Many publishers who are struggling to optimize their interstitial revenue are focused on the wrong problem. They are chasing minor improvements to their creatives, testing colors, CTAs, and imagery, when the real opportunities lie in timing. The timing of the ad, the timing of the next ad, and the readiness of the user to engage with the ad all contribute significantly to ultimate engagement. Getting these wrong hurts retention in pursuit of higher revenue per session.
What is, precisely, a “natural pause” for ad presentation?
The ad industry has vague language around “a natural transition point in the narrative”, but this is rarely quantified with precision. Natural pauses are not simply moments between two events; they are periods where the cognitive load of the user has dropped to zero following an event.
In a mobile game this means between levels, when the level has been completed and the next has not yet begun. In a utility app it means immediately after the task is completed and before the download/initiation prompt. Long-form editorial is the same concept as between pages when initiated by the user. In all these examples the user is in a period of zero cognitive load. Their mental energy is not being expended on a decision, and that is when you should interrupt them.
Why is the timing of the ad trigger so critical?
Most publishers that have monetized their products are implementing the logic incorrectly or leaving significant revenue on the table. The standard interstitial frequency capping is one impression per user per 24 hours, and one per 12 hours in heavier categories like casual games.
There is math behind this – the average mobile session length is between 4-8 minutes. Two sessions per day, capped at 1 impression, means your user is being interrupted once every 10-16 minutes. That is a tolerable frequency for an interstitial as it does not intrude on the continuity of the experience.
Once you get beyond 3-4 impressions per day, the math gets more complex, but the principle holds true that there is a tolerable upper bound before ad fatigue turns into negative perceptions of the product itself.
eCPM also follows a non-linear curve in regards to frequency. Many publishers assume that two impressions equal twice the revenue of one, but in practice the second impression has a much lower CPM than the first, due to a combination of factors relating to demand-side bidding and the lower perceived engagement of frequent impressions.
Why are landing page interstitials a bad idea?
Putting a full-screen advertisement between a user and your content is counterproductive in more ways than one. Google has stated that mobile pages using intrusive interstitials see a bounce rate increase of more than 40% compared to pages using delayed interstitials.
This penalty is primarily psychological, as a user who has clicked through to a page expecting to fulfill a specific need is presented with a full-screen advertisement as soon as they arrive. They are immediately aware on a conscious level, or within 2 seconds at the latest, that this product is prioritizing its own revenue above their time. Any subsequent engagement with the page is tainted by this realization.
What is the difference between an active trigger and a timed trigger?
Publishers who use “after 30 seconds” ad triggers are taking an educated guess at when the user is ready to engage. In many cases this is incorrect as the period between user actions is much shorter than the publisher assumes. When an ad is triggered by the user hitting “Next”, submitting a form, or clicking a nav element, the publisher is capitalizing on the transition the user has already initiated. The user is much more receptive to an interruption when they are already in the middle of navigating.
The difference in conversion rates between active and timed triggers is significant enough that it should inform your entire trigger selection strategy. Active triggers not only have better click-through rates but better overall performance since the user is much more receptive to the interruption and more likely to follow through on any actions requested.
<p>For publishers comparing the best interstitial ads across both web and app, flexibility around trigger timing should be a key consideration. Networks that can tie your interstitials to active triggers without customization on a per-placement basis give you the best option for implementing the active trigger methodology.
Pre-loading assets and the ghost click problem
The performance characteristics of interstitial advertisements are frequently misrepresented due to technical limitations that cannot be measured in creative optimization. For example, if an interstitial is triggered by a transition but takes 800ms to load, a user who had clicked “Next” and then clicked again in the same location while waiting for the ad to load would register as a ghost click. The second click is interpreted by the system as a request to close the ad, but the ad had not yet appeared on screen.
Ghost clicks are problematic as they artificially inflate CTR while providing no value to the advertiser. Worse, the user is being sent to a location they were not requesting to see. The damage to your product’s perceived quality from such a click is difficult to quantify, but it is almost certainly hurting your retention.
The technical solution to ghost clicks is to remove the technical limitation which makes them appear in the first place. Your next interstitial should be pre-loaded, with its assets cached locally, rather than waiting for the trigger event to begin downloading them. When the transition trigger fires, the interstitial will appear with less than 100ms latency due to local caching of the assets.
Segmenting frequency capping by user value
A blanket policy of ad frequency capping ignores the obvious fact that certain users engage with your product more than others and therefore have a higher tolerance for interruptions. It would make little sense to interrupt a power user who engages with your product every day over a user who has only ever engaged once, despite both being treated as equivalent users.
Instead of a flat frequency cap consider implementing a segmented system which uses a user’s value to your business as a reference point. High-value users who engage frequently, make purchases, and spend more time on your product are appropriate candidates for reduced ad frequency. These users will generate much higher revenues across their lifetime, and an interruption-heavy experience will hurt your ability to capture that revenue.
Low-value users, such as occasional visitors or those who landed on your product via a bad referral link and left immediately, are appropriate candidates for increased ad frequency. This is especially true if they were never likely to convert in the first place. Your ability to capture revenue from them is extremely low regardless of advertising frequency, and interrupting them more often will not hurt your retention in the long run.
This is not about giving certain users fewer ads versus more ads; it is about maximizing the lifetime value of your users as they engage with your product. An interstitial shown at the right time to the right user is a win-win for everyone involved, whereas the same interstitial shown to your power users too frequently is a lose-lose.
Close buttons are not the enemy of engagement
There is a certain amount of counter-intuitive thinking that goes into user acquisition and engagement, but one myth that persists is that the presence of a close button encourages users to dismiss your ad. In reality a close button which appears within one or two seconds after ad load actually encourages engagement.
The reason for this is simple: A user who knows they can close your ad at any time is much more receptive to its presence and content than one who cannot close it for five seconds or longer (or who has to find the close button in a tiny corner of the screen). Users who see a close button as an exit are not annoyed by its presence and are in fact often encouraged to observe the ad content to see if it has useful information before closing it.
A slow appearance of the close button or its absence encourages hostility in the user which will reduce their receptivity to the information being presented and subsequently damage their retention.
What you should be measuring when you run a test with different timing windows?
Most advice on A/B testing interstitial ad timing revolves around the “15 seconds vs 45 seconds” test, which is useful but lacks nuance. Prior to designing the test, decide what you want to measure. Scroll depth at the point of trigger? CTR? Duration of the session immediately following the ad? 7-day retention?
A two-point test is fine for establishing a baseline, but identifying the time window at which user attention intersects with tolerable interruption frequency is more complex. It depends on your content type, your device, and your sources of traffic. One timing window will work best for a mobile game between levels, and it will be different for a long-form web article. The underlying principle of finding a local minimum of user engagement, however, will remain the same. The point where users are most receptive to a disruption is the window in which you should show the interstitial.
The information you need to base your decision on is not merely the raw click-through rate figures but a combination of CTR and the following engagement. A 15 second delay may get you more CTR’s but fewer engagements with your app/store/content, and vice versa for a 45 second delay. Which has the bigger impact on the metric you care about most—retention or revenue—depends on whose data you use and what your ultimate goal is.
The architecture makes the revenue
Interstitial ad performance is a systems problem, not a creative one. Getting the fundamentals right – the interstitial timing, frequency, trigger, and presentation – makes a significant difference to ultimate engagement. The publishers who are seeing the highest revenue gains from their interstitials are the ones which approach it from the perspective of a systems optimization rather than an afterthought. And, critically, they are not sacrificing long-tail retention in the process.
