ROAS: Return on ad spend.

ROAS tells you whether user acquisition is paying for itself. A ROAS of 100% means the players from a campaign have earned back exactly what they cost. Below 100%, you are still in the red. Above it, the campaign is profitable on that revenue basis.

The ROAS formula

ROAS = revenue from the campaign's users ÷ campaign spend × 100%

In mobile games you almost always add a time window:

DN ROAS = cumulative revenue from the cohort up to day N ÷ spend on that cohort

A worked example

A studio spends $5,000 on a campaign in one week. The players it brings in generate:

Day since install	Cumulative revenue	ROAS
D7	$1,500	$1,500 ÷ $5,000 = 30%
D30	$3,500	$3,500 ÷ $5,000 = 70%
D90	$5,600	$5,600 ÷ $5,000 = 112%

The campaign breaks even somewhere between day 30 and day 90, and by day 90 it has returned $600 more than it cost. That point is the payback period.

Why early ROAS matters most

You cannot wait 90 days to decide whether to keep a campaign running. So teams learn, from past cohorts, how early ROAS relates to later ROAS, and set a target for day 7.

Example. In this game's past cohorts, D90 revenue has been about 3.7 times D7 revenue. To break even by day 90, a campaign needs D7 ROAS of at least 100% ÷ 3.7 ≈ 27%. The campaign above reached 30% on day 7, so it was on track, and the D90 result confirmed it (30% × 3.7 ≈ 111%, close to the 112% actually reached).

That multiplier comes from your own data and changes by game, country, platform and channel. Recheck it every month.

Gross vs net ROAS

Be clear about which revenue you use:

• Gross: what players paid, plus ad revenue.
• Net: after the app store's commission on purchases, and sometimes after taxes.

A campaign at 100% gross ROAS can still lose money once the store takes its share of purchase revenue. Many teams set targets on net revenue for that reason.

ROAS, LTV and CPI

ROAS and LTV describe the same thing from two angles:

ROAS on day N = LTV on day N ÷ cost per install

If D30 LTV is $1.40 and you paid $2.00 per install (eCPI), D30 ROAS = $1.40 ÷ $2.00 = 70%.

That makes the two levers obvious: earn more per player (higher LTV), or pay less per player (lower CPI, often through better creatives with a higher IPM).

Where ROAS data comes from

To calculate ROAS by campaign you need three things joined together: spend from each ad network, installs attributed to each campaign, and revenue from each player. An MMP usually does this for you, including ad revenue at the user level where networks report it. On iOS, much of the data arrives through SKAdNetwork in aggregated form, so iOS ROAS often relies on modelling.

Common ROAS mistakes

• Judging too early. D1 ROAS is noisy. Use the earliest day that reliably predicts your payback, often D7.
• Blending everything. One ROAS for all countries and channels hides the campaigns that lose money.
• Ignoring organic lift. Paid campaigns can also raise organic installs. Strict campaign ROAS does not count them, which can undervalue some campaigns.
• Forgetting ad revenue. In hybrid games, ad revenue can be a large share of early revenue.

See our ads and user acquisition service if you want help setting ROAS targets.