Discrepancies in Microsoft Partner Center Acquisition Reports

Some time ago it has come to our attention that data in Microsoft’s Partner Center (formerly Dev Center) Acquisitions reports doesn’t seem to be realistic. Specifically, the “Page views” numbers.

Customers noticed that the number of clicks they get on their ad campaigns differs by order of magnitude from what is reported for page views in the Partner Center. And these numbers have to be pretty much identical (minus a tiny percentage of people who cancel navigation mid-redirect, etc.)

We did our own experiment, and for a campaign that got more than 11,000 clicks, only 506 page views were registered in Partner Center. That’s more than 20x difference!

page-views-by-channel

The table right underneath this chart gave us the first clue of what could be wrong:

page-view-conversions

As you can see, according to this report it seems that our test app is enjoying a whopping 44% click-to-install ratio. Any user acquisition professional would sell their soul for a rate like that. An industry standard is somewhere in 1 to 5% range. Based on our click data (11,000), and assuming that Microsoft’s conversions number is correct (224) we get a click-to-install ratio of around 2%, which is in line with what one should expect.

After additional experimentation, we were able to verify that the cause of this massive discrepancy is a bug in Microsoft’s Store that only counts page views when a user clicks that “Get” button in the web store and proceeds to the Store app. This is obviously logically incorrect, as this click on “Get” is a clear intent to download – hence the magical 44% click-to-install rate.

Not only this is incorrect from the common sense perspective, but it is also in contradiction with Microsoft’s documentation (emphasis mine):

A page view means that a customer viewed your app’s Store listing page, either via the web-based Store or from within the Store app on Windows 10.

We have reported the issue to Microsoft some time ago but are yet to get confirmation and an estimated ETA for the fix. We will update this post once we hear some additional information and/or when the issue is fixed.

In the meantime, keep in mind that the page views number is inaccurate and either use an intermediary (like an URL shortening service) to track visits to your app’s page in the Store or just exclude these numbers from your campaign effectiveness assessments.

AppRaisin: Time to say goodbye

On November 17th, 2015 we’ve launched AppRaisin – an app and a service to help Windows 10 enthusiasts discover new apps and games. On August 10th, 2016 we decided to stop active development for reasons I described in a lengthy blog post. On May 6th, 2019 AppRaisin’s story will come to an end…

Since we stopped development the usage went down slowly but steadily. Yet, the distribution between mobile and desktop users didn’t change much – AppRaisin is still de-facto mobile-first app even though Microsoft’s Windows Mobile platform is not really an active thing anymore. We tried to maintain and moderate the service for as long as it provided value for the community and wasn’t a big strain for us. In recent months it just doesn’t feel like we are providing an adequately sized value for the effort.

Even though our efforts were limited in recent years, AppRaisin managed to maintain a 4.8-star rating in the Store which means that many people still find value in it that the official Microsoft Store and 3rd party websites can’t satisfy. So, while this decision was on our minds for more than a year now, we knew that it will upset some people and just couldn’t find the courage to pull the plug. It’s time though…

I would like to thank everyone who worked on AppRaisin over the years, everyone who used the app, submitted and upvoted app news, rated it in the store and provided valuable feedback, every developer who made great apps that enabled AppRaisin to have content, and every person and media outlet who helped us spread the word about our little project.

Thank you!

Alan Mendelevich.