Choosing predictive events that still have volume

Ashley Black, founder of Candid Consulting and former longtime Googler, breaks down how (and when) subscription apps should switch Google App Campaigns from CPA to tROAS, the pitfalls that stall performance, and how to feed better signals (activation/retention events) for durable scale. We also dig into iOS vs. Android realities, exclusions that actually matter, and why “automated” ≠ “set-and-forget.”

What you’ll learn

• The most common mistakes when moving from CPI/CPA to tROAS (targets too high, windows too long)
• How to set a realistic ROAS target (start ~20% below goal) and ramp it without killing volume
• Volume prerequisites for value bidding (why you need revenue events, not just trials)
• When tROAS fits (risk tolerance, trial length, budget) and when to stay with CPA
• Android vs. iOS with Google: inventory, tracking constraints, and creative needs (YouTube/Shorts)
• The right exclusions to apply (existing users, brand, re-installs) and why CPM rising can be good
• Using early activation/retention events to improve optimization when trial-start isn’t predictive

Key Takeaways

• Don’t over-ask early. Setting day-7 ROAS targets too high and using 30–90 day windows starves delivery. Start with a short window (≈7 days) and a lower target, then stair-step up.
• You need real revenue signals. For tROAS to learn, pass purchase/subscription events—trial-start alone won’t cut it. Rule of thumb: aim for ≥10 post-install revenue events/day (often more).
• Trial length matters. 30-day trials delay signals; tROAS may burn spend blind. Shorter trials or earlier monetization events make tROAS viable.
• Expect a ramp-up. Some accounts stabilize in days; aggressive targets can take weeks to unlock. Be patient and ready to lower targets to gain learning volume.
• Scale vs. profit trade-off. CPA often scales easier; tROAS can be more profitable once learned. Consider geo split tests to compare mixes.
• Inventory shifts under tROAS. Eligible placements are the same, but you may see more search/Play and higher CPMs—often a sign of higher-quality traffic, not waste.
• Exclude smartly. Add exclusions for current users, brand queries, and (optionally) re-installs to protect incrementality.
• iOS = different game. Google’s iOS performance lags Android; expect more YouTube/Shorts traffic and lean on strong UGC-style video. Treat iOS Google as a later-stage test.
• Optimize for activation. If trial-start users don’t retain, bid to an early in-app action (e.g., completed tutorial, first message) that correlates with D1/D7 retention and occurs fast enough for learning.
• Automation needs adults in the room. UAC/PMAX aren’t fire-and-forget—active tuning (targets, assets, exclusions) still moves the needle.

Links & Resources

• Ashley Black — Candid Consulting: https://www.candidconsultinggroup.com/
• Ashley’s guide to tROAS for subscription apps: https://www.botsi.com/blog-posts/value-based-bidding
• Connect with Ashley on LinkedIn: https://www.linkedin.com/in/ashleym-black/