When your ROAS looks great, you're probably underspending
Every quarter it's the same conversation. Blended ROAS is sitting at 4 to 1, comfortably above target, the growth dashboard is green, and the founder wants to know why revenue has been flat for two quarters.
The answer is usually hiding in the same sentence. The ROAS is 4 to 1.
A return on ad spend that looks perfect almost always means you stopped buying right before it got interesting. You are the driver watching the fuel economy gauge instead of the road, proud of 52 miles per gallon, doing 40 in the slow lane, two hours late to where the revenue is.
The first dollars you spend are always the cheapest
Every ad platform serves your easiest conversions first. The opening tranche of budget reaches people who were already reaching for their wallet: branded search, warm retargeting, users a friend already told about you. That spend comes back at 6 to 1 because it barely did anything. Those people were going to convert anyway.
Push daily spend higher and you start paying to reach people who need convincing. That tranche comes back at 2 to 1. Blend the two and the report reads 4 to 1, and 4 to 1 reads as healthy, so nobody asks the next question.
Here is the next question. How many profitable customers are standing in a queue you have decided not to serve?
I once cut a games studio's best line item, the one with the cheapest installs and the highest ROAS in every budget meeting. We tested it the only way that survives bad attribution, paused it in matched regions and watched revenue. Almost nothing happened. The channel was re-buying players who would have installed anyway and taking the credit. In one public dataset of 225 incrementality tests, branded search came in at a median 0.70x incremental ROAS, below breakeven, a dollar in and seventy cents out. The prettiest ROAS in that account was the least real.
The only ROAS worth a decision is the marginal one
Blended ROAS tells you what your spend so far returned. It is the average speed for the whole trip, and it is useless when the question is whether to press the accelerator right now. The question that decides whether to scale is narrower: what would the next 500 dollars a day return?
That number, the marginal ROAS, is always lower than the blended one, and it is the only one that matters. If the next tranche of spend still clears breakeven, you are underspending. Not "you could consider spending more." You are leaving money on the table right now.
Breakeven is 1 divided by your contribution margin. At a 60 percent margin, that is 1.67x. Anything you are turning away that clears 1.67x on an incremental basis is profit you chose not to make.
One mobile-gaming studio I worked with had held UA spend flat for two quarters because ROAS was above target and nobody wanted to touch a good number. That instinct is understandable. You fought for that number, the board finally stopped asking hard questions, and now someone wants you to risk it. A marginal test, raising one campaign's budget in matched regions and watching incremental revenue, showed the next block of spend still cleared breakeven by a wide margin. The flat revenue line had never been a market limit. It was a decision, made quarterly, by not deciding.
When a high ROAS is a real thing to protect
This is not "always spend more." A high ROAS is a question, not an answer, and twice the honest answer is to hold.
The first case is retention. If your day 7 curve is still falling and has not flattened by day 30, scaling into it just means paying more to lose the same users faster. I once froze a client's budget entirely while we rebuilt onboarding and the first ten minutes of the product. Day 7 retention went from 16 percent to 23 percent, day 30 ROAS improved 34 percent, and only then was the thing underneath worth amplifying. Spending during those weeks would have bought a more expensive version of the same churn.
The second case is incrementality, the branded-search problem above. If a geo-holdout shows revenue barely moves when you pause a channel, its ROAS is not a signal to scale. It is a line to cut.
So the rule is not "green means go." The rule is that a ROAS comfortably above breakeven is telling you to go find out why.
The move
- Stop using blended ROAS for scaling decisions. Pull the marginal ROAS instead: raise one campaign's daily budget by about 30 percent, hold everything else constant, and measure incremental revenue over two to four weeks. Use matched-region holdouts if you have the volume.
- Write your breakeven line on the media plan. It is 1 divided by your contribution margin. Everyone on the team should be able to say it out loud.
- If marginal ROAS is above breakeven and retention is stable (the curve flattens by day 30), scale that campaign until the marginal ROAS approaches breakeven. That point, not your target, is the actual ceiling.
- If marginal ROAS collapses fast, or the channel fails a geo-holdout, you have found either a real market limit or a channel that was never incremental. Both are useful. Neither is a reason to scale.
- Re-run the marginal test every quarter. The ceiling moves with your creative, your product, and the season.
The source
I have built UA budgets for consumer apps and mobile games for over a decade, across more than 150 titles. The single most common problem I find is not a founder overspending. It is a founder protecting a ROAS number that was quietly telling them to spend more.
Bottom line
A perfect ROAS is not a trophy. It is a speedometer stuck below the limit.
Find the marginal dollar. Spend to the breakeven line. Watch the road, not the gauge.
FAQ
What is a good ROAS for a consumer app?
There isn't one. A blended ROAS well above breakeven usually means you are underspending. The number to watch is marginal ROAS, the return on your next tranche of spend, measured against your breakeven line of 1 divided by your margin.
How do I calculate breakeven ROAS?
Divide 1 by your contribution margin. A 50 percent margin gives a 2.0x breakeven. A 60 percent margin gives 1.67x. Below that line you are losing money on every incremental sale, above it you are making money even if the number looks unimpressive.
What is marginal ROAS and why does it matter more than blended?
Marginal ROAS is the return on the next block of spend, not the average across everything you have spent so far. It is always lower than blended because the cheap, high-intent demand gets bought first. It is the only version of the metric that tells you whether scaling makes money.
Should I always scale when ROAS is above target?
No. Two exceptions. If retention is not stable, meaning the curve is still falling at day 30, you will pay more to lose users faster. If the channel is non-incremental, meaning a geo-holdout shows revenue barely moves when you pause it, you are paying for conversions you would have gotten free.
How do I test marginal ROAS without a full incrementality setup?
Raise one campaign's daily budget by roughly 30 percent, hold every other variable constant, and compare incremental revenue over two to four weeks against the extra spend. It is rougher than a matched-region holdout but it answers the scaling question.
My ROAS drops every time I increase spend. Is that a bad sign?
Not on its own. It is expected. Efficiency falls as you reach colder audiences. The only question that matters is whether marginal ROAS stays above breakeven. If it does, higher spend at lower efficiency still means more total profit.