Most app teams we meet can recite their tap-through rate from memory and have no idea what an installed user is worth in month three. That gap is where budget quietly disappears. The apple search ads metrics that appear first in the dashboard are the ones easiest to collect, not the ones that decide whether the channel deserves more money next quarter. Impressions, taps and installs describe the top of a funnel that ends somewhere else entirely, in a subscription renewal or an in-app purchase that happens weeks after the click that started it.
We run App Store campaigns for clients whose whole business is the app, so the reporting has to survive a conversation with someone who signs cheques. That means every number we put on a slide should connect, by an argument we can spell out, to revenue or retained users. Some metrics do that well. Some are useful for diagnosis but worthless as a target. And a few are actively misleading when read alone, because they improve for reasons that have nothing to do with the health of the account. Sorting them into those three groups is most of the work.
Start from app economics, not ecommerce habits
An ecommerce campaign can be judged in days because the purchase happens near the click. An app rarely works that way. Someone taps an ad, installs, opens the app once, hits an onboarding screen, and then either becomes a paying user or does not, often after a trial period. The distance between the ad spend and the money is measured in weeks. If you judge the channel on cost per install alone you are judging the first two minutes of a relationship that pays out over a year.
So the frame we use is install to paying user, then paying user to retained paying user. Cost per install is an input to that chain, not a result. A campaign with a higher cost per install that converts twice as many of those installs into subscribers is the better campaign, and the dashboard will not tell you that on its own. You need the in-app events flowing back, attributed to the campaign and keyword that produced them, before the numbers mean anything.
The apple search ads metrics worth reporting
Once attribution is in place, a short list does most of the explanatory work. We keep reporting narrow on purpose: a dashboard with forty tiles gets skimmed, a dashboard with six gets read. These are the ones we defend in client meetings, and each one exists because it answers a question someone actually asked.
- Cost per paying user, not cost per install, as the headline efficiency number for each campaign and each high volume keyword.
- Day 7 and day 30 retention, split by the keyword group that brought the user in, since intent at search time predicts whether the app gets opened again.
- Trial to paid conversion rate by campaign, which often varies more between keyword themes than the install cost does.
- Share of spend on branded versus discovery terms, so nobody mistakes demand you already had for demand you created.
- Contribution after store commission and refunds, because gross revenue flatters every calculation built on top of it.
Notice what is missing. Tap-through rate is on our diagnostic list, not our reporting list. It tells us whether creative and keyword match, which is genuinely useful when a campaign underperforms, but it moves for reasons like competitor absence that have nothing to do with our work. Reporting it as an achievement invites the wrong conversation.
Reading the numbers by search intent
The strongest lever in an App Store account is usually not bid strategy, it is separating intent. Someone searching your brand name has already decided. Someone searching a category term is comparing. Someone searching a competitor is curious or dissatisfied. Those three groups convert at different rates, retain at different rates, and deserve different bids and different creative. When they share a campaign, the averages hide all of it and the apple search ads metrics you report become a blend that nobody can act on.
We structure accounts so those intents never mix, then read performance per group over a window long enough to include the subscription decision. It is slower to draw conclusions this way. It is also the only way we have found to answer the question clients really care about, which is where the next thousand units of budget should go rather than how yesterday looked.
Attribution limits you should state out loud
Privacy changes on iOS mean some of what you would like to measure arrives aggregated, delayed, or not at all. Pretending otherwise produces reports that look precise and quietly are not. We prefer to name the uncertainty: which conversions are modelled, which windows are truncated, where a cohort is too small to say anything. A finance team will accept a number with stated limits. They will stop trusting you entirely the first time a confident figure turns out to have been guesswork.
How we work through this with clients
A typical engagement starts with measurement rather than bidding. We check that in-app events fire correctly, that campaign structure matches intent, and that somebody has agreed what a paying user is worth before we argue about cost per install. Only then do we touch budgets. It is less exciting than launching on day one, but it means that when we do scale spend, we are scaling something whose payback we can show rather than something that merely looks busy in a dashboard.
Keep reading: Apple Search Ads · Advertising