This was originally written early September. The patch was live late September. And 4 weeks of results later, this post. My apologies if I mixed past and present tense
The number of add-on downloads was always an interesting figure to me. Over several months, one of my add-ons was getting over 100 downloads a week without any promotion. But my ADU remains almost constant every week. It may increase by a few, but compared to the number of downloads, it only seemed like I got an abnormally low number of conversions. Less than 15% of users who downloaded my add-on turned into a ADU.
This didn’t seem right, so for 0.7 release of Safe extension, I added some code that sent me some info whenever someone uninstalled from the Add-ons Manager.
According to the add-ons dashboard, I had 78 ADU users for version 0.7 (submitted 2009-08-14. Averaged over a period of 1 week.).
How many downloads since it was uploaded? About 659 (I likely uploaded it late evening, so it might be less then that). So about 11% of the users that downloaded the extension didn’t uninstall it. But how many users chose to uninstall it using the Add-ons manager?
Ten.
So what of the other 500+ users? How did they uninstall my extension? I wondered whether web crawlers were downloading my extension. 500 seems like an awfully big number. Maybe a broken crawler? Well, actually, it turned out that AMO’s robot.txt did not exclude robots from downloading extensions at all. And how often a crawler visits is determined by the provider. Google cache takes a snapshot around once every week for me.
I realized that these numbers won’t be accurate because there are more ways to uninstall an add-on, including new profile or deleting the add-ons folder from the profile. But these seem like rare cases. Certainly not responsible for such a misleading number.
I filed and fixed a earlier bug to have robots.txt exclude crawlers from the /downloads/ path of each of our localized versions of AMO.
After 4 weeks, my stats have leveled off and the number of downloads have dropped by about 10 a day. My add-on downloads dropped from >10 to about 3 a day. This won’t be considerable to popular add-ons (AdBlock Plus shows no difference, they are still getting 80,000 downloads a day). Looking at the statistics on AMO, total downloads have dropped over 100,000 a day since the stuff went live last month. Good stuff.
October 26th, 2009 at 11:49
This is actually VERY interesting (and important).
This might explain the 10% number I referenced in my contributions post.
http://www.kaply.com/weblog/2009/10/21/one-add-on-developers-perspective-on-contributions/