In one line
Revenue ÷ total users. That one line is all ARPU (Average Revenue Per User) is: what an average user generated over a period.
Do not confuse it with ARPPU
The names are close but the denominators differ.
- ARPU = revenue ÷ all users
- ARPPU = revenue ÷ paying users
In an app with a 3% payer rate, ARPPU lands around 30× ARPU. The two lead to completely different conclusions, so check which one a report is actually showing.
They answer different questions: ARPU is the profitability of the whole user pool, ARPPU is how much payers spend.
A rise is not automatically good
ARPU is a ratio, so it also rises when the denominator falls. When new-user volume drops, the remaining base skews toward heavy users and ARPU climbs on its own.
Reading ARPU alone and declaring improved monetisation is therefore risky. Read user count and revenue together, and split by cohort where you can.
How it feeds LTV
LTV ultimately comes from cumulative ARPU curves per cohort: stack cumulative ARPU at D7, D30 and D90 after install and see where the curve converges.
Watch the observation window. Recent cohorts have immature curves, and extending them in that state produces badly wrong LTV.
Go deeper
The relationship between LTV and CAC is covered in LTV:CAC; reading cumulative cohort curves is covered in cohort analysis.