As competition intensifies and CPIs continue to rise, sustainable growth in mobile gaming demands more than short-term ROAS optimisation.

In this interview, Anton Zaluskyi, founder of Smart Project, explains how predictive modeling, full automation, and a relentless focus on day-180 pROAS are reshaping the company’s approach to user acquisition.

From budget allocation to creative strategy, he outlines why long-term value, not early performance spikes, defines real UA success.

How do you determine whether a UA campaign is creating durable long-term value rather than short-term ROAS?

We use a predictive model that is based on two key metrics: Retention and ARPU day-to-day growth rate. These allow us to build an automatically accurate long-term forecast for each campaign on a daily basis.

After the model finishes its learning period, we can quickly understand whether a campaign is generating long-term value or short-term ROAS only. This happens very early in the campaign lifecycle, without waiting for many months of real data.

Based on this forecast, the system automatically decides what bid level is appropriate for each campaign. Campaigns with a strong long-term potential receive less aggressive bids, while short-term campaigns are the opposite. This approach helps us focus our budget on predicted profit no matter whether the campaign has long-term value or short-term ROAS.

How has your approach to budget allocation across channels changed as CPIs continue to rise?

Today, our budget allocation is fully automated and happens on a daily basis.

We use a self-developed system that distributes budget based on the performance of each channel and each UA campaign. Performance is calculated automatically using a predictive model that forecasts ROAS on day 180.

Because of this, we don’t rely on manual decisions or fixed budgets per channel. The system continuously moves the budget toward campaigns with the highest potential profit. This allows us to stay efficient even as CPIs grow and competition becomes stronger.

Which single metric matters most for UA success in long-lifecycle games - and why?

For us, the most important metric is pROAS on day 180.

Our main goal is maximum efficiency of spend, and pROAS d180 gives the clearest view of real profitability over a long lifecycle. Short-term metrics can look good, but they don’t always reflect the true long-term value of a player.

By focusing only on long-term predicted ROAS, we make more accurate decisions and avoid scaling campaigns that look good early but fail over time.

How do you connect creative insights back to product?

We have an automated system that tracks creative performance for each source and game on a daily basis. This data is shown as a daily ranking of top-performing creatives and is available to the entire creative team.

The motivation system for the creative team is also tied directly to creative performance. This creates a clear connection between results and rewards.

Because of this setup, insights from UA quickly flow back into creative production and product thinking. We clearly see which mechanics, visuals, and messages work best and can use this knowledge both in new creatives and future product decisions.

Looking ahead, which UA decisions do you expect to become fully automated in the next few years?

We believe that UA is moving toward almost full automation.

In this case, UA managers and their experience are mainly needed to decide the order of channel launches based on game genre, to visually monitor how well the system is working, and to give feedback to improve algorithms. All operational decisions, such as bid calculation and budget distribution, are handled automatically by the system.

This allows teams to focus less on manual optimization and more on strategy and product-level decisions.

How tightly integrated are UA and product teams at Smart Project when launching new features or titles?

Our creative team is involved in the product from the idea stage. They understand the core mechanics and key features long before release.

Some team members have direct access to the product source code and can record gameplay creatives even before a feature goes live. This allows us to test new gameplay hypotheses very early.

As a result, we can quickly validate ideas, prepare creatives in advance, and launch UA tests immediately when a feature is released. This tight integration significantly speeds up learning and iteration.

Which growth channel has surprised you most in the last two years for your games - and why?

AppLovin has surprised us the most.

Its player value prediction system is extremely strong and allows targeting with wider attribution windows. The ROAS forecasts are very accurate and adapt quickly to product changes.

Because of this, campaigns require much less manual control from the marketing team while still delivering stable and predictable results.

Looking ahead, where do you see the biggest growth opportunity for Smart Project: new markets, new formats, or new UA models?

We continue to improve our automated marketing management system and strongly believe that further AI development will allow us to automatically generate and iterate creatives in the near future.

Our automated marketing system significantly reduces both the time and the number of people needed in the marketing team. This allows us to concentrate talent in marketing and analytics instead of spending time on routine manual tasks.

For example, we can efficiently manage multi-million monthly budgets across many projects with a team of just two or three people. This clearly shows how automation changes the role of UA teams and why more decisions will be handled by systems rather than humans.

In addition, we see strong potential in working with indie developers through our publishing department. By combining indie products with our UA expertise and automated marketing system, we believe we can unlock significant growth for both sides.