How We Beat Drimage’s BTS Island ROAS D21 KPI by 46% from Week 1
Download the PDF version of the case study
We partnered with DRIMAGE to retarget lapsed players of their match-3 puzzle game, BTS Island: In the SEOM Puzzle, on iOS in the US and Japan.
The three-week campaign was timed to the game’s fourth anniversary, when limited-time events, rewards, and in-game campaigns created a strong opportunity to bring previously active players back to the game.
Results at a Glance
46%
ROAS D21 above the KPI on Week 1
2.1x
ROAS growth from D7 to D30
25%
Higher ROAS D30 from IDFA-less audiences
95%
Of budget on full-screen video creatives
Campaign Goal
The primary goal was to re-engage lapsed players and bring them back to BTS Island to make additional in-app purchases.
We targeted two segments: payers who had previously made in-app purchases but had since become inactive, and non-payers who had been active in the game but had not made any purchases.
The campaign was measured against a set of ROAS-based KPIs that varied by maturation cohort.
Unlike always-on retargeting, DRIMAGE’s campaigns are activated around key in-game events, creating a limited window to identify and re-engage the right users.
The value of a strong retargeting partner is not simply in delivering against a campaign brief, but in understanding the objective behind it and finding opportunities to improve performance. Over the years of our collaboration with Persona.ly they have consistently demonstrated that approach, helping us maximize the impact of event-driven campaigns while delivering beyond our ROAS targets.
Yoshida Mayu – UA Manager at DRIMAGE
Programmatic Retargeting Process
From the start of the campaign, we leveraged Persona.ly’s proprietary Live Audiences segmentation engine to build custom audiences.
Live Audiences classifies users across multiple signals, including activity level, purchase history, inactivity window, and intent. Directly integrated with Persona.ly’s programmatic DSP, the engine allows audiences to be updated and suppressed in real time.
For BTS Island, we built two audience segments: previous purchasers who had lapsed, and previously active non-payers who had been inactive for a specified period.
The short campaign window left little room for a traditional learning phase. Our ML-driven DSP began optimizing from campaign launch, using early signals to identify which audience segments were more likely to click, re-open or re-install the game, and make an in-app purchase.
As our bidder accumulated enough data to identify which audience segments and placements would bring the strongest performance, it started shifting the budget and adjusting bids to achieve the highest ROAS.
This approach allowed us to outperform the ROAS D7 KPI by 20% from the onset and maintain the performance throughout the campaign
Ensuring Campaign Incrementality
The campaign coincided with a major anniversary event, ensuring that paid re-engagement was incremental was particularly important. Our approach across re-engagement campaigns includes prioritizing click-through attribution and highly-engaging creative formats.
Click-through Attribution
We focused on click-through attribution rather than relying on view-through attribution.
For a title with a highly engaged fan audience, view-through attribution can mistakenly credit campaigns for users who would have returned to the game organically for the anniversary event.
By requiring an active interaction with the creative, click-through attribution provided a stronger signal that the campaign contributed to the re-engagement.
Full-screen Video Creatives
More than 95% of campaign spend ran on full-screen video creatives.
The format commands active attention from the user and helped us avoid accumulating impressions alongside sessions that users were likely to initiate organically.
Within the first few days our ML-driven DSP collected enough engagement data to identify creative champions (the strongest-performing creatives), and shift delivery toward them, contributing to the success of the campaign.
Campaign Results
ROAS Progression by Market
Across the three-week campaign, we consistently outperformed the ROAS-based KPIs. Performance strengthened further as cohorts matured, with ROAS growing 2.1x from D7 to D30 and continuing to rise beyond D30.
Across the two markets, the ROAS curve evolved differently over time.
The US audience showed stronger early ROAS, outperforming the D7 ROAS KPI by 42% and reaching 2.9x the D7 ROAS KPI at lifetime.
The Japan audience started more conservatively at D7, but continued to mature over time. By lifetime, it reached 3.5x the D7 ROAS KPI, ultimately delivering the stronger long-term performance.
IDFA-less vs. Deterministic Audiences
Reaching both probabilistic and deterministic audiences allowed us re-engage a large share of the churned user base, despite the limitations of iOS identifiers.
Both audience setups (probabilistic and deterministic) exceeded the D7 ROAS KPI. However, the difference became more pronounced as cohorts matured.
By D30, probabilistic audiences delivered 25% higher ROAS than deterministic audiences, and held that advantage through lifetime.
The probabilistic audience continued to outperform as the cohorts matured, demonstrating that IDFA-less audience retargeting could uncover high-value opportunities beyond deterministic matching.
Creative Performance
Our ML-driven DSP identified top-performing creatives and shifted delivery toward them within the first week of the campaign.
Combined with real-time audience optimization and dynamic bid adjustments, this approach helped maximize performance.
About DRIMAGE
DRIMAGE is expanding the boundaries of HYBE’s entertainment spectrum by developing games and conducting publishing operations to interact with users worldwide. Through the combination of cutting-edge interactive media technology with music, artists, and entertainment elements, DRIMAGE presents a variety of content and services that offer new perspectives and experiences.
We place a lot of value on building long-term partnerships with leading Korean developers. Our partnership with DRIMAGE is a great example of that.
This campaign reflects the approach we’ve built around iOS retargeting: combining multiple signals with machine learning to identify valuable users when signals are limited. It’s a key reason we’ve been recognized by AppsFlyer as a top media
for iOS retargeting in US gaming and by Singular as the top ROI-driving partner.
Rachel SEO – Business Development Director, Korea at Persona.ly
About Persona.ly
Persona.ly is a data-driven product company specializing in mobile user acquisition and retargeting, powered by proprietary machine learning algorithms and a robust first-party data management platform.
We help leading mobile companies including King, Tencent, Papaya, NextNinja, Tilting Point, Nexon and many others reach and exceed their growth goals by accurately predicting which users are most likely to engage and convert. Our machine learning models are optimized to target high-value users based on predicted LTV, driving strong ROAS and long-term user quality.
Our in-house programmatic DSP is directly integrated with top-tier ad exchanges and OEMs, processing over 4 million ad requests per second globally.
By combining real-time predictive analytics with programmatic scale, we empower our partners to unlock smarter growth, higher LTV, and measurable incremental impact.
With IMPACT, our incrementality measurement platform, advertisers can go beyond attribution to understand the true incremental value generated by Persona.ly DSP. IMPACT provides transparent measurement across audiences, segments, and campaigns, connecting media exposure with incremental conversions, revenue, and iROAS.
Thank you for your interest!
Our team will get in touch with you shortly.
