Applying Our First Interstitial Ad

Boosting ad revenue with low impressions

1

Generating a nice chunk of revenue with their high prices, interstitial ads are incredibly tempting, but they're also hard to just casually drop into our product because they can easily disrupt the user experience. The fact that they cover the entire screen and force the user to watch an ad for a few seconds (usually around 3 seconds) is enough to cause frustration.

Aside from the negative impacts that show up in metrics (like a drop in user retention), there's a huge concern that it could ruin the overall image of the product. It can easily give off the vibe of, "Oh, this app is just trying to squeeze money out of me with ads however it can."

Because of these worries, we forgot about interstitials for a long time. Then, last quarter, I came across some analysis data on interstitials, and I realized we couldn't ignore them anymore. The revenue impact was projected to be not just a "nice chunk," but a massive "game changer." Plus, seeing roughly with my own eyes that similar services were already making "huge revenue" by serving interstitials in "small volumes" made me feel like we were falling behind. I really wanted to figure something out.

First approach: Let's reduce the impression rate

When trying to generate revenue from ads, we naturally tend to plan for higher impressions per user. That's because getting 10 impressions per user can bring in 10 times the revenue compared to just 1 impression. With interstitials, we took the opposite approach. Since their unit price is 30 times higher than regular banners, reducing the impression volume didn't seem like it would hurt revenue. Since our basic banners and native ads were already bringing in a steady amount of revenue, we viewed interstitials as a supplementary revenue source and decided to start a bit conservatively. Our estimates showed that even if we only showed them to 1% of total users, we could secure a pretty decent extra revenue stream.

Second approach: How should we reduce the impression rate?

Should we show it at a 1% random probability during a high-traffic user journey, or should we show it at 100% probability during a user journey that makes up about 1% of traffic (e.g., visiting the settings tab, visiting a specific feature screen)? We decided to go with the latter. Thinking back on my own experience with interstitials in other apps, you quickly learn the conditions under which they appear. We figured that if users could learn and predict the flow, it would feel less frustrating and more familiar. Now we just needed to decide which journey to put it in.

Third approach: What reward should we give for showing it?

When we discussed this internally, everyone agreed that since an interstitial covers the entire screen and blocks actions for a few seconds, we needed to give users a specific reward in exchange. Typical services show interstitials when users try to use a certain feature. If we showed an interstitial right when someone was trying to dismiss their alarm, it would be a total disaster. After exploring various moments, we decided to tie it to snoozing the alarm.

In our product, there's a feature that limits the number of times you can snooze an alarm. For example, a user who sets it to 3 can only snooze exactly 3 times. At this point, we offer a one-time "Watch ad to snooze one more time" option and show an interstitial. A user who desperately wants to snooze just one last time chooses to watch an ad in exchange. The moment they watch the ad, the alarm is snoozed once more, and they can't snooze it again after that.

The results?

It worked like a charm. From the start, an amazing CPM of 10 USD held steady, bringing in quite meaningful revenue even with low impression volume. Even though it's shown to less than 0.3% of total users, it generates 1,000 USD a week. Thankfully, there were no noticeable user complaint tickets, (just a few on Android at launch due to an interstitial bug...) and the user retention rate didn't drop either.

If the proportion of users who limit their snoozes grows, the percentage of users seeing the "Watch ad to snooze one more time" option will also expand from the current 0.3% to 1% or 3%. If ad revenue scales proportionally, we expect this to create an extra monthly revenue stream of 40,000 USD.

Actually, we didn't set up any mediation for the interstitial placement in this experiment. We were hitting that 10 USD unit price from a single ad source. Recently, our DARO team ran mediation optimization for the iOS interstitial placement, and the unit price has skyrocketed and stabilized at a whopping 25 USD.

The first interstitial introduced in our product...! I'm so glad the results were highly positive.

If applied by designing the right impression volume and the right user journey, interstitials can become a rock-solid revenue source.

They might seem a bit scary at first, but why not gather some courage(?) and take a look to see if there's a good moment for them in your product?

If you're interested in ad monetization, feel free to reach out to DelightRoom anytime~