APUS built one of the world's most widely used Android launchers. The app speeds up devices, lets users customize home screens with themes and wallpapers, manages apps and files, and saves battery life. By the time APUS came to Epom, it was already the leading utility publisher in China – with a user base large enough that the next phase of growth had to come from somewhere else entirely.
That somewhere else was 23 international markets all at once.
The goal wasn't just to get installs; any campaign can pay for installs after all. APUS wanted users who would stay – people who found the app genuinely useful, came back to it, and became what the team called brand advocates. All of this while holding to a strict cost per acquisition that left no room for wasted spend.
The brief was simple enough to state but really hard to achieve: scale globally, get loyal users, and keep costs under control.
One App, One Billion Potential Users, One Problem
Expanding into 23 markets at once creates a specific kind of advertising challenge. It's not about budget or logistics – it's a relevance problem.
The APUS launcher does different things for different people depending on what phone they're using. For someone with an older Android device, the most important feature is speed. The phone gets faster, apps stop crashing, and battery lasts longer. That's a real, immediate, felt benefit.
For someone with a newer device, speed isn't the issue. What they care about is personalization – themes, wallpapers, making the phone feel like their own.
Same app. Same features. Two completely different reasons to use it.
You can't expect a single campaign creative to cover both users at once. And if you run it anyway, you're paying for impressions that just aren't going to resonate – and you'll lose users who would have converted if the message had been right
This was the problem APUS brought to Epom.
The Device in Their Hand Told Us What to Show Them
The solution started with a single insight: the device itself is the targeting signal.
Using Epom’s dynamic optimization, campaigns read device characteristics in real time and served the matching message. Older Android device – the ad focused on speed, performance, battery life. Newer device – the ad led with customization, themes, and interface features.
This wasn't a segmentation exercise done in a spreadsheet before the campaign launched. It happened dynamically, impression by impression, across 23 markets.
| User context | Ad message | Feature led with |
| Older Android device | Your phone, faster | Speed, battery, performance |
| Newer Android device | Make it yours | Themes, wallpapers, customization |
Every user who saw an APUS ad saw one that reflected something about their actual situation – not a generic pitch that assumed everyone had the same phone and the same problem.
Custom Smart Banners Built for This Campaign
Personalized messaging needs creative that can carry it.
Epom's in-house studio developed custom smart banners specifically for the APUS campaign. These weren't repurposed assets adapted for mobile – they were built for user acquisition across mobile ad channels, designed to work across the range of device types and market contexts in which the campaign would run.
Each banner was tailored to its audience segment and continuously refined throughout the campaign based on performance data. What worked in one market was tested against what worked in another. Underperforming variants were replaced. The creative evolved as the campaign ran.
And the result was consistent execution across 23 markets without sacrificing the local relevance that made the targeting approach work.
Optimizing for Retention, Not Just the Install
Most mobile acquisition campaigns are optimized for one thing: the install. You get the download, the event fires, the campaign takes credit. What happens to the user after that is someone else's problem.
APUS measured differently. The campaigns were optimized continuously against two metrics: user retention and ROI – not just install volume. And that changes the whole budgeting dynamic.
As performance data accumulated, spend shifted toward the traffic sources and audience segments that generated returning users. And sources that just produced installs but not retention started to get cut back or deprioritized.
| Optimisation approach | What it produces |
| Install-only | High volume, variable retention, CPI pressure |
| Retention + ROI | Higher lifetime value, CPA efficiency over time |
What the Numbers Looked Like at the End
25,000 installs per day. Cost per acquisition 50% below what APUS had originally budgeted for. 70% user retention rate. Campaigns running across 23 international markets.
The CPA result is the important one to focus on. APUS came in with a strict acquisition budget and a number they needed to stay under. And when the campaign ended, they'd done it – not by cutting corners, but because device-level personalization reduced wasted impressions and retention-focused optimization made each acquired user worth more in the long run.
The 70% retention rate is what makes that install number actually mean something. 25,000 daily installs from users who leave within a week is a different business outcome from 25,000 daily installs from users who stay, keep using the app, and become brand advocates. The campaign delivered the second kind.
"The APUS case shows what happens when dynamic optimization and strong creative work together. The install numbers are the visible result. The retention rate is the proof that the right users were being acquired – not just the easiest ones."
