Best Ad Networks for Apps, Compared

For most apps the best ad setup is not one network. It is a mediation platform with several networks bidding inside it. In our benchmark estimates AppLovin MAX comes out on top for Tier 1 price and fill, AdMob is the easiest place to start, and Unity LevelPlay is the natural fit for games built in Unity. The rest earn their place as extra bidders, not as a first choice.

The question people type is "what is the best ad network", and it hides a mix-up. Some of the names you will meet are ad networks (they sell ads), some are mediation platforms (they run an auction between networks), and some are both. Pick the platform first, then decide which networks to put inside it.

This guide compares the eight networks in our data, says who each suits, and shows why a mixed stack usually beats a single source. All dollar figures are modelled estimates from our benchmark tables, not figures from any network's reporting.

Networks, platforms and the names that confuse people

An ad network brings advertisers and pays you for showing their ads. A mediation platform sits in your app and asks several networks for ads, then shows the best one. Mediation versus a single network goes through the trade-offs of adding that layer.

Here is how the main names fit:

• AdMob is Google's network and also a mediation platform. Google's mediation help page says you can run bidding, where ad sources compete in a real-time auction, or a waterfall, where sources are called one by one in order of the average eCPM you set. You can mix both in the same group.
• AppLovin MAX is a mediation platform. AppLovin states on its monetization page that MAX mediation is free for publishers, that bidding networks pay AppLovin a fee on the ad revenue they generate, and that every network, AppLovin included, competes in the same unified auction.
• Unity LevelPlay is Unity's mediation platform. Unity's Ads and ironSource page describes it as connecting 25+ ad networks, with A/B testing and reporting. Unity Ads and ironSource Ads are separate networks that bid inside it, and they can also bid in other platforms.
• Chartboost Mediation is a platform too. Its documentation calls it a bidding-based product with more than 33 demand sources.
• Meta Audience Network, Liftoff Monetize (formerly Vungle) and InMobi are networks. You normally reach them through one of the platforms above.

The practical effect is simple. You choose one platform and add the networks you want. Our guides to AdMob mediation setup, AppLovin MAX and Unity LevelPlay cover the setup for each.

What the networks pay in our tables

The table gives our Tier 1 Android baselines for the three main formats, the Tier 3 rewarded figure and blended fill. iOS runs about 1.3 times higher in our model, and every figure has a range around it.

Two things stand out. The spread at the top is smaller than people expect: AppLovin's rewarded estimate is about 21% above AdMob's. And InMobi, last in Tier 1, jumps to second in Tier 3 rewarded. The full set, including native and app open, lives in the mobile eCPM benchmarks.

Price alone misleads. Multiply by fill and you get the value of each request. For Tier 1 rewarded, AppLovin works out at 14.50 x 0.90 = $13.05 per thousand requests and AdMob at 12.00 x 0.88 = $10.56.

Network by network

AdMob

The default for a first app. Setup is quick, the dashboard is clear and Google's advertiser pool is broad. Together with AppLovin it is one of only two networks in our tables with all five formats, including app open at $5.50 in Tier 1 and native at $3.50. Google's own docs now mark the older Google Mobile Ads SDK as in maintenance mode and point new work to the GMA Next-Gen SDK, so check which one the tutorial you follow uses.

AdMob also runs mediation, so you can add other networks later without changing platforms.

AppLovin MAX

The highest Tier 1 figures and the highest fill in our tables. MAX is built around the unified auction. AppLovin's getting started guide lists seven steps from account to reports, and it recommends prioritizing bidding networks because they cut manual waterfall maintenance and latency. The Mediation Debugger checks that each network is set up correctly.

The cost is attention. A MAX stack has more moving parts than a plain AdMob one. The head-to-head page for AdMob and AppLovin shows our numbers side by side.

Unity Ads and ironSource (LevelPlay)

Unity Ads has the second-best rewarded and interstitial estimates, at $13.00 and $9.00 in Tier 1, though it has no native or app open in our data and weaker fill at 84%. ironSource Ads sits between Unity Ads and AdMob. Both lean on game advertisers, which suits games and does less for a utility app.

If you build in Unity, LevelPlay has a Unity SDK and it is a sensible mediation choice. Ads in Unity games covers the Unity side.

Meta Audience Network

Meta's developer docs centre on app bidding, an open auction where each ad opportunity is offered to several demand sources in real time. In practice you add it as a bidder inside a mediation platform. Its native estimate of $4.20 in Tier 1 is the highest in our tables, ahead of AppLovin's $4.00. Overall fill is 80%. Read our Meta Audience Network guide for the setup.

Liftoff Monetize (Vungle)

Google's integration page states that Vungle is now Liftoff Monetize. Liftoff's rewarded and interstitial estimates sit close to AdMob's, at $12.00 and $8.20, and Google lists both bidding and waterfall support for it in AdMob mediation. It is a reasonable bidder to test in games and in apps with video-heavy placements.

InMobi

The weakest in Tier 1 and one of the strongest in Tier 3. If your audience sits in India, Southeast Asia, Latin America or Africa, it deserves a slot. The InMobi guide has the numbers, and our Tier 3 figures are in the benchmarks guide.

Chartboost

Chartboost's own pitch is mediation: its docs say bidding partners pay no demand fees. As a network in our tables it comes in at the lower end on price, with 78% fill. That does not make it a poor addition, just a secondary one. Test it as a bidder and watch its win rate before you decide.

The ones we do not model

AdMob's list of mediation sources is longer than ours. It includes Pangle, Mintegral, Moloco, DT Exchange and BIGO Ads, among others, and any of them might beat a network in our table for your particular audience. We have no benchmark data for them, so we cannot rank them. If a platform recommends one, treat it as a test, not a given.

Which network for which kind of app

Our category multipliers sit on top of the network figures: hypercasual 0.8, casual 0.95, puzzle 1.0, utility 0.9, social 1.0, mid-core 1.35 and finance 1.6. They shift the price, not the ranking, so the order of networks stays similar across categories. What changes is how much each one matters.

• Hypercasual games: high volume, rewarded and interstitial heavy. Run MAX or LevelPlay with several bidders and keep the SDK count sensible. Our best ad network for hypercasual games page gives the short answer.
• Casual and puzzle: mediation again. The audience is global, so fill matters as much as price.
• Mid-core and strategy: the 1.35 multiplier makes every point of eCPM worth more, which justifies the work of testing bidders.
• Utility and education: low volume and a small number of placements. AdMob alone is fine until daily users reach the thousands.
• Finance: the highest multiplier at 1.6, but a narrower pool of advertisers. Test before you assume the number holds.
• Kids apps: the options shrink. Ads in apps aimed at children come with their own rules, so read ads in kids apps before adding any SDK.

Why a mixed stack beats picking one

A single network runs one auction among its own advertisers. A mediation stack runs several, then picks the winner, so two things improve at once: fewer requests go unfilled and the winning price rises.

Here is an example with assumptions. An Android game sends 100,000 rewarded requests a day from Tier 1 countries.

• AdMob alone: 88% fill gives 88,000 impressions. At $12.00 per 1,000, that earns about $1,056 a day.
• AppLovin alone: 90,000 impressions at $14.50 earn about $1,305 a day.
• An AdMob-led stack with extra bidders: assume it reaches 93% fill and a 10% price lift, so $13.20 per 1,000. That is 93,000 impressions and about $1,228 a day, roughly 16% more than AdMob alone.

The last line rests on two assumptions, not results, and real stacks vary a great deal by audience. AppLovin alone still beats the assumed AdMob stack in this example, a reminder that the lift you assume is the soft part of any forecast. What the arithmetic does show is that letting more buyers compete rarely costs you anything, and the gain tends to be larger in Tier 2 and Tier 3, where one network's demand is thinner.

Bidding versus waterfall explains why modern stacks lean on bidding. Reading your own reports then tells you what the stack is really doing.

Reading the results once the stack is live

Give a new network a few weeks before you judge it. Look at three numbers per network: how often it fills, what it pays on the impressions it wins, and how much of your total revenue it contributes. A network with a high eCPM on a handful of wins is not helping much, and one with modest eCPM and steady volume may be carrying more of your income than you think.

Compare like with like. Split the reports by country and format before you compare networks, because a rewarded impression in the United States and a banner in Brazil are different products. Our fill rate explainer shows how to read requests, matches and impressions together.

Change one thing at a time. Add a bidder, wait, then look at revenue per daily user and day-one retention, not just eCPM. An extra interstitial from a generous new bidder can lift today's revenue and cost you players next month.

What to check before adding a network

Every network is another SDK, another adapter and another set of rules.

1. App size and crashes. Adapters must match the SDK versions they were built for. Update them together and test on real devices.
2. Seller files. Each network wants its line in your app-ads.txt. A missing line costs you demand from buyers who only purchase from authorized sellers. The Google Play app-ads.txt guide shows the placement.
3. Consent and privacy. The consent flow has to reach every network in the stack, and kids' apps have stricter limits.
4. Payouts. Networks that pay you directly each have their own threshold and timing, so expect several smaller payments. AppLovin's setup guide, for example, has you enter payment information in your account.
5. Policies. Each network has publisher policies, and a violation can hit your earnings with that network or, in the worst case, your account.

Common questions

What is the best ad network for a new app?

AdMob, for the first release. It is quick to set up and has the broadest format coverage. Add mediation when your daily users and your Tier 1 share make the extra work worth it.

Which network pays the most?

In our tables AppLovin MAX leads in Tier 1, InMobi is second on Tier 3 rewarded price, and Meta leads in Tier 1 native. Your own audience will rearrange this, which is why you test.

Should I use AdMob or AppLovin MAX for mediation?

Both support bidding. Our tables favour MAX on price and fill, and AdMob mediation is simpler if you are already on Google's stack. Run your own traffic through the AppLovin MAX calculator and the AdMob revenue calculator before you decide.

How many networks is too many?

There is no fixed number. Every network adds SDK size and upkeep, so judge each one by how often it wins and what it pays when it does. Drop any that rarely wins.

What to do next

Start with where your users are. Open the mobile eCPM benchmarks and check which tier most of them sit in. If it is Tier 1, put AdMob and the AppLovin MAX calculator side by side with your daily users and format mix.

Then choose a platform. For a first app, stay on AdMob and add the biggest bidders one at a time. For a Unity game, try LevelPlay or MAX. Whatever you pick, add one network per release so you can see what each one changes, and keep an eye on how to monetize an app for the income that does not come from ads.