Written by Krisztián StarkKey takeaways
- ARPU is total revenue divided by users in a period; always state the period and which users you counted.
- ARPDAU uses daily active users and moves fast; ARPPU counts paying users only; purchase ARPU equals ARPPU times the share of users who pay.
- Cohort ARPU after 30 or 90 days, divided by your target ROAS, is the most you can pay per install.
What does ARPU mean?
ARPU stands for average revenue per user. It is the total revenue from a product in a period divided by the number of users in that same period. If an app earns $3,000 in a month from 20,000 monthly active users, its ARPU for that month is $0.15.
ARPU is one of the first numbers app developers, game studios, and website owners look at, because it turns revenue into a price per person. That makes it easy to compare with what one user costs to acquire, and to compare months, countries, and monetization methods on the same scale. It is also in our advertising glossary.
The ARPU formula
ARPU = total revenue in the period / users in the periodThe formula is simple; the definitions behind it are where numbers go wrong. Decide two things before you calculate:
- Which revenue. Purchases, subscriptions, ad revenue, offerwall earnings, or all of them. Many games report all sources together, while a publisher may want ARPU for one format only.
- Which users. Monthly active users, all installs, or everyone who ever signed up. Dividing by registered accounts instead of active users makes ARPU look smaller; dividing by a flattering definition of “active” makes it look bigger.
GameAnalytics makes the same point in its guide to these metrics: be honest about what counts as an active user, or the number stops meaning anything.
ARPU vs ARPDAU vs ARPPU
| Metric | Formula | What it tells you |
|---|---|---|
| ARPU | Revenue / users in the period | Average value of a user over a week, month, or quarter |
| ARPDAU | Revenue in a day / daily active users that day | A fast daily read on monetization, useful for spotting changes |
| ARPPU | Revenue / paying users | How much the users who pay spend, on average |
These definitions match the ones used by analytics firms such as GameAnalytics and AppsFlyer. The three are linked by one identity that is worth remembering, counting purchase revenue only:
Purchase ARPU = ARPPU x share of users who payThat identity shows the two ways to raise purchase ARPU: get more users to pay anything at all, or get paying users to spend more. They usually need very different work, so it helps to know which one moved.
Illustrative month: a word game with 50,000 players
A word game has 50,000 monthly active users. It earns $2,500 from rewarded video and $4,000 from purchases made by 800 paying players. ARPU is $6,500 / 50,000 = $0.13. ARPPU on purchases is $4,000 / 800 = $5.00, and the paying share is 1.6%. If a better starter pack lifts the paying share to 2% at the same ARPPU, purchase revenue rises to $5,000 and ARPU to $0.15. The game is made up; the arithmetic is what you reuse.
ARPU vs LTV
ARPU covers a fixed period with a clear start and end. LTV, or lifetime value, covers everything a user brings from the first visit until they leave. AppsFlyer’s glossary draws the line the same way. The two connect through cohorts: take all users who installed in the same week and track their total revenue per install after 7, 30, and 90 days. That cumulative, per-install figure is cohort ARPU, and over a long enough window it becomes your working estimate of LTV. CAC vs CPA vs LTV covers the lifetime side in depth.
How advertisers use ARPU to price users
Cohort ARPU is the bridge between what a user earns you and what you can pay to get one:
Maximum cost per install = cohort ARPU after N days / target ROAS
ROAS after N days = cohort ARPU after N days / cost per installFrom a $0.42 day 30 ARPU to a price per install
Users who installed in March earned you $0.42 each by day 30. You want installs to pay back 1.2 times within that month, so the most you can pay is $0.42 / 1.2 = $0.35 per install. If a channel sells installs at $0.28, its day 30 ROAS is $0.42 / $0.28 = 1.5, and it is worth scaling. Sample numbers, chosen to keep the division easy.
Calculate ARPU per country as well as overall, because it varies as much as ad prices do. Cost per install by country shows published install prices to compare against, and What is ROAS? explains how to pick a target. The ROAS calculator runs the numbers for a live campaign.
How publishers use ARPU and ARPDAU
For a website, app, or community that earns from ads and offers, ARPU is the fairest way to compare formats that pay in different ways. A banner pays per impression, rewarded video per view, and an offerwall per completed offer, so eCPM alone mixes up very different things. Revenue per user, or per 1,000 active users, puts them on one scale. eCPM vs RPM and eCPM and EPC explained cover the per-impression and per-click side.
ARPU also turns a monetization idea into a revenue estimate before you build it. The offerwall earnings calculator shows revenue per 1,000 active users from your traffic, the share who open the wall, and average payouts.
One caution: offerwall revenue arrives in bursts, and some results are confirmed days later or reversed. Judge offerwall ARPU over a week or a month, not one day, and use credited earnings rather than pending ones.
Common ARPU mistakes
- Comparing ARPU figures that use different user definitions, such as one per active user and one per registered account.
- Mixing periods: a monthly ARPU is not comparable with a daily ARPDAU multiplied by 30, because different people are active on different days.
- Reading a rising ARPU as good news when it rose because casual users left and only heavy spenders remained.
- Using gross revenue when app stores or partners take a share first. Use the revenue you actually receive.
- Setting bids from an overall ARPU when most revenue comes from one or two countries.
Working out offerwall ARPU from your Moonetix app page
Each app in the Moonetix publisher dashboard shows four numbers for the last 30 days: Active users, the users who opened that offerwall; Monetized users, the users with at least one credited result; Clicks; and Earnings, your credited earnings. From these you can calculate two kinds of ARPU yourself:
- Revenue per wall user = Earnings ÷ Active users. It shows how well the wall earns from the people who open it.
- Revenue per paying user = Earnings ÷ Monetized users. This is the offerwall version of ARPPU.
Revenue per wall user will be higher than an ARPU calculated over everyone who used your app or site that month, because not everyone opens the wall. Both are useful: the first tells you how good the wall is, the second how much it adds per user of your whole product. To see which offers the money comes from, open Reports, Conversions.
Estimating offerwall ARPU before your publisher application
Run your monthly active users through the offerwall earnings calculator to get revenue per 1,000 users, then compare it with what your current ads earn per 1,000 users. That comparison tells you whether a wall is worth a slot. Publisher applications to Moonetix are open; see how the Moonetix offerwall works to prepare.
Frequently asked questions
How do you calculate ARPU?
Divide total revenue in a period by the number of users in the same period. $3,000 of monthly revenue from 20,000 monthly active users is an ARPU of $0.15.
What is the difference between ARPU and ARPPU?
ARPU divides revenue by all users; ARPPU divides it by paying users only. Purchase ARPU equals ARPPU multiplied by the share of users who pay.
What is ARPDAU?
Average revenue per daily active user: revenue in one day divided by the users active that day. It is the daily version of ARPU and is common in mobile games.
What is a good ARPU?
There is no universal figure, because ARPU depends on the category, the countries, and the business model. A good ARPU is one that stays comfortably above what you pay to acquire a user.
Is ARPU the same as LTV?
No. ARPU covers a fixed period, while LTV covers a user’s whole lifetime. Cohort ARPU over a long window, such as 90 or 180 days, is often used as a practical estimate of LTV.
