Mobile game UA reports are full of short acronyms, and it is easy to stare at a dashboard without knowing which number should drive a decision. This is a plain guide to the main mobile game UA metrics: what each one measures, what it hides, and how they fit together. We also cover the mobile game user acquisition trends that change how you should read them in 2026.

The short version is that no single metric is the answer. Cost and quality metrics each tell part of the story, and the decision comes from reading them in a chain.

The chain of UA metrics

Think of a campaign as a funnel with a price at each step:

  1. An ad is shown. This is an impression.
  2. Some viewers install. The install rate of that step is captured by IPM.
  3. You pay for those installs. That is CPI.
  4. Installed players come back and play. That is retention.
  5. Players earn you money through ads and purchases. That is LTV.
  6. Compare what they earn with what you paid. That is ROAS and payback.

If you read only one step, you can be fooled. A cheap install that never returns is worse than an expensive one that pays back. A creative with a high install rate can attract the wrong audience. So each metric below comes with a warning label.

CPI: cost per install

CPI is what you pay, on average, for one install. It is the most quoted metric in the industry and the easiest to understand, which is why it gets misused.

CPI is useful for budgeting and for comparing like with like, such as the same game, same country and same platform over time. It is much less useful for comparing different channels or different geos, because players from those sources can be worth very different amounts.

What it hides: quality. A low CPI only matters if the players are worth more than they cost. Cost per install also varies a great deal by genre, country and platform, and it tends to be highest for mid-core and RPG titles. Treat any CPI figure from another game as context, not as a target. The drivers are genre, geo, platform and creative.

IPM: installs per mille

IPM is the number of installs per thousand impressions. It measures how well an ad turns attention into installs, so it is mainly a creative metric.

A rising IPM tells you a creative is persuading more viewers. Because many networks price on a mix of impressions and bids, a higher IPM often leads to a lower CPI and more delivery for that creative.

What it hides: intent. An ad that over-promises can drive a high IPM and then disappoint at day 1. Always pair IPM with early retention. Our creative testing framework shows how to read IPM alongside quality signals when you decide which creatives to keep.

Retention: do players come back?

Retention is the share of an install cohort that returns on a given day, such as day 1, day 7 or day 30. It is the first quality check on any campaign.

Compare retention by channel, geo, creative and campaign, because a cohort that churns early is unlikely to earn enough later. Retention is also a leading indicator. It arrives days before revenue data is reliable, so it lets you catch a bad source early.

What it hides: monetisation. A game can retain players who never spend or watch ads. That is why retention sits between cost and revenue, not at the end.

LTV: what a player is worth

LTV, or lifetime value, is the total revenue a player brings over their life in the game. In practice you work with a window, such as 7, 30 or 180 days, and a predicted value beyond what you have observed.

For hybrid games, LTV combines ad revenue and in-app purchase revenue, and each part is built from cohorts. We walk through the calculation in our post on LTV for mobile games.

What it hides: uncertainty. Early LTV is a forecast, and forecasts can be wrong. Keep the assumptions visible, compare predictions with actual cohorts as they mature, and avoid committing large budgets to a single projection.

ROAS: return on ad spend

ROAS is revenue from a cohort divided by what you spent to acquire it, usually quoted at a point in time such as day 7 or day 30. A ROAS of 100 percent at a given day means the cohort has earned back its cost by then.

ROAS is the metric that links spend to outcome, and it is what most networks now let you optimise toward. The question is always: ROAS at which day, and is that enough? Our post on what is a good ROAS for mobile games covers targets in more detail.

What it hides: timing. A strong day 7 ROAS may flatten out, and a weak day 7 ROAS may keep climbing. Read ROAS as a curve, not a single point.

Payback: when do you get your money back?

Payback, or breakeven time, is the number of days until a cohort’s cumulative revenue equals what you spent on it. It is ROAS expressed as a date.

Payback is the metric a studio can plan its cash around. A short payback lets you recycle money into new spend quickly. A long payback ties up cash and raises the risk that the forecast is wrong. Choose a payback window that matches your runway, then let it set how much you will pay for players.

What it hides: scale. A campaign can pay back well at small spend and fall apart as you push volume. Recheck payback every time you raise budgets.

How the metrics fit together

Here is a worked way to read a campaign without any figures from a particular game:

  • IPM is low. The creative is not persuading. Test new hooks before touching budgets.
  • IPM is high but retention is weak. The creative may be misleading or the audience is wrong. Check what the ad promises against what the game delivers.
  • CPI is high but retention and ROAS are strong. The traffic is expensive and worth it. Hold or raise spend if payback is inside your window.
  • CPI is low but ROAS is weak. You are buying cheap players who do not monetise. Shift towards a return-based goal or a different source.
  • ROAS is on track but payback is long. Fine if you can fund it. If not, cut spend or improve early monetisation.

Notice that every case ends with an action. A metric that does not change a decision is only decoration.

Several shifts in 2026 affect how much you can trust each number.

Costs are under pressure. Industry commentary describes acquisition costs as rising for structural reasons, including less identifiable data and a small number of networks holding most of the inventory. That is one more reason to judge CPI against expected revenue, not against last year’s number.

Return-based optimisation is the norm. Networks increasingly let you bid toward return on ad spend. That makes ROAS and predicted LTV central to how campaigns are steered, and it raises the importance of sending clean revenue events.

Creatives wear out faster. A creative that drove a strong IPM last month may decay quickly. Track IPM by creative age, and see our post on creative fatigue for the warning signs.

Measurement is more aggregated. Privacy changes mean some data arrives late, in groups, or modelled. Our post on UA after IDFA explains what that means for reporting. In practice, lean on cohorts and trends over a few days instead of single-day swings.

A simple weekly routine

You do not need an elaborate dashboard. A short weekly review can cover most of what matters:

  1. Check IPM and CPI by creative to find winners and tired ads.
  2. Check day 1 and day 7 retention by channel and geo.
  3. Check ROAS and projected payback by campaign.
  4. Decide which campaigns to raise, hold or pause, and write down why.
  5. Compare last week’s predictions with what actually happened.

Over a few weeks, that last step makes your forecasts more honest.

Summary

CPI tells you what you paid, IPM tells you how persuasive the ad is, retention tells you whether players stay, LTV tells you what they are worth, and ROAS and payback tell you whether the whole thing pays. Read them as a chain, keep each metric’s blind spot in mind, and let payback decide how much you are willing to pay.