App marketing is one of the few advertising markets where the headline number can be produced almost at will, which makes cost per install a poor basis for choosing a supplier. What matters is how many of those people open the app a week later and do the thing your business depends on. A company that cannot discuss retention cohorts, post-install events and the limits of mobile attribution is selling volume rather than growth.
Optimise toward the event that represents value
Networks find whatever you tell them to find, so an account reporting installs will deliver installers. Define the event that means something, a completed booking, a subscription, a second session, and send it back through a measurement partner. This is usually the highest-value work in the first quarter and it is frequently outside a media proposal's scope.
The store listing is the conversion step nobody owns
Every advertisement sends people to a store page, and the icon, screenshots, video, title and reviews decide whether they install. Improving it lifts the return on all media at once. Establish whether the supplier does store optimisation, who writes the listing, and how variants are tested, because a poor listing wastes every dollar of media behind it.
Understand what attribution can see now
Privacy frameworks limit user-level measurement, so some performance arrives aggregated, delayed or modelled. That is the environment, not a supplier failure. What you should expect is an explanation of which measurement they trust for which decision, and the use of holdouts and incrementality tests rather than treating every reported conversion as caused.
Fraud checks belong in the agreement
Click injection, click spamming and bot installs inflate exactly the numbers the supplier is judged on. Ask what anti-fraud tooling is used, who pays for it, how disputes with networks are handled, and whether you see rejected volumes. A company that has never disputed a network invoice has either been lucky or has not looked.
Questions people ask about mobile app marketing companies
What should we pay per install?
It is the wrong target. Work out what a retained active user is worth and what proportion of installs becomes one, then derive an acceptable acquisition cost. Two campaigns with the same install cost can differ several-fold in value delivered.
Do we need a mobile measurement partner?
For meaningful spend, yes. It is the neutral record between you and the networks and makes attribution and fraud disputes tractable. Own the account and the contract yourself rather than having the agency hold it.
Is store optimisation a separate service?
It is sometimes sold separately and should at least be coordinated, because it converts the traffic the media buys. Whoever owns it should test listing elements systematically and report conversion from impression to install alongside media performance.
How long before app marketing can be judged?
Long enough to see a retention cohort mature, which for most apps means at least a month after a spend period, plus time for campaigns to exit their learning behaviour. Judging on week-one installs measures how fast the money was spent.