set while an experiment is running on that page, contains nothing but a variant number and is not used on other sites. Analytics stay cookieless. The cookie is described in our cookie policy. In some countries, rules may treat a testing cookie as non-essential, so check what applies to your audience before you start.