Growth is the label agencies use when they want to be judged on outcomes rather than deliverables, which is a good instinct and an easy thing to fake. The test is simple: did the agency name a metric, a baseline and a horizon before it began, and is that metric one the business actually cares about?
The metric has to be agreed in advance, with a baseline
Sign-ups, qualified enquiries, activated accounts or revenue, measured against the trailing period, with the seasonality acknowledged. An agency that proposes a metric after three months has chosen the one that moved. Write the baseline into the contract along with what will be reported monthly, and agree who pulls the numbers, because an agency reporting on its own work from its own dashboard is marking its own homework.
Growth work usually crosses into the product
The highest-return changes are often onboarding, pricing presentation, activation emails and the signup flow rather than advertising. That means engineering time you have to supply, and an agency that has no route into your backlog will fall back to the channels it can control. Settle that access before the engagement or expect a media plan with a growth label on it.
Experiments are the method, and most of them fail
A real programme runs a queue of tests, expects most to be flat or negative, and keeps the winners. Ask for a test log from a previous client: how many ran, how many won, what the median effect was. A portfolio of only successes is a portfolio of the successes, and it tells you nothing about how they work.
Be careful with performance-based fees
Paying on outcomes sounds aligned and often is not: it rewards claiming credit for demand that already existed, and it makes attribution the contract rather than a diagnostic. If you do use it, define the attribution rule precisely, cap the upside, and make sure the agency cannot influence the measurement it is paid against.
Questions people ask about digital marketing growth agency
Is a growth agency different from a performance agency?
In principle yes. A performance agency buys media efficiently; a growth agency should be willing to change the funnel, including parts of it that are not advertising.
How long before a growth engagement shows anything?
A quarter is the usual minimum to run enough tests to learn something. Anything faster is a channel change, not growth work.
Should we hire in-house instead?
Eventually, usually. Agencies are good at starting a programme and at bringing testing discipline; the knowledge is worth retaining once the cadence is established.
What is the clearest warning sign?
A proposal with no baseline in it. If nobody has written down where you are starting, nobody can tell you where you got to.