Achieving a 6x revenue lift with predictive analytics and incrementality

rocket studio success story - screen

6x

increase in revenue

42%

rise in day 7 ROAS

24 hours

time within which predicted values were delivered

Overview

  • Vietnamese game developer Rocket Studio was struggling with data visibility and campaign measurement on iOS as a result of SKAdNetwork changes.
  • As a hyper-casual gaming provider with growth ambitions, effective user acquisition was key. The team also wanted to identify players with high lifetime value in order to boost ROAS.
  • Using AppsFlyer’s Predict solution, and working closely with Facebook Gaming, Rocket Studio was able to use predictive metrics to successfully optimize its campaigns and acquire more valuable users. 

Background 

Rocket Studio is a Vietnamese gaming studio that primarily develops arcade action games, including the popular and highly-rated “Galaxy Attack: Space Shooter”. The company aspires to become the leading game studio in every category it participates in, bringing the best experience to players while bolstering the position of Vietnamese games on the global stage.  

As one of the first hyper-casual gaming studios in Vietnam, Rocket Studio initially focused on exporting its games to the US, which still represents its largest and most valuable market. Since launching in 2015, the company has expanded into more than 50 countries around the world and established itself as a highly influential player in the gaming landscape. 

In addition to its core business of game development, Rocket Studio has started to tap into the game publishing space, acquiring the rights to publish games from smaller studios in the Vietnamese market. 

Challenge 

Rocket Studio had been successfully using AppsFlyer’s measurement suite since 2017 as part of OneSoft, one of the largest gaming studios in the Asia-Pacific region. In addition to measuring the effectiveness of their UA campaigns, the team was focused on acquiring players with high lifetime value (LTV), achieving high Return on Ad Spend (ROAS), and lowering Cost Per Install (CPI)

But when Apple introduced its SKAdNetwork (SKAN) protocol and App Tracking Transparency (ATT) framework, Rocket Studio, like many advertisers, faced a number of new measurement challenges for its iOS campaigns. And with iOS accounting for nearly a third of Rocket Studio’s overall traffic, the team needed to find a solution that would help them avoid major revenue loss. 

SKAN’s restricted 24-hour postback window proved especially complex for Rocket Studio’s longer payout window, making it challenging to identify potential high-value players and accurately measure campaign effectiveness. The team also had difficulty balancing their ROAS and CPI, and scaling their UA campaigns on iOS, due to limited measurement and data visibility. 

Solution

Rocket Studio collaborated closely with AppsFlyer and Facebook Gaming to design a holistic solution that would enhance the team’s measurement capabilities and improve ROAS, resulting in tangible revenue outcomes. 

The teams developed a two-pronged approach. First, they identified AppsFlyer’s predictive analytics solution (Predict) as an ideal starting point for overcoming SKAN limitations. Second, they leveraged Facebook Gaming’s value optimization capabilities to translate predictive insights from AppsFlyer into acquiring high-value players.

Predictive insights enabled Rocket Studio to turn SKAN campaign performance measurements from a campaign’s initial 24 hours into projected day-30 performance metrics. These covered day-30 ROAS, average revenue per user (ARPU), predicted percentage of paying users, and rolling retention. 

These insights are made available through a unique predictive model that is built around each app’s business logic. It correlates initial user engagement measurements and their eventual results, based on advanced analysis of Rocket Studio’s historical data. 

The Predict solution then communicates these values to Apple’s SKAN protocol directly, turning them into SKAN conversion values that encapsulate all the measurements from the 24-hour window and delivering feedback directly to integrated media partners. 

Armed with predictive data from AppsFlyer, it was time to put it to the test. Together, AppsFlyer and Facebook Gaming designed a holistic, geography-based framework to measure incremental lift in ad revenue. This would enable better ad delivery and measurement for Rocket Studio, and enhance optimization capabilities for its iOS campaigns. 

AppsFlyer’s incrementality solution relies on a time-based regression methodology that estimates the incremental, causal effect that an advertising intervention has on a response metric — in Rocket Studio’s case, ad revenue. The geography-based, comprehensive experiment design ensured that test and control groups were properly balanced and highly correlative so that the results of the test could be attributed to the incremental effect of the campaign and not to differences within the groups or other variables.

Using the predicted revenue metric from AppsFlyer’s Predict solution, the teams worked to optimize their activity, taking into account the predicted lifetime value of players. Delivered within just 24 hours, the predictive values provided Rocket Studio with actionable insights into projected campaign performance within Apple’s SKAN framework. The predictive values were communicated seamlessly to Facebook Gaming’s campaign optimization engine through the SKAN framework itself.

The last step was to devise a campaign geared towards acquiring higher purchases and compare the results with Rocket Studio’s typical campaigns that center around app install optimization, to observe the potential improvement in ROAS. To accomplish this, Rocket Studio worked with Facebook Gaming to configure the predicted values as value set ranges to facilitate value optimization.  

Results 

The framework that the teams developed proved to be highly effective in boosting Rocket Studio’s numbers. 

The Facebook Gaming value optimization campaign that was based on predictive insights from AppsFlyer generated a 6x improvement in Rocket Studio’s revenue, compared to the test’s control group. The company also measured a 42% average boost in day 7 ROAS compared to its usual app install campaigns, cementing this strategy as a success. 

AppsFlyer’s Predict Solution is a key component of our iOS campaign optimization and retention strategy, helping us acquire more valuable users with a high payout probability.

Amari Nguyễn Diệu, UA Marketing Leader, Rocket Studio 

As a longtime customer and an early adopter of the AppsFlyer Predict beta solution, Rocket Studio provided invaluable feedback to improve the solution’s capabilities. Their successful outcomes are the result of a longstanding, productive partnership with both AppsFlyer and Facebook Gaming and set a promising precedent for leveraging Predict with Facebook for iOS campaign value optimization.

Facebook Gaming has always been one of our go-to UA channels. Post-ATT, we’re confident that optimizing for revenue in Facebook Gaming ad campaigns using AppsFlyer Predict and Incrementality is a reliable approach for bringing in high-quality users and boosting ROAS. 

Amari Nguyễn Diệu, UA Marketing Leader, Rocket Studio
Thanks for your download!
Background
array(4) {
  ["title"]=>
  string(35) "Ready to start making good choices?"
  ["buttons"]=>
  array(2) {
    ["button1"]=>
    array(7) {
      ["enabled"]=>
      bool(true)
      ["text"]=>
      string(13) "Contact sales"
      ["url"]=>
      string(37) "https://www.appsflyer.com/start/demo/"
      ["style"]=>
      string(34) "btn btn-primary btn-primary--white"
      ["size"]=>
      string(6) "btn-lg"
      ["target"]=>
      string(5) "_self"
      ["id"]=>
      string(0) ""
    }
    ["button2"]=>
    array(7) {
      ["enabled"]=>
      bool(true)
      ["text"]=>
      string(12) "Sign up free"
      ["url"]=>
      string(32) "https://www.appsflyer.com/start/"
      ["style"]=>
      string(39) "btn btn-text btn-arrow btn-arrow--white"
      ["size"]=>
      string(3) "btn"
      ["target"]=>
      string(5) "_self"
      ["id"]=>
      string(0) ""
    }
  }
  ["image"]=>
  bool(false)
  ["section_settings"]=>
  array(1) {
    ["hide"]=>
    bool(false)
  }
}
Ready to start making good choices?