Should you double your UA budget? Check for a retention leak first
They asked me to double the UA budget. I told them to stop spending.
A mobile-gaming studio brought me in to build a scale plan. Real revenue, a growing install base, a game people liked. The ask was simple: more budget, more installs, more revenue. Actually, the first ask was even simpler: find a cheaper channel.
I ran the audit before I touched the budget. Here is what it found, and why the scale plan died.
Week one looked clean
Every campaign in that account was profitable the day we scored it. CPI was fine. Day 1 retention was fine. The dashboard was green.
That was the problem. A campaign gets judged in week one. The players it bought were leaving between day 30 and day 90. Those two clocks never meet, so the scoreboard was not lying. It was reading the wrong clock.
My own first read stood on the wrong clock too. It took looking at the cohorts two and three months out to see it: players installed, played, then drifted away, long after the campaigns that bought them had been called a success.
Acquisition was not the constraint
Acquisition was working. The game was losing players faster than the campaigns could replace them. Every new cohort landed in the same place the last one had already left.
So doubling the budget would not have doubled revenue. It would have bought a more expensive version of the same month.
The cause: the game ran out of reasons to come back
Nobody could point at a bad level or a broken feature. The regulars had simply finished the menu. There was no content variety deeper in the game, no events, no tournaments.
The fix was liveops. Every studio already has a liveops roadmap. What changed was how we chose what to ship: by A/B test, not by opinion. Whoever argues hardest in the meeting is not a testing method.
The result
Average daily revenue went from $3,524 to $4,081. A 15.8% lift, with no increase in acquisition spend.
One studio, one quarter-on-quarter comparison. It shows how the diagnosis worked, not what to expect for your game.
The part you will argue with
Creative is not a CPI lever. It is a cohort-quality lever. The ad writes the expectation the first session has to meet, and a cheap click buys a cheap expectation.
So the cheapest creative in your account is not automatically your best one. You only find out at day 60.
The move
- Before any scale plan, pull retention by cohort out to day 90. Look at the shape, not the day 1 number. If the curve is still falling at day 60, you have a leak.
- Cut your cohorts by creative, not just by campaign. Read them at day 60 instead of day 7. If your cheapest asset is also your best at day 60, you are fine.
- Fix the leak on the budget you already have. No new spend, no new channels, no new vendors.
- Choose fixes by test. Ship the liveops or product change to a test group first, and let revenue per user decide.
- Only then write the scale plan. Now every new dollar buys players who stay.
Bottom line
A scale plan is the most expensive way to find out your product has a problem. Budget does not fix retention. It just buys more people to lose.
Diagnose. Then decide. Then spend.
FAQ
Should I increase my UA budget if my campaigns are profitable?
Not until you have checked retention past day 30. Campaigns are usually judged in the first week, but many games and apps lose most of their value between day 30 and day 90. If that curve is still falling, more budget buys more of the same churn.
How do I know if I have a retention problem or an acquisition problem?
Compare cohorts over time. If new cohorts arrive at a fair cost but keep dropping off at the same point, the constraint is retention. If cohorts stay but cost too much to acquire, it is acquisition. Look at day 30, 60 and 90, not only day 1 and day 7.
What fixed the retention leak in this case?
Liveops: content variety, events and tournaments, chosen by A/B testing rather than opinion. Average daily revenue rose from $3,524 to $4,081, a 15.8% lift, with no change in acquisition spend.
Why judge creative at day 60 instead of by CPI?
Because the ad sets the expectation the product has to meet. A creative that overpromises buys cheap installs who leave early. Day 60 cohort revenue by creative shows which ads bring players who stay.