Marketing an app is a different trade from marketing a website, and hiring the wrong kind of specialist is the most common expensive mistake in the category. The traffic does not land on pages you control, the store listing is the conversion step, attribution works through a measurement partner rather than a pixel, and the whole economics depend on what a user does in week two rather than on the install. An app marketing agency should be fluent in all four of those. This guide sets out what the work covers, what moves the price, and the questions that separate an operator from a media buyer with an app deck.
The three jobs, and which one you are short of
App marketing separates into store optimisation, paid user acquisition and retention, and most disappointing engagements happen because the client bought the wrong one. Store optimisation is the work on the listing itself: the title and subtitle, the keyword field, the screenshots and preview video, the description, ratings and reviews. It is cheap relative to media and it multiplies everything else, because it lifts the conversion rate from store visit to install for traffic you are already paying for. Apple documents how App Store search works and Google publishes guidance on getting discovered in Play search, and any agency you hire should be able to discuss both without hand-waving. Paid acquisition is the media layer across search ads on both stores, social platforms and programmatic networks, and it is where the budget goes. Retention and lifecycle work covers onboarding, notifications, messaging and the product changes that decide whether an install becomes a user. Ask a candidate which of the three they are strongest at and check whether it matches the thing that is actually broken.
Attribution is the technical gate
Nothing about app marketing can be judged without a working measurement setup, and this is where inexperienced agencies are exposed fastest. Installs are attributed through a mobile measurement partner rather than through the ad platforms alone, in-app events have to be defined and instrumented by your engineers before any campaign can optimise toward them, and privacy frameworks on both platforms mean a share of your data arrives aggregated, delayed or not at all. The practical consequence is that platform-reported installs will not reconcile with store console numbers or with your measurement partner, and an agency that cannot explain why is going to report the friendliest of the three every month. Before signing anyone, establish which measurement partner you use, who owns that account, which events are instrumented, and what the agreed source of truth is for cost per install and cost per activated user. Put the answer in the contract, because arguing about it in month four is how these relationships end.
What moves the price
Fees follow media spend, channel count and the amount of creative production included. Most agencies charge either a monthly retainer, a share of managed media spend, or a hybrid with a floor, and the share model deserves scrutiny in a category where scaling spend is easy and scaling efficiently is not. Creative volume is the largest hidden line: app advertising is won on the number of concepts tested rather than on bidding, so a programme producing a steady stream of new video and playable assets each month costs materially more than one recycling a handful. Store optimisation is sometimes a separate project fee and sometimes folded in. Two things should always be separate from the fee: the media budget itself, and any measurement partner or creative production costs billed at pass-through. Ask for a staffing plan and for the monthly creative output in units, because those two numbers explain almost every difference between quotes.
Judging performance without fooling yourself
Cost per install is the metric that flatters everyone and predicts nothing. Cheap installs are easy to buy from low-quality inventory and they churn within a week. Insist that reporting is built on downstream events instead: activation, day seven and day thirty retention, and revenue or subscription conversion by cohort and by channel. Ask for cohort tables rather than monthly averages, because monthly averages hide the fact that a good cohort and a bad one arrived in the same period. Agree a payback window that matches your business and judge channels against it. And check incrementality where you can, particularly on branded search and retargeting, since both channels are excellent at claiming credit for users who would have arrived anyway. An agency that raises incrementality before you do is one that intends to be judged honestly.
Questions people ask about app marketing agencies
Should we hire an app specialist or our existing agency?
Almost always a specialist, unless your existing agency can demonstrate live app accounts, a working measurement partner setup and store optimisation work they have actually shipped. The attribution and store mechanics are specific enough that a strong web performance team will spend your first quarter learning them.
How much of the budget should go to store optimisation?
A small fraction of what you spend on media, and it should come first. Improving the conversion rate from store visit to install lifts the return on every campaign you subsequently run, so doing paid acquisition into a weak listing is paying full price for traffic you then discard at the door.
Is a share-of-spend fee reasonable?
It is common, and workable if the contract also names a target efficiency and a downstream measure such as cost per activated user. Without those, the model pays your agency more for spending more, which is only aligned with you while the economics hold. A hybrid with a floor and an efficiency target is usually the fairer structure.
How long before we can judge the work?
Store optimisation changes can show within weeks. Paid acquisition needs enough spend to exit learning on each channel plus a full retention window, so realistically one to two quarters before cohort economics are readable. Agree in advance what evidence would justify continuing, pausing or reallocating.