The numbers looked fine. CPI was under target. Installs came in. The reward tasks got completed.
But then nothing happened.
Retention nor revenue hit the target. The cohort LTV drew almost a perfect horizontal at D30.
The honest answer for this, in most cases, is not the network. It is not the creative. And it is not the audience — at least not entirely.
It is the event itself.
Here are the five mistakes that show up most consistently in reward campaigns that fail to convert — and the thinking behind why they happen.
Contents
Mistake #1: The Task Was Too Easy
The most common mistake is designing a task that anyone can complete in 90 seconds.
“Complete the tutorial.” “Log in once.” These tasks exist because they are easy to verify and generate clean CPE data. They also generate installs that churn immediately.
When the task requires no actual engagement with the game’s core loop, the user who completes it has no relationship with the product. They came for the incentive, and once they received what they came for, they left.
The error here is optimizing for task completion rate rather than for user quality. A 95% completion rate on a trivial task is not a sign of campaign success. It is a sign that you paid for a behavior that required only a little action.
A more useful question: does completing this task require the user to have understood why this game is worth playing?
Mistake #2: The Task Was Too Hard — in the Wrong Way
The opposite problem also exists, and it is less obvious.
Some campaigns set tasks that are technically achievable but require several hours of daily play from a cold-start user within a 7-day window. On paper, this filters for quality. In practice, it filters out almost everyone, including users who would have been good long-term players if given slightly more time.
The issue is not difficulty itself. Difficulty is useful — it creates investment. The issue is task design that does not match the actual onboarding curve of the game.
If the game’s natural progression puts users at Stage 10 by Day 5 organically, then “Reach Stage 10” is a reasonable CPE target. If organic users typically reach it at Day 14, then a 7-day CPE window is selecting for users who rushed through content, which is not a proxy for high LTV.
- Audit your organic D7 milestones before setting CPE targets
- Set the task 1–2 steps behind where organic payers typically convert
- If the window is tight, reduce the task difficulty — do not extend the window
Mistake #3: The MMP Was Misconfigured
This one is uncomfortable to admit, but it happens more often than the industry acknowledges.
MMP misconfiguration in reward campaigns tends to fall into a few categories:
- Postback windows set incorrectly for the reward network’s attribution model
- Re-engagement and re-attribution events firing when they should not, inflating CPE counts
- In-app event schema not matching what the reward publisher is reading
- Device-level fraud filters rejecting legitimate completions from high-risk regions
The result is usually one of two failure modes: you pay for completions that never happened, or genuine completions go unverified and users do not receive their reward — creating churn before the relationship even starts.
Before launching a reward campaign on any new network, it is worth doing a small test cohort of 200–300 installs specifically to verify that the full event tracking chain is working correctly end to end.
Mistake #4: The Audience Did Not Match the Game
Reward campaigns often reach narrow audiences — often much narrower than paid social or search. That reach is the point. It is also the risk.
A common mistake is running a reward campaign for a mid-core RPG on an offerwall that primarily serves casual users completing simple daily tasks. The task gets done. The game does not retain anyone.
This is a targeting mismatch, and it is often invisible in aggregate data. CPE looks fine. Installs look fine. Only when you segment the cohort by source and look at D14 behavior does the problem become visible.
Some questions worth asking before launch:
- What is the typical user profile on this reward network? Casual? Mid-core? High-intent?
- Does the reward incentive attract users who want the reward, or users who might genuinely want the game?
- Is this network known for delivering quality on this genre, or just volume?
Not every reward network is wrong for every game. But matching the network’s user base to the game’s retention profile is a prerequisite — not an afterthought.
Mistake #5: There Was No Bridge to Monetization
This is the structural failure that underlies most of the others.
Even when the task is well-designed and the audience is right, reward campaigns frequently fail because the event ends at task completion. The user finishes the required steps, collects the incentive, and exits the engagement loop — because there is no next step.
Reward UA works best when the event sequence is designed with a monetization pathway in mind, not bolted on afterward. That means:
- The final task should bring the user to a natural decision point — a place in the game where spending feels optional but compelling
- The event reward itself should ideally give the user partial progress toward something they now want to complete
- The timing of the first IAP offer should be tied to event progression, not to a fixed day counter
The difference between a reward campaign that generates installs and one that generates revenue is almost always this: whether the event was designed to end, or designed to transition.
Summary
None of these mistakes are rare. Most UA teams will recognize at least two or three of them in campaigns they have run.
The pattern tends to repeat because reward campaigns are still often treated as a volume play — a way to hit install targets quickly. The framing makes sense for some objectives. It does not make sense when the objective is revenue.
If the goal is more than simple installs, the event needs to be designed for users who are likely to generate revenue from the start: tasks that filter for investment, tracking that actually works, audiences that match the product, and a structure that creates a reason to stay after the reward is claimed.
- Too-easy tasks optimize for completion rate, not user quality
- Too-hard tasks filter out the wrong users when not calibrated to organic progression
- MMP misconfiguration wastes budget silently and often goes undetected
- Audience mismatch is invisible in aggregate metrics but obvious at cohort level
- Without a monetization bridge, even well-designed events produce installs, not revenue
