How many networks are needed to optimize mobile ad monetization?

Posted on July 17, 2017 by Eric Benjamin Seufert

This guest post was written by Jonathan Raveh, the Director of Monetization at CallApp, a caller ID app. Read Jonathan's previous Mobile Dev Memo post, 8 Simple Tips for Effective Ad Mediation.

When you venture into the world of ad monetization, you quickly find yourself surrounded by many similar propositions. Every network is bound to talk to you about eCPM and explain why and how they can produce the highest rate for you. In most cases, you found yourself wishing you only had to deal with 1 or 2, which is why so many publishers choose mediation to manage their ad inventory. However, that does not solve the issue entirely.

CallApp, a caller ID app with over 35M installs worldwide, is an app relying on mainly ad revenue to sustain itself, so choosing the right numbers of partners to integrate is crucial for us. Having limited resources to help with the integration, let alone the day-to-day management of an elaborate waterfall, we aligned ourselves with Mopub as a mediation tool. However, we quickly found that mediation itself did not provide us with a full solution, especially when venturing outside tier 1 geographies: with 3 billion ad requests per month, our revenue potential was not being fulfilled. There was a learning curve to optimizing monetization, and we were eager to learn, so we put aside everything else and took a deep look into ad revenue alone. We then asked ourselves what characteristics CallApp possesses that have any relevance to ad monetization and specifically ad networks. We found 4 main parameters:

  1. Format – We only use 1 ad format: native ads. Our app usage does not really support full page ads of any sort, and we try to maintain a great UX throughout the app;
  2. OS – Android. CallApp is strictly Android focused.
  3. Regions – Not a similar distribution, but we see high volume of users & ad impressions in 4 different regions: North America, LATAM, EAMA and APAC;
  4. Maximization – We were not willing, as a company objective, to settle for less than achieving 100% of CallApp's revenue potential. 80% is not enough.

This was the starting point of the 'FORM' model we created.

The 'FORM' model (Format, OS, Regions, Maximization) is a very simple model to help estimate the number of ad networks you need to reach your revenue aspirations. This model does not address eCPMs, revenue & optimization. In other words – on its own, it won't get the job done in terms of reaching 100% revenue potential. It will, however, suggest the number of tools you need to get the job done.

FORM model is comprised of 4 parameters, 3 of which are mostly objective, and 1 which is purely subjective:

  1. Formats – if you're using one main ad format (such as interstitials), that would limit the number of networks you need to use. Since large apps usually utilize more than one format, the number of networks you required should grow with DAU scale;
  2. OS – This is pretty self-explanatory. You'll find that most networks have an orientation towards a specific OS. Not all of them can monetize your traffic well if it comes from both Android and iOS. Another reason to diversify in its own right;
  3. Regions – Sure, Google, FAN and AdColony/Applovin/Chartboost might get you covered in US & Canada, but what about other tiers? Most popular apps tend to have a large user base somewhere, and a lot of other users scattered around. Those users are out of focus when it comes to monetization -- a very common publisher mistake;
  4. Maximization – This is the bottom line, and the bottom line is usually very subjective: where do you aim to get? Do you have enough manpower and expertise to manage multiple networks? Are you satisfied with achieving 70%-85% of your revenue, or are you willing to work hard to reach over 90%?

The FORM model is very simple:

  • Each format, OS, Region is worth 1 point, as well as a decision to reach over 90 percent of revenue;
  • Add your points up and you get a number. Create a range of +-1/2 around that number, and you have your range;
  • When comparing your ideal range to what you currently have: if you are using mediation, consider the number of active, revenue generating networks inside the mediation that are getting access to your traffic.

To make things very clear - This is by no means an accurate model. Looking at the 4 parameters above, most seem pretty obvious and easy to answer, but that's not always the case. Everybody wants to capture 100% of their revenue potential, but does every business have the resources available to invest to get it done? This question becomes even harder  to answer when these resources could have been spent to generate growth. The same goes for deciding whether 8% ad impressions coming from Africa constitute as a separate region or not.  

In other words, there might not be a golden number, but there most definitely is a golden range.

Now, let's break the FORM model into some examples:

If you're willing to sacrifice about 10%-25% of your revenue for the easy life, and your distribution is highly concentrated around tier 1 countries, you might just need Facebook and Google. 1-3 networks could be enough. This is as low as it gets.  

With 2 main regions to monetize, 2 ad formats, 1 OS and a desire to reach over 90 percent of your revenue potential, you need about 4-7 ad networks.  Add another ad format or region, and you're looking at 5-8 networks.

Taking this to the extreme: 4 ad formats (i.e. native, interstitials, playable ads, banners), a pretty equal 3-region distribution (LATAM, Africa, APAC & North America, 20%-30% each), 2 OSs and a desire to max out your earning potential (over 90%) will require 8-11 networks.  

Lastly, let's put CallApp to the test with the FORM model:

  • 1 ad format
  • 1 OS
  • 4 strong regions
  • 1 willingness to do whatever it takes to maximize revenues

7 points totals, which put us in the range of 5-8. With Mopub mediation, we only had about 4 serious active networks, so we know we needed to add a few to, at the very least, figure out if we are on the right track. After careful selection and negotiations, we added additional external partners. At this point in time, Google & Facebook still serve as our premium networks, generating most of our revenue, strengthened with Mopub (marketplace), Yahoo, Mobvista, Appnext, and Pubnative: A total of 7. Each added network has a different objective, most being region-oriented, which is Callapp's main optimization challenge. Having to deal with so many SDK's, with most of which weighing over 500K, updating versions and managing such an extensive waterfall becomes a huge, time-consuming task. As mentioned before – this is probably NOT every developer's first choice. For some developers, the extra work will make this infeasible. For other developers, an extra 5MB in SDK's will be a deal-breaker. Making a decision to work with so many partners demands dedication.

Now, is that number even enough? Honestly, we don't know. We are considering adding 1 or 2 more if and when the time is right. We are also fully aware that any new network will not likely provide us with 10% incremental revenue. It's also possible that by better managing the waterfall we could drop 1-2 existing networks and continue with our current revenue level. These considerations are reflected very well in the range we have set up using the FORM model. At CallApp, we prefer to work with an extra network that can quickly cover up any loss by a different ad network. Fluctuations in the number of campaigns & eCPM are very common, and being dependent on too few networks is the worst possible situation when things go sour. The range calculated using the FORM model does acknowledge the differences between managerial styles and organizational appetite for risk.

Still, the bottom line is that most developers play the ad network game as they go along, without putting considerable time and effort analyzing their own app performance, their organizational goals, and the number of 3rd party tools they need to reach those goals. Selecting the right networks is a must, of course, but for that extra 10%-25% revenue, it's not only quality that a developer should consider, but quantity as well.    

Jonathan Raveh is the Director of Monetization at CallApp, a caller ID app. Jonathan has 10 years' experience in the mobile industry in various content & advertising positions, with previous position as the Director of Partnerships & Community at Appnext, a mobile ad network. Jonathan operates his own Hebrew language mobile industry blog.