What five in-game ad experiments taught us about player behaviour and revenue

• Changing the interstitial order boosted ad revenue by 24%.
• Simple tweaks in UI drove a rise in rewarded video engagement.
• Across all experiments results showed that minor frontend changes can drive major backend shifts.

This article from Božo Janković was originally published as part of the GameBiz Consulting Ad Monetisation newsletter.

In our last GameBiz admon piece, we took a closer look at mediation platforms and ad networks, sharing several test cases that challenged familiar assumptions in ad monetisation. Those experiments revealed just how unpredictable small adjustments on the backend can be. 

Whether it involved shifting bidding strategies, tweaking waterfall instances, or running what appeared to be straightforward A/B tests, the findings showed that even established practices can produce outcomes that are anything but certain.

This month, we are shifting focus to another side of the story: the in-game environment and the player-facing adjustments that may seem simple at first glance, yet often lead to equally surprising and sometimes confusing results.

From interface changes to modifications in how ads are presented, these experiments illustrate how sensitive player behaviour can be, and why every decision deserves close examination.

1. How did adjusting rewarded and interstitial placement order deliver a 30% uplift?

One of the classic friction points between product teams and monetisation teams is the use of interstitial ads. For that reason, we considered it a real win when we convinced one of our clients to introduce interstitials for non-paying players who had been active in the game for more than a month.

In the initial setup, interstitials were shown after a match, but only once the player had been offered the option to watch a rewarded video ad that doubled their earned resources. Introducing interstitials resulted in a +23% increase in ad ARPDAU.

We were particularly encouraged by this outcome because the same publisher had previously attempted to add interstitials independently, only to see retention fall without any meaningful boost in revenue. So where is the surprise?

The client continued experimenting with interstitials and tried a relatively small adjustment in the flow. Instead of the original sequence:

Match finish → results screen → rewarded video to double the reward → interstitial ad

They tested a slightly altered order:

Match finish → interstitial ad → results screen → rewarded video to double the reward

The result was striking: up to a 24% increase in interstitial ad revenue. No changes in frequency caps, no signs of added user frustration, only a significant uplift from a subtle shift in placement order.

2. How cosmetic changes can dramatically boost rewarded video engagement rate

When you are responsible for driving revenue, whether as part of an in-house team or as a consultant, you quickly learn that every small detail matters. Often, what looks like a minor tweak can end up delivering results far beyond expectations. Here is one case that demonstrates just that.

The game in question was a casual, IAP-heavy title, where only 9% of total revenue came from ads. Rewarded video was the sole ad format, integrated into three placements: free Gold in the Shop, free Lives, and a post-match reward.

Here are the specific changes we implemented across each placement:

Free gold in the shop: The Gold balance was always visible at the top of the Main screen and also functioned as the entry point to the shop. However, players had no indication that something free was waiting for them there. We added a simple “cloud message” above the shop, reading “FREE GOLD!”

Free lives: Two key changes were made here. First, we made the placement available anytime a player had fewer than the maximum number of lives, instead of only when they had none left. Second, we added a small label above the Lives status bar on the Main screen, reading “Free life available.”

After-match placement: For this one, we made a few design adjustments. The button colour was changed from green to yellow, and we added the word “free” next to the existing play icon. This created a clearer distinction between the IAP option (located on the right-hand side of the post-match pop-up) and the ad option (on the left-hand side).

Additionally, a five-second countdown timer was introduced to create a sense of urgency.

The results were significant and are summarised in the table below. Which of these outcomes surprises you the most?

3. When the safest ad format lets you down

In our experience, offerwall is often among the easiest ad formats to integrate into a game, provided you know how to set it up correctly. Over time, we have established a set of guiding principles that usually lead to strong results:

• Game type matters: The game must be IAP-heavy, with high demand for the premium currency. Offerwall is not a good fit for hyper-casual games.

• Access for all: Offerwall should be available to both paying and non-paying users.

• Placement: The primary location should be within the shop, with visibility reinforced in one or two additional areas of the game.

• Exchange rate: Define the rate carefully. If Gems are the reward, you must determine the number given per $1 paid by the network. Too generous, and you risk undermining IAP revenue. Too stingy, and players will not engage.

By following these rules, we have successfully implemented offerwall in a variety of games across genres, consistently seeing no negative impact on retention, IAP performance, or rewarded video ad engagement. A few examples include:

• An idle game where offerwall generated 8.4% of overall ad revenue after integration.

• A simulation title where it contributed 14% of ad revenue.

• Another simulation game where offerwall drove 20% of all ad revenue.

• Several RPGs where the share ranged from 13% to 36%.

But as with any system, there are exceptions. In one case, the results were not what we expected. In the experimental group, IAP revenue dropped by 1.5%.

Digging deeper, we saw that conversion rates remained stable, players were still making purchases, but the average transaction value fell. Offerwall revenue was not enough to compensate for the decline, which meant the overall outcome was negative for paying users.

This led us to analyse the non-paying segment more closely, splitting them into two groups: players who regularly watched rewarded video ads, and those who did not.

• Rewarded video viewers (non-payers): Our concern was that offerwall might reduce their ad viewing. However, impressions per DAU and engagement rates remained consistent across both test and control groups. The only variance was a slight (-1.39%) lower eCPM in the test group, which we dismissed as an AB testing anomaly, since offerwall cannot directly impact eCPM for rewarded videos. Any revenue from these users, therefore, represented incremental gains.

• Non-viewers (non-payers): With no IAP or rewarded video activity to cannibalise, any revenue generated through offerwall was purely additional.

While the initial results were not ideal, we determined that offerwall could still be rolled out to non-paying players. The next logical step would have been to refine the setup for paying users. Unfortunately, the publisher was discouraged by the findings and chose to put the project on hold.

4. When the first iteration does not go well, try again

This was a truly hybrid monetisation game, with revenue split evenly between IAP and rewarded video ads. No other ad formats were in use, and the gameplay itself was casual.

Our mission was to improve ad monetisation within this title. The objectives were clear: increase engagement rate (and in turn impressions per DAU and ARPDAU) on rewarded video ads, and broaden the monetised user base by adding interstitials for non-paying players.

In our first attempt, we tried to do too much at once. We introduced a new placement, experimented with segmentation, and added interstitials. The results were difficult to interpret, and the parts we could read were not encouraging.

The new rewarded video ad placement (offered after failing a level to allow players to continue for free) generated noticeable additional revenue. However, for certain user groups, this was not enough to offset the IAP cannibalisation we observed. Interstitials did not harm retention, but the revenue uplift was lower than anticipated.

We then decided to take a step back and focus on one change at a time. We redesigned the rewarded video placement, delaying its introduction until players had completed more levels.

We also applied different caps depending on the user’s country. This iteration solved the cannibalisation problem, and the placement went on to account for 4% of overall ad revenue. With further adjustments, this figure eventually rose to 7%.

Additional rewarded video placements were introduced later, but none were able to outperform the original ones or the continuation placement established during these tests.

For interstitial ads, we optimised by increasing caps for certain countries and showing ads earlier in those markets. This boosted their effectiveness, leading to interstitials contributing 17% of total ad revenue.

5. To block or not to block competitors

The final example is somewhat borderline. Although the change is applied on the network side, it directly influences users because it alters the ads they see and can ultimately affect retention. This is one of the longest-running debates in free-to-play mobile games that monetise with ads. 

Hypercasual titles are excluded, since competitor ads make up the bulk of their revenue, and blocking them would be unsustainable. For other genres, particularly games where IAPs are the primary revenue driver, the decision carries far more weight.

We worked on refining the competitor blocking strategy for one of our clients, Playstudios, and their casino slots titles.

Before we stepped in, the team had blocked hundreds of games with similar mechanics. We chose to challenge that approach. Our process involved a thorough review, including consolidation of block lists across different ad networks. The lists had been inconsistent, which meant competitor ads still slipped through on certain networks.

We also conducted an analysis of churn data using Unity LevelPlay’s Ad Quality tool, now available for free to all publishers (as reported earlier).

After careful evaluation, we decided to unblock all competitor ads. We then tracked churn, eCPM, and the share of competitor ads both before and after unblocking. The detailed results are presented in the table below, but the key takeaway is clear: across four apps, churn did not increase. In fact, in one of the titles, eCPM rose by 20% and ad ARPDAU improved by 15%.

Final thoughts

The main takeaway from these experiments is straightforward: you cannot predict in advance how changes like blocking or unblocking competitors will affect eCPM and overall ad revenue without actually testing.

Real outcomes only reveal themselves once the experiments are carried out.

If you would like to dive deeper into the competitor blocking case, you can find more details in our dedicated case study.