App installs increase 167% for popular 3D avatar mobile app
THE CHALLENGE
IMVU is a 3D avatar-based social experience with the largest 3D world and catalog to customize an avatar’s look. Users meet new people, chat with friends, and roleplay within the app. IMVU reached out to Liftoff to help them acquire new quality mobile users in the USA on iOS. Their main KPI was scaling in-app purchases at a set CPA cost.
THE SOLUTION
The IMVU iOS campaign started by analyzing the install to in-app purchase user flow, defining KPIs, and setting expectations. The user flow for IMVU is install > register > in-app purchase. In-app purchases available include over 20 million user-generated VR-ready goods to customize the looks of avatars within the app. To feed Liftoff’s machine learning platform, IMVU started sending all relevant postback data from AppsFlyer, a leading mobile app tracking & attribution analytics platform. Postback data included installs, registrations and in-app purchases.
From the start of the campaign, Liftoff conducted ad creative A/B testing, optimizing for click-thrus, CPIs, and in-app purchases. The team created multiple tests across banners, videos, and interstitials. Ad creative tests included comparing different avatar characters, copy variations, avatar groups, avatar hairstyles, and more.
The campaign started on a CPI pricing model (optimize for app installs). Liftoff’s machine learning (ML) platform programmatically started to bid on ads solely to learn which users were most likely to install the IMVU app. We exited the learning phase of the campaign just days after launching the campaign, moving to a CPA focus, aggressively seeking users most likely to make an in-app purchase.
As the campaign shifted to CPA event optimization, scaling was initially a challenge. First in-app purchase events happened too infrequent, not providing our machine learning platform enough data to optimize. After switching from a CPI to fixed CPA pricing model, the campaign gained significant traction, resulting in increased daily spend and spend consistency.
The Liftoff team has continued to exceed our expectations by always being on top of the campaign, proactively identifying new growth opportunities to scale up performance, minimizing any fraudulent activity and resolving any issues that came up around misalignment on data attribution. It’s refreshing to work with a partner who is actually as motivated as us to making this partnership successful.
Lomit Patel, Vice President, Growth at IMVU
THE RESULTS
As the campaign shifted to CPA event optimization, scaling was initially a challenge. First in-app purchase events happened too infrequent, not providing our machine learning platform enough data to optimize. After switching from a CPI to fixed CPA pricing model, the campaign gained significant traction, resulting in increased daily spend and spend consistency.
- In-app purchases increased by 276%
- Installs increased by 167%
- Install-to-action rates increased by 41%
On the creative side, we uncovered some noteworthy performance insights. A video test that compared avatars switching outfits vs switching hairstyles resulted in a 202% improvement in in-app purchases. Another A/B test pitted static native images of avatars with different fashion styles against a more generic image with fewer avatars and large copy. The ad with more avatars consistently outperformed all other creatives, resulting in an 85% improvement in the impression-to-app install rate.
Once we began optimizing for in-app purchases against a target cost-per-action cost, Liftoff was able to ramp up and scale ad spend, increasing the average daily ad spend by 217%. This is a great example of the benefits of optimizing for post-install conversions, where a clearly defined CPA target backed by ML and creative tests resulted in a high performing mobile user acquisition campaign.
This article first appeared on Liftoff.
