Why the app store changes the calculator's inputs
On the default page, a click lands on a page the advertiser controls and converts there. Install campaigns put a third party in the middle. The ad sends people to a store listing whose layout, ratings, screenshots and review count shape the decision, and the install happens inside the store. The conversion rate the calculator needs is therefore the share of store page visits from the ads that end in an install, not a website conversion rate and not the store's overall rate across all visitors.
The outcome counted is also earlier than it looks. An install is a download; it is not a registered user, still less a paying one. Many installs are opened once, some never. Reading the calculator's CPACost of each conversion. as the cost of a user overstates what the budget bought, and reading it as the cost of a customer overstates it further. The guide to install campaigns explains why many advertisers move to optimising for in-app events once install volumes allow.
Measurement adds a final complication. Some platforms report installs late, in aggregate or partly by model, because of privacy frameworks on mobile operating systems. The install count behind the conversion rate may therefore be incomplete, especially for recent days.
The inputs for app install campaigns
| Calculator input | On the default page | For app install campaigns | Where to find it |
|---|---|---|---|
| CPM | Cost per thousand impressions | CPM of the install campaign by platform and placement | Platform CPM |
| CTR | Click-through rate | Click-through rate from the ad to the app store page | Platform CTR |
| Conversion rate | Conversion rate | Click-to-install rate: installs ÷ store page visits from the ads | Store console or attribution partner |
| Budget | Budget | Install campaign budget | Campaign budget |
The conversion-rate row is where this use case differs. Take installs attributed to the campaign from the attribution partner or the store console, and divide by the store page visits that came from the same ads in the same period. Where the platform reports only installs and clicks, use clicks as the denominator and say so. Do not borrow the store's overall rate, which blends in search and browse visitors who arrived already looking for an app like yours. The guide to app attribution and SKAdNetwork explains where install counts come from and why they differ between systems.
Entering the install campaign inputs step by step
Separate the campaign by platform and placement
Install costs and store page behaviour vary between networks and placements. Enter one platform, or one placement group, at a time rather than an account-wide average.
Read CPM and click-through rate to the store
Use the campaign's CPM and the rate at which impressions became clicks to the store listing. Exclude clicks on other parts of the ad that do not open the store.
Measure the store page install rate from paid visits
Divide attributed installs by store page visits from the ads for the same dates, using one reporting source throughout. Leave out the most recent days if installs are still being reported late.
Enter the inputs and add the activation step
Type the four figures into the calculator. Divide the CPA it returns by the share of installers still active a week after installing, and record both cost per install and cost per active user.
A worked install campaign with illustrative inputs
An invented language app checks what an active user costs
Lingomoor, an invented language-learning app, runs an install campaign with a CPM of $7. Its ads earn a 1.2% click-through rate to the app store page, and 18% of store page visits from the ads end in an install. On a $12,000 budget, the calculator returns a CPA of $3.24, about 3,700 expected conversions and an implied CPC of about $0.58. Those conversions are installs. Lingomoor's own data shows that 35% of installers are still using the app on day seven, so roughly 1,300 active users came from the budget and the cost per active user is $3.24 ÷ 35% ≈ $9.26.
CPA Calculator
Open the CPA Calculator →Prefilled with the example's inputs: CPM $7 · CTR 1.2% · Conversion rate 18% · Budget $12,000 → CPA $3.24 · Expected conversions 3,703 · Implied CPC $0.58.
Lingomoor and its numbers are invented. Raise the conversion rate in the embedded calculator a little, as a better store page might, and compare the effect with a similar improvement in click-through rate.
Interpreting cost per install for app campaigns
Read the CPA output as cost per install, then divide it by the day-seven active rate to get cost per active user. The second figure is closer to what the business wants to buy, and it lets campaigns with very different install quality be compared fairly. A cheap install that is deleted at once costs more per active user than a dearer one that sticks.
Read expected conversions as installs, before any opens, registrations or purchases. The figure is a ceiling on what the budget can produce further down the funnel, not a forecast of users.
Next, treat the store page as a lever in its own right. Because it sits inside the conversion-rate input, a better listing moves cost per install as much as a cheaper CPMCost per 1,000 impressions. or a stronger CTRShare of impressions that result in a click.. The guide to app store optimisation covers the elements that persuade store visitors, and the calculator's decomposition shows how much a given improvement in the install rate, or CVRShare of visits or clicks that convert., is worth against the other two levers.
Finally, allow for attribution limits. On platforms where installs are reported late, in aggregate or by model, the counted installs can fall short of the real ones, so cost per install may look worse than it is in recent data and better once modelled installs are added. Compare like periods from the same reporting source before concluding that a campaign has changed.
The calculator also cannot see what happens after the first week. Two campaigns with the same cost per active user can differ sharply in how many of those users subscribe, buy or are still around a month later. Once the app has enough history, carry the same division further down the funnel, to cost per registered or paying user, and let that figure decide where the budget goes.
Pitfalls when install numbers go into the calculator
- Treating installs as customers. An install is a download. Budgets set on cost per install alone tend to favour sources whose users never return.
- Using a conversion rate from organic store traffic. Organic visitors often arrive already searching for the app, so their install rate is higher than paid visitors' and makes cost per install look lower.
- Ignoring modelled or delayed install reporting. Recent days can be missing installs that arrive later, which makes fresh campaigns look worse than they are.
- Optimising cost per install where low-intent placements inflate installs. Some placements produce cheap installs from accidental taps or incentives. They lower cost per install and raise cost per active user.
- Mixing counts from different systems. Installs from an attribution partner divided by clicks from the ad platform can produce a rate that neither system would recognise.
