+57% CVR: What Avakin Life Learned Running 50 Experiments at Once

When Lockwood Publishing set out to improve conversion rates for Avakin Life on Google Play, they took a different approach. Instead of hiring more designers, adding new tools, or running a single creative refresh, they focused on optimizing their experimentation process.

They automated the entire experimentation loop.

This led to a 57% increase in conversion rate. (Publishing, 2024) The improvement didn’t come from a single creative, but from running over 50 experiments at once across icons, screenshots, and descriptions, with each cycle building on the last.

Here’s a look at what they tested, what they learned, and why their approach made the difference.

The starting point

Avakin Life is a well-established title in the life simulation genre, with millions of downloads and a strong user base. The design team was already producing high-quality creative assets.

The challenge wasn’t creative quality. Lockwood’s designers were already delivering solid work. The real challenge was the speed of testing.

Like many publishers, their testing process was slowed down by manual steps. They could only run one experiment at a time, with weeks between cycles. As a result, many hypotheses sat in the backlog, waiting to be tested.

The store listing was performing reasonably well. However, in a competitive genre, not testing regularly means missing out on potential installs.

What they tested

The experimentation scope covered every testable element of the Google Play store listing:

Icon variants included changes to color palettes, character poses, background treatments, and framing. These were subtle adjustments designed to test specific ideas about what attracts attention in browse and search results. They tested the balance between gameplay footage and aspirational imagery. Text overlay placement. The number of screenshots visible before scrolling. The story arc across the screenshot set.

Feature graphics tested composition, color dominance, character positioning, and text weight.

Short descriptions tested benefit-first vs. feature-first framing, social proof inclusion, and character count optimization.

A key detail: these experiments ran at the same time, not one after another. Over 50 experiments across different elements generated data simultaneously, all contributing to the learning process. (Kolayan, 2025)

The results: 57% CVR lift

The 57% increase is impressive, but the details behind it are even more valuable.

No single experiment delivered the full 57% lift. Instead, improvements added up over multiple cycles. An icon test brought a few percentage points, a new screenshot layout added more, and optimizing descriptions and feature graphics contributed further gains.

With each cycle, the system learned what worked and generated better hypotheses. Experiments improved over time by building on previous results.

This is the compound effect at work. The first round of experiments brought modest gains, and each following round improved as it built on earlier results. By the tenth cycle, the system had a clear understanding of what drove conversion for Avakin Life and was able to focus on the most effective opportunities.

The 57% increase wasn’t the result of a single creative idea. It was the outcome of a system that learns and builds over time.

Velocity, not creativity, made the difference.

One insight that stood out to the Lockwood team: the winning variants were not dramatically different from the originals.

The winning icon was a subtle color shift and a small change in character positioning, not a complete redesign. The best-performing screenshot layout adjusted one element and changed text weight. The top description changed the opening line and added a social proof element. Any designer could have created these variants. The key was not creative talent, but the ability to test many options, learn quickly, and apply those learnings to the next round. Running one experiment per month could eventually yield the same results, but it would take years.

By running 50 experiments at once, the system found the best variants in just a few weeks.

Testing speed is the multiplier, while creative quality is the foundation. Both matter, but most teams don’t invest enough in increasing their testing velocity.

Applying this to your portfolio

While these results are specific to Avakin Life, the same principles apply to other titles and genres. Your numbers may vary, but the approach remains relevant.

Start Begin by looking at your current conversion rate. For most titles with over 100,000 monthly downloads, even a 10% improvement can mean thousands of extra installs each month. Consider what a 20% lift could mean for your app over a quarter or a year. (Curry, 2025) Look at your testing cadence. How many experiments did you run last quarter? How many hypotheses are sitting untested? What is the gap between your current velocity and what the Google Play experiment infrastructure actually supports?

The difference between those numbers is your opportunity. Every month of slow testing is also a risk, since competitors can use the same data to improve their results.

Avakin Life’s 57% conversion rate lift began with one decision: move from manual experiments to a fully automated testing process.

The same system that delivered +57% CVR for Avakin. The same system that helped Avakin Life achieve a 57% increase in conversion rate is available for your titles. Book a demo to see PressPlay in action.

References

Publishing, L. (2024). How AI-Driven A/B Testing on Google Play Scaled Avakin Life’s Conversion Rates by 57% in Two Months. Phiture. https://phiture.com/success-stories/pressplay-avakin-life-57-percent-cvr-increase/

Kolayan, A. (January 1, 2025). Phiture’s success with PressPlay game. LinkedIn. https://www.linkedin.com/posts/arthurkolayan_we-at-phiture-launched-our-ai-driven-ab-activity-7214184500468064257-13Lr

Curry, D. (2025). Mobile Game Conversion Rates (2025). Business of Apps. https://www.businessofapps.com/data/mobile-game-conversion-rates/