If you monetize your Android app with ads, you have probably noticed that no single ad network fills every impression at the highest possible price. Advertiser demand fluctuates by region, time of day, user demographics, and dozens of other factors. AdMob mediation exists to solve this problem by letting multiple ad networks compete for each impression, driving up the effective price you receive.
For indie developers, mediation can mean the difference between a modest eCPM and a significantly higher one, sometimes by 20 to 50 percent or more. Understanding how it works helps you make informed decisions about which networks to add and how to configure them.
What Mediation Actually Does
At its core, mediation is a system that sits between your app and multiple ad networks. When your app requests an ad, the mediation layer decides which network gets the opportunity to fill that request. Instead of relying on a single network like AdMob alone, mediation creates competition among several networks for each impression.
Google's AdMob platform includes a built-in mediation feature that supports dozens of third-party ad networks. You integrate the AdMob SDK as your primary SDK, then add adapter libraries for each additional network you want to include. The mediation logic runs server-side, so you can adjust your configuration without pushing an app update.
The result is that each ad request goes through a selection process. The network willing to pay the most for that specific impression wins, and the user sees their ad. Your app receives the highest available bid rather than whatever a single network happens to offer.
Waterfall Mediation
The traditional mediation model is the waterfall, sometimes called daisy-chaining. In a waterfall setup, you rank ad networks in order of their expected eCPM, typically based on historical performance. When an ad request comes in, the mediation layer calls the top-ranked network first. If that network cannot fill the request or its bid falls below a preset floor price, the request cascades to the next network in the list, and so on down the waterfall.
The advantage of the waterfall is simplicity. You set up tiers, assign eCPM floors, and the system works through them sequentially. The disadvantage is that the ranking is based on averages. Network A might have the highest average eCPM, but for any given impression, Network B might actually be willing to pay more. Because the waterfall calls Network A first, Network B never gets the chance to bid.
Waterfall mediation also introduces latency. Each network call takes time, and if the first several networks cannot fill the request, the user might wait noticeably before seeing an ad. This latency can affect user experience, particularly for interstitial and rewarded ad formats where the user is actively waiting.
Bidding Mediation
Real-time bidding, sometimes called in-app bidding or open bidding, addresses the limitations of the waterfall. In a bidding setup, all participating networks receive the ad request simultaneously and submit their bids in real time. The highest bidder wins the impression regardless of their historical ranking.
This approach is more efficient because every impression goes to the network that values it most at that specific moment. There is no sequential calling, so latency is reduced. And because the competition is genuine for every single impression, average eCPMs tend to be higher than in a pure waterfall configuration.
AdMob supports a hybrid model where bidding networks compete alongside waterfall networks. Bidding partners submit real-time bids, and those bids are compared against the waterfall tiers. The highest value wins, whether it comes from a real-time bid or a waterfall tier. This hybrid approach lets you benefit from bidding while still including networks that do not yet support real-time bidding.
Popular Mediation Networks
The mediation landscape includes dozens of networks, but a few consistently perform well for Android apps. Meta Audience Network remains one of the strongest demand sources, particularly in North America and Europe. Unity Ads is a dominant player in the gaming vertical. AppLovin and ironSource (now merged under the Unity umbrella) offer strong performance across multiple ad formats. Pangle, operated by ByteDance, has grown rapidly and offers competitive bids in many Asian markets.
For most indie developers, starting with three to five networks is practical. Adding too many networks increases SDK bloat and complicates testing. Focus on networks that are strong in the regions where your users are concentrated and that support the ad formats you use most heavily.
Configuration Tips for Better Results
Once mediation is set up, ongoing optimization is where the real revenue gains happen. Review your mediation performance weekly, paying attention to fill rates and eCPMs by network and by country. A network that performs well in the United States might have poor fill rates in Southeast Asia, and vice versa.
Set floor prices carefully. A floor that is too high reduces fill rate because fewer networks can meet it. A floor that is too low means you are leaving money on the table. Start with conservative floors and adjust based on observed performance. Many developers use A/B testing to compare different floor configurations across segments of their user base.
Keep your ad network SDKs updated. Outdated adapters can cause fill rate drops, crashes, or policy violations. Each network releases updates that improve bid accuracy, fix bugs, and comply with evolving privacy regulations. Schedule regular dependency updates as part of your release cycle.
Monitor for anomalies. A sudden drop in one network's fill rate might indicate a configuration issue on their end, a policy change, or an SDK incompatibility after an app update. The sooner you identify these issues, the sooner you can adjust your mediation stack to compensate. AdMoola lets you monitor your AdMob earnings, eCPM trends, and impression data directly from your Android device, making it easier to spot the impact of mediation adjustments without waiting to get back to your desk.