The classic growth agency retainer sold capacity: a number of hours a month, spent on campaigns, reported in a monthly deck. When the contract ended, the client kept the deck.
That model is dying, and not because agencies got worse. It is dying because the work changed shape. The valuable part of growth is no longer executing campaigns — it is building the machine that executes them.
What clients noticed
Three things, roughly in this order.
First, the deliverable evaporated. A campaign run in the agency's tools, on the agency's data, with the agency's templates, leaves nothing behind. Second, the same work was being redone monthly because none of it accumulated. Third, and most damaging: the agency's incentive was to remain necessary.
What replaced it
The engagement that works now looks more like an infrastructure build than a marketing retainer:
- The audience definition, sequences, ad audiences, data pipelines and reporting live in the client's accounts, under their billing, from day one.
- The engagement is structured in phases with an explicit end state — build, operate, hand over — rather than an open-ended monthly fee.
- Success is measured on blended cost-per-opportunity and system uptime, not on activity volume.
- Documentation and runbooks are deliverables, not favours.
A campaign run in the agency's tools leaves nothing behind. A system leaves an asset.
The obvious objection
If you build a system the client owns and can run without you, why would they keep paying?
Because most do, and for a better reason. Once a system exists, the interesting work is extending it — a new segment, a new channel, a new market, a new signal. That is higher-value work than campaign execution and it is genuinely optional, which means it gets renewed on merit rather than on inertia.
The engagements that end are the ones that should end. That is a feature. It is also the only honest way to sell a system: if you have built it properly, the client could walk away, and the fact that they do not is the actual evidence it works.
How to buy it
If you are evaluating a growth partner, three questions sort them quickly. Whose accounts will this live in? What exactly do we own if we stop working together in six months? And what is the number you want to be measured on?
A retainer shop will find all three uncomfortable. That is the point of asking.