When an account is judged on blended return, returning-customer revenue quietly carries the numbers and new-customer acquisition is allowed to stall. Blue Illusion's growth was capped there, not on the platform.
The constraint.
The reported numbers looked acceptable because blended performance was being flattered by existing customers. Judge acquisition that way and you cannot see that the new-customer engine has slowed, which is exactly the engine that funds real growth.
Budget was also too fragmented to learn cleanly, so the creative and audience signal the account needed was never strong enough to act on with confidence.
What we changed.
We separated acquisition from retention and judged it on new-customer economics, not blended MER. That alone changed which campaigns looked healthy and which were quietly losing. We concentrated the testing budget so it could exit learning, and rebuilt the account structure around acquiring new customers profitably.
With the structure pointed at the right job and the measurement honest about it, spend could move without breaking the economics.
The outcome.
Google new-customer growth doubled and Meta lifted 79%, inside the first 90-day engagement. Both came from removing a structural and measurement constraint, not from spending more for its own sake.
Acquisition has to be judged on new-customer economics, not blended return.
The moment returning-customer revenue is allowed into the read, a stalling acquisition engine looks fine until the spend ladder breaks something. Separate the two and the real picture appears.