App Store Optimization

App Store Optimization (ASO): A/B Testing Screenshots to Improve Installs

App Store Optimization (ASO) requires ongoing management. To maximize your conversion rates for the lowest possible user acquisition cost, you must understand optimizing the sequence of screenshots on your app store pages.

Previously, we explored designing screenshots, picking the best ASO keywords for your app, designing your app’s icon, and optimizing your icon through A/B testing. In this post, we’ll expand on A/B testing and finding the best screenshots to optimize conversion rates (CVR). Be mindful of the user experience on each device. Web scroll behavior theory applies to mobile as well. Many users will not bother swiping or scrolling too far to the right, so it’s critical that your high-converting screenshots are visible right away.

Here’s how you find those screenshots that will increase downloads under any budget.

How to A/B Test Screenshots for Every Mobile OS

The quickest way to find your A/B test winner is through sequential A/B testing. However, how you execute this method differs by app store:

  • Google Play: Android developers have it easy thanks to Google Experiments, which lets you set up tests and collect results within Google’s dashboards.
  • iOS App Store / Others: Everyone else needs to go the paid route. Similar to A/B testing icons, you can use paid Facebook campaigns.

02 - Google Experiments

Experiments allow for easy A/B testing on Google Play.

Tools (And Lessons) From a Mobile App Expert

We reached out to Ilya Shereshevsky, former Director of Performance Marketing at DeNA, who has extensive experience A/B testing apps on multiple platforms. Throughout his career, he regularly experimented with different methods to find the best fit for each unique product.

03 - Godus

Godus’ screenshots were A/B tested while Shereshevsky was at DeNA.

“We typically used Facebook tests prior to an app’s soft launch, which required us to use the campaign objective ‘Click to Website’ where the clickthrough took users to the company’s website. The clickthrough rate was the evaluation criteria that we would base our decisions on.”

However, Shereshevsky found that Facebook didn’t truly emulate an app store. “What we were missing was mimicking the true consumer experience that occurs when a user sees your screenshots on the iOS App Store or Google Play.”

For a more authentic app store experience, consider services like StoreMaven. “When working for a client, I used StoreMaven to A/B test screenshots. The account management, sharing of best practices, and reporting provided by the StoreMaven team combine for a robust testing capability.”

One important caveat to this approach, however, is budget: “Unfortunately there was a big drop-off from the ad click to the app store page which made the cost to acquire each sample user much higher than expected: greater than $10 on Google Play.”

04 - Ilya Shereshevsky

Ilya Shereshevsky, former Director of Performance Marketing at DeNA.

On Google Play, Shereshevsky adopted Google Experiments as soon as it rolled out. “Now, my preference is to A/B test utilizing Google Play’s developer console. Google Experiments reduces the reliance on click-to-website ads to drive traffic and most closely mimics the real user experience. It’s simple to set up and can run in parallel with your normal user acquisition efforts.”

In closing, Shereshevsky offered an additional iOS alternative, “For iOS-first clients, I plan to also test Splitmetrics, as I’ve heard that their platform for A/B testing is robust as well.”

Combining A/B Testing With Design

Always lead with the screenshots that appeal to your priority target market, accounting for common behavior across devices where your user views your store page. When they come across your page, they could be on their desktop, laptop, phone or tablet. To maximize your conversion rates, you must not only have the best set of screenshots, but also know the correct order to present them.

05 - Instagram

Instagram on Google Play has clear screenshots on the web which scales down for mobile.

After you upload your screenshots to the app stores, you’ll need to compare how they look on web (laptop and desktop) and mobile (phone and tablet). You can upload screenshots for each screen size. This will prevent them from looking perfect on one device/app store, but stretched or dense on another. For example, the iPhone 6 Plus would need screenshots sized 1242x2208px, while an iPhone 5 would only require ones that are 640x1136px (here’s a handy guide to common device resolutions).

The Wrap-Up

We did our own testing, and we found that  a 10% improvement in your CVR can lead to 10% more app installs or 10% less user acquisition spend to reach your target DAU/MAU. In short, no matter what, optimizing your screenshots will improve your user acquisition. You can also use Advertising Analytics to see the results of your marketing optimization.

There are many methods and services to A/B test, and the “right answer” won’t be the same for every company. Be diligent and keep testing. In time, you’ll discover which screenshots are leading to the most downloads, and you’ll see the fruits A/B testing efforts.

Key Takeaways:

  • Optimize the order of your screenshots so the highest-converting one (among your target users) is first. Use tools to determine this.
  • A/B test with tools(we listed the ones in this article below) to save your user acquisition team time.
  • The “best” A/B testing tool depends on your budget and the app store your app lives on.
  • Be mindful of screen sizes — your users could find you on mobile or desktop, and you want your screenshots to look good on both.


Want More? App Annie Has It

Subscribe to our blog to see more app industry reports, conference updates and product news. We love hearing from our readers, so share your questions — or tips you’ve discovered yourself — below or on Twitter.

Want to perform your own analysis? Create a free Store Stats account or request an Intelligence demo today.

October 2, 2015

App Store Optimization
Market Data

Related blog posts