Please note that as of October 24, 2014, the Nokia Developer Wiki will no longer be accepting user contributions, including new entries, edits and comments, as we begin transitioning to our new home, in the Windows Phone Development Wiki. We plan to move over the majority of the existing entries over the next few weeks. Thanks for all your past and future contributions.
Windows Phone ads providers performance
In this article I try to explain and clearly show my experiences with different ad providers in Windows Phone. As a result of this study, I have concluded that AdMob offers the most revenue for apps that have a target audience outside the USA and Canada, but that other systems offer different benefits for different use cases. The rest of the article explains my findings.
For the last month I was performing an experiment in the Windows Phone Store. The objective was to get a picture of what monetization model we can use for free apps. I think a combination of free app, ads and in app purchase could be the best, following some rules:
- The user doesn’t need to pay for the app complete features.
- We can show in app Ads and monetize the app this way with ads revenue.
- The user has the choice to remove ads by using In App Purchase, paying 0.99€.
- In no case, the user is forced to buy app features.
Three ads providers were tested during this month:
- Microsoft PubCenter
- Adduplex ads network
- Google Admob
The business model with Pubcenter and Admob is straight forward: every time a user gets an ad, we get money. In the case of Adduplex, the idea was to interchange ads with other apps and attract users to the app so IAP buy rates are high enough to cover the loss of ads revenue.
As a result of this study, I have concluded that AdMob offers the most revenue for apps that have a target audience outside the USA and Canada, but that other systems offer different benefits for different use cases. The rest of the article explains my findings.
The application I used for this experiment was PHOTOLab. This app allows a user to pick a photo and, using Nokia Imaging SDK 1.1, apply different filters and effects to it. Finally the user can share the result or save it in the local gallery. During this month of tests, PHOTOLab gets around 35.000 downloads, 413 reviews (with a medium of 4 stars) and was showcased in Windows Phone Central, Windows Mobile Power User, Windows Phone Apps, and some other regional blogs. The App was translated to English, Spanish and Portuguese. Also Microsoft has promoted the app in the Windows Phone Store in Saudi Arabia, United Arab Emirates, Macedonia, Poland, Indonesia, Malaysia, South Africa, Canada and United Kingdom.
On average, this app gets about 1250 daily downloads. We can see in the chart above some high download days, coinciding with the app being showcased on the Store or on popular websites. Now that we have a solid context of what is happening with the app, is time to start taking a look at ad performance. We are going to review three different periods:
- From 02/26/2014 to 03/03/2014 we used PubCenter as our ads provider.
- From 03/04/2014 to 03/17/2014 we used AdDuplex as our ads provider.
- From 03/17/2014 to 03/22/2014 we used AdMob as our ads provider.
As we can see, the AdDuplex period is much larger than the others. I want to wait until downloads for the app decrease to levels equals to the initial one before using AdMob, so I can compare the revenue in PubCenter and AdMob in equivalent conditions. AdDuplex was used during the most active period and, thus far, has the biggest numbers of this little study. How were ads placed inside the application? PHOTOLab has 5 screens:
- Main screen: select gallery or camera, view 5 last photos in your gallery, access for settings and about. Menu option for ads removal.
- Photo editing screen: Apply 32 different filters to the photo, view original to compare against edited, zoom photo…
- Save and share screen: After finishing your edit, this screen allows you to see the final result and save to phone gallery or share using the system contract.
- Settings: Allows the user to enable or disable a secondary tile for the app.
Only Photo editing and Save and share screens have ads. The average user session in the app sees 4 screens:
The average session duration is about one minute and 20 seconds. Of that time, the user spends around 37 seconds in the Edition screen and 10 seconds in the save screen, so we have an average of 47 seconds per sessions showing ads in the application. With an ads refresh configuration of 30 seconds, we can show 2 ads in the first 37 seconds and 1 ad in the next 10-second screen. This means, in a 100% fill rate world, we can have around 4000 ads impressions in the app. As we will see, this isn’t the situation in every case. But let's start with the In App Purchase across the total period of this last month, and how advertising provider can affect IAP.
In App Purchase
In PHOTOLab, IAP was introduced only as an option to remove ads from the application. To do this, the user has an option in the main screen ApplicationBar menu and a “X” button is shown above the ads in the edition and save screens. In one month, 12 people buy the “ads removal” IAP item placed in the application. For an application with 35.000 downloads, an average of 32.000 users and more than 69.000 user sessions, 12 is a very low number. First, I think IAP for ads removal is really a way for the user to directly pay a developer for her/his great job with the application. I talked with a lot of users and many of them think about IAP ad removal as a way to reward the developer. Other users think about IAP ads removal when the ads are intrusive or placed in a way that causes the app user experience to suffer. In the case of PHOTOLab, the ads don't distort the user experience and aren’t present in all screens, so the effectiveness of IAP is smaller.
Now it's time to start analyzing the Ad revenues and performance. We are going to start with PubCenter.
Microsoft PubCenter is the default and easiest method to include ads in your app in Windows ecosystem. It is included with the SDK and is easy to use. These are the PubCenter performance numbers from 02/26/2014 to 03/03/2014:
As we can see in the table and graph above, we have an average active user count of 967, with 1650 daily ad impressions. This led to an average eCM of 0.07€ and an average daily revenue of 0.11€. If, as we previously analyzed, we have an average of three ad impressions per user session, and an average 967 user sessions, we are talking about: 967x3 = 2900 ads impressions per day average. But in the real world, we only manage to have an average of 1650 daily impressions. And this is the first and foremost problem with PubCenter: The fill rate outside USA and Canada is really bad. Every day we ask for near 3000 ads and only receive near 1650. That’s a fill rate of 55%; 45% of time the app displays a blank placeholder in the app. This can be worse if your main market target is Spain, South America or Asia. In Spain you can barely see one ad per day in the app.
The second problem is the eCPM rates. On average you only get 0.07€ per mile impressions. So you need to do a lot of impressions to earn even 1€ but, as we saw, you aren’t going to have a big impression number unless your market is centered in USA or Canada. This could also affect IAP. As you aren’t displaying ads, the user doesn’t have any reason to pay to remove them.
The second ad provider we are going to analyze is AdDuplex. AdDuplex doesn’t pay you by impressions; it's an ad exchange network. For each 10 ads of other apps showed in your app, they show 8 ads of your app in other apps. Here the trick is to be able to display a lot of ads, so you receive in exchange a lot of impressions in other apps. The goal is that you get a lot of new users who use and rate your app, and also use IAP to remove ads, so your IAP revenue can growth. As we can see, that’s not what happens in every case. The period analyzed was from 03/04/2014 to 03/17/2014:
|App version||Date||Sessions||Impressions in app||App impressions||Clicks|
In this case, we have 2100 average users daily, and get 5180 average ads to show in our app. This for a fill rate of 82%. That’s actually a great fill rate. The problem here with AdDuplex is, we only receive an average of 30 new downloads of our app per day. In my honest opinion, when your app is getting 2100 average new users per day, getting 30 more is not a significant enough increment to spend a 10% of your screen real estate displaying ads. We are talking about getting only a 1.5% increment of users. Maybe in other apps, or for apps with lower daily users, it’s a valid promotion system. In this experiment and in my opinion, I'd rather remove the ads because I’m showing ads in the app and not getting a great revenue.
Finally we are going to review AdMob ad network performance. I stop promoting the app and updates to get the app downloads lower so I have a number of users like the one we had in PubCenter. In fact, during the last period the number of users was a bit lower than the PubCenter period:
As we can see in the table, during this period, the average user number drop to 941, compared to Pubcenter period average of 967 and AdDuplex period of more than 2000. But, surprisingly the average impressions numbers are higher compared to that on PubCenter. With an Average of 947 users, and our estimation of 3 ads per user session we get an average daily impressions of 2841. The AdMob impressions average is of 2767, for an awesome 97.4% Fill rate. This is the best fill rate of the three ad networks. Almost every time a user asks for an ad, it receives an ad. Also we can see a higher eCPM rate in the AdMob network. In average we get 0.07€ eCPM on PubCenter. In AdMob we get 0.09€ average eCPM. If we combine the higher eCPM with the higher fill rate, we can see our average daily revenue climbing from 0.11€ in Pubcenter to 0.30€ in AdMob. We are talking about nearly 3x performance in our revenue. If we were using AdMob during the AdDuplex period, when we had 3x more users than now, our revenue could have been amazing in that period.
Some conclusions and disclaimer
As we reach the end of this little “hobbyist” study, I'd like to put in clear text some conclusions:
- PubCenter is the easy and most robust ad provider SDK to integrate in your apps, but it’s weak if your target audience is outside the USA or Canada. Even in USA or Canada, the eCPM is so low it barely gives you money enough to justify the inclusion of ads in your app.
- IAP is great as a reward system so your users can pay for the app if they feel you deserve it. At the same time your app is free and is listed in the free category, where the vast majority of users look for apps. If apps with trials can be listed in the free apps section, it could be equally as good as IAP.
- AdDuplex is awesome for first launch apps, to get new users and spread the word about your app, but with high numbers of downloads you don’t get a real “revenue” for your inversion. Also AdDuplex could be great in combination with PubCenter to supply ads when PubCenter doesn’t return anything, using a control like ad rotator.
- If you want to get the most revenue in a global market with a Windows Phone 8 app, today AdMob is the way to go. Better fill rate and better eCPM led to better daily revenue. It is really sad they haven't released AdMob for Windows Store apps.
Also, I want to state clearly that I’m not an expert in statistics and I don’t have the only truth. The concepts expressed here are my own opinions and experiences based on one experiment I made in the last month. Maybe for someone PubCenter is better, maybe someone gets more new users in AdDuplex or maybe you don’t want to use Google products. My only intention here is to show my experience and be clear with my app data so you all have another point of view when deciding how you are going to monetize your application.
Hope this little study can help with your monetization strategy.