Recently, you may have noticed an issue with the new AMO stats dashboard. When designing the new pages, we wanted to reduce the noise level from mislabeled or malformed data. This was especially necessary for our “Usage by Application” data, which contains many applications we do not recognize.
To accomplish this, we filtered the raw data to only index versions of applications we were aware of. Unfortunately, the data source for these applications was incomplete, so when we indexed the data into the new system, we missed out on a large amount of legitimate data. Included in this data were the minor versions of Firefox, such as 4.0.1, 5.0.1 etc. As a result, a few developers noticed abrupt changes in their usage data.
We’ve relaxed the filter to be more permissive, and will be re-indexing statistics data for all add-ons to increase accuracy. Developers can expect to see a noticeable increase in their “Usage by Applicaton” statistics, where the sharp drop-offs shortly after each major version will now appear as a smoother curve. Total add-on usage numbers should remain unchanged as a result of this re-indexing.
Please let us know if you have any questions or concerns, and apologies for any inconvenience.