Engagement is what happens between opening the app and putting it down. Publishers usually measure it with a number that counts openings and calls it a session, then wonder why the retention curve looks like theirs does. Honest engagement measurement starts somewhere less flattering and ends somewhere more useful: it measures what the person did, what they got for it, and whether they came back.
The mechanics that actually build a habit
Three mechanics do most of the work in mobile engagement, and they are worth separating because they fail in different ways.
A streak binds the product to a calendar. It is the strongest habit mechanic because it creates a reason to open the app tomorrow that has nothing to do with what is in the app tomorrow. It is also the most brittle, because a broken streak is punishment, and punishment is only motivating when the person is already motivated.
A progress bar binds the product to a task. It is the most reliable mechanic for a new user, because it answers the question every new user has - what do I do first - and it makes the first session end with something completed rather than something open.
A daily set of varied actions binds the product to a habit of returning without binding it to a single task. This is the mechanic a rewarded wall naturally produces, and it is the reason a task-based app can be opened daily without the app ever having to be interesting on its own.
The mistake is using one mechanic to carry the whole retention plan. A streak without progress will lose the people who have not yet built the habit. Progress without a streak will lose the people who have not yet built a reason to come back tomorrow.
Rewarded ads and rewarded offers are not the same thing
Both hand the person something in exchange for attention, and the difference is who is on the other side of the trade.
A rewarded ad is an ad. The person watches a video or looks at a brand's material and receives a reward for tolerating it. The demand comes from the brand's media budget. A rewarded offer is a task: the person installs an app, completes a survey or plays a level, and the reward is the point. The demand comes from the advertiser's acquisition budget, and the reward is the compensation for a real action rather than for attention.
The behavioural evidence for the ad format comes from Adjust's research on rewarded video, which is unusually consistent. Across the surveys Adjust cites, 75% of players see rewarded video as less disruptive than other formats, and 50% would be less satisfied with a game if rewarded video ads were removed. 71% of players prefer in-game video ads to in-app purchases, and 84% would adopt a new game advertised with tangible rewards. Completion rates for rewarded video run around 75.8% on the data Adjust publishes - high enough that the format works, which is not the same as saying the format suits every app.
Session length and what it means
Session length is the least reliable engagement metric, and it is the one most often used. It measures how long the app was in the foreground, which is a measurement of the session's length and not of its quality.
A person who opens a rewards app for four minutes to complete a survey is more valuable than a person who idles in the app for forty minutes watching a video, even though the numbers look worse on the dashboard. And a person who opens the app, completes nothing and leaves has a session of thirty seconds, which looks like churn even if they will come back tomorrow for a mission.
Sensor Tower's State of Mobile 2025 puts the scale of the attention being fought over: 4.2 trillion hours spent in apps over the past year, consumer spend reaching 150 billion dollars for the first time, and gaming spend climbing to 81 billion dollars, up 4% year on year. There has never been more app time to compete for, which is precisely why an engagement number that does not describe a behaviour is worth nothing.
Measuring engagement without fooling yourself
Most engagement dashboards have three common problems, and all three make a publisher look better than they are.
The first is counting the session rather than the outcome. A session that ends with nothing done is a session, and it inflates every rate built on top of it.
The second is averaging away the users who leave. A change that lifts the median session while the bottom half of the cohort churns faster is not an improvement, and the aggregate number will not show it.
The third is mistaking rewarded activity for engagement. When the task wall is the product, completion rates measure how good the offers are, not whether the app is worth opening. A publisher can raise completion from 20% to 40% by making the tasks easier and watch their engagement metric double while their retention falls, because the easy tasks stopped being worth anything to the person.
The discipline that fixes all three is to measure a chain rather than a rate: did the person do something, did they get something, did they come back and do something more demanding. The middle link is where most of the fraud shows up too, because a completion that was never verified inflates the reward side of the chain while the revenue side never arrives.
What to do next
Pick one mechanic and tune it properly before adding a second. The most common engagement failure is a feature list where nothing was tuned.
Make the first session finish something. A person who completes a task on their first visit is a different population from one who browses, and the difference shows up in every metric afterwards.
Measure the chain - action, reward, return - and instrument the credit step, because that is where both the honest measurement and the fraud live.
And when the numbers say the wall is competing with the app, shrink the wall. The publisher who keeps the experience intact is the one who still has an app to measure in five years.
Related reading
Run rewards in your app
Tell us the platform and roughly how many daily active users you have, and we will tell you whether rewarded games, surveys or both are the better fit.
Questions
What is the difference between a rewarded ad and a rewarded offer?
A rewarded ad pays a person for tolerating advertising, and the money comes from a brand's media budget. A rewarded offer pays a person for completing a task - installing an app, answering a survey, finishing a level - and the money comes from an advertiser's acquisition budget. The ad format is well suited to games with a natural pause; the offer format is well suited to an audience with task-shaped intent.
Is session length a good engagement metric?
Not on its own. It measures how long the phone was in the foreground, which is a property of the session rather than a description of what the person did. Completion rate, the share of sessions ending in a completed action, reward latency and repeat-action rate all describe behaviour, and they can move in the opposite direction to session length.
How do I know my engagement numbers are honest?
Measure a chain rather than a rate: did the person do something, did they get something, and did they come back and do something more demanding. That structure exposes the two usual distortions - counting sessions that end in nothing, and averaging away the cohort that churns - and it also surfaces unverified completions, which inflate the reward side of the chain while the revenue never arrives.
How do rewarded ads perform compared with other ad formats?
Adjust's research reports that 75% of players see rewarded video as less disruptive than other formats, 50% would be less satisfied with a game without it, 71% prefer in-game video ads to in-app purchases and 84% would adopt a new game advertised with tangible rewards. Completion for rewarded video runs around 75.8% on the data Adjust publishes. The format works, but it suits games where watching something is a natural pause rather than an interruption.