Most agency relationships don't fail loudly. They fail quietly — the account hits its agreed numbers, nobody's technically doing anything wrong, and growth just... stops. This is the pattern, and how to check for it yourself.
Most retainers pay a flat fee for a flat scope. Once an agency hits the agreed target, pushing further costs them more hours for the same fee — so the rational move, from their side, is to stabilise at "good enough" rather than keep pushing. Nobody signs a contract that says this. It just happens, quietly, once the account stops being a problem worth solving.
The tell isn't bad performance. It's flat performance — hitting the same number month after month, with no real attempt to move past it.
The most common way this shows up: an account's blended ROAS looks healthy because branded search — traffic that would have converted anyway, for close to nothing — is doing most of the work. Non-brand prospecting, the actual growth engine, quietly stagnates underneath a headline number that still looks fine.
See the full worked example on this mechanic →None of these prove anything on their own. Two or three together is a real pattern.
| Signal | What it suggests |
|---|---|
| Same campaign structure for 12+ months | No real strategy refresh, just maintenance |
| Brand search's share of spend or revenue is rising | Coasting on near-guaranteed ROAS instead of prospecting |
| Performance plateaus exactly at the agreed KPI, never past it | Working to the target, not beyond it |
| No changelog of tests run or changes made | Nothing's actually being tested |
| Creative or ad copy unchanged for 6+ months | No real experimentation happening |
| Agency only initiates contact at contract renewal | Reactive relationship, not a proactive one |
The cycle repeats because nothing actually gets checked in between.
This isn't about assuming your agency is doing a bad job — most aren't. It's an independent check, using the same four-job scoring as everything else: is the account actually being pushed to Scale, or resting on Defend-level brand spend labelled as growth? Is anyone running a genuine Learn test, or has testing quietly stopped? If something's slipped, is Recover happening proactively, or only once you notice and ask?
This sits inside the Forensic Audit and the Initial Audit — it's not a separate product, just one of the areas that can be in scope when an agency relationship is part of what's being reviewed.
15 minutes. No pitch, just a look.
Book a free chat →