A notice to users of my apps, starting with KEYBOX 2.2.1, Accent 1.0 for iPad and Accent 1.1 for Mac I am collecting install base statistics.
The Story
Recently I had difficulty reconciling the rankings of my Accent app in the Mac App Store and the sales results in iTunes Connect portal.
On its opening day Accent was ranked #1 in the Mac App Store’s Reference category in Japan, even beating out renown apps like Delicious Library 2 among a whole host of others.

However, when I saw the sales numbers I was shocked. I expected to see hundreds of sales but I only got 2 in Japan. Yes, two. I let it slide for a day thinking the results would be reflected by the next day. No such deal.
I approached Apple support to ask which wasn’t working, their ranking algorithm or their sales tallying. I’ve had apps rank #1 in a different content aggregator before and know what to expect. Surely an app with only 2 sales could not have reached #1 placement I thought.
Following this thought, an app at #1, even if accidentally placed there, should see a flood of sales the next day because of its high placement. This flood never happened.
Apple gave me the below response that ensured their ranking algorithm was in working order but neglected to answer whether the sales count was correct.

Since this time I’ve come to understand a bit about how Apple ranks apps and have arrived at the unescapable conclusion that either their system is broken or the Reference category under which Accent was categorized is a near ghost town. It’s not in Apple’s best interest to tell me which is the case of course.
This isn’t the first time I’ve had issues with the slightly buggy iTunes Connect portal but every time I approach Apple about errors/discrepancies I go in not knowing anything. The whole portal is a black box to me and to every other developer.
I’ve decided to arm myself with information from now on! Introducing the statistics collector!
How Statistics Are Being Collected
This is how it works: My server is notified each time one of my apps is executed for the first time or when it is upgraded. Just to be absolutely clear, this notification contains ZERO private information. I don’t see anything that would identify you.
So what does get sent?
- app name
- app version
- device type (iphone, ipad, mac)
- os type (ios, osx)
- os version
- locale
- first execution time
- is upgrade? (Y or N)
That’s it! No UDID. No snooping around in your address book. No thumbing through your photos.
Here’s the web request showing me running KEYBOX 2.2.1 for the first time on my iPhone (fresh install):
GET xxxxxxxxxxxxxxxxxxx?app-name=keybox&app-ver=2.2.1&first-exe=20120514072247&dev=iphone&os=ios&os-ver=5.1.1&loc=ja_JP&upgrade=n HTTP/1.1
As you can see, there is nothing private or sensitive here.
What I Learn From These Statistics
This system isn’t perfect. It doesn’t accomplish my goal of knowing how many sales I made. It only tells me when my software is installed on a user’s Mac or iOS device. For instance, users who purchase KEYBOX on their iPhones are also entitled to install it free of additional charge on their iPads and iPod touches. In this case I’ll see 3 installs for one purchase. In fact I won’t even be able to tell if the same person is installing KEYBOX in all 3 devices.
However, if I see 1000 installs but only 5 sales then I know something is not right and can approach Apple with this information. After all, nobody has 200 iOS devices at home. This system is about seeing such wide discrepancies.
In Closing
I have updated the KEYBOX privacy guarantee to reflect the new behavior mentioned above.
I appreciate everybody’s understanding in this matter and I hope I’ve addressed any concerns you might have by this change.
If you have any questions please do no hesitate to contact me at jay [AT] this here domain.
Thank you in advance,
Jason Fuerstenberg


My Flow:






