The early days
Since the early days of the iTunes App Store, there has always been a free-to-paid loophole in the popularity ranking algorithms that allowed paid applications to appear to be more popular than they were. When the App Store first opened, developers could send their apps to the top of the top-100 list by gathering sales as a free app and then changing their app to a paid app. All of their sales as a free app, which were typically 10x-20x of what they would have received as a paid app, were included in their total sales. The top 100 list was created based off of total sales. Apple initially corrected the list manually, but eventually the use of this loophole become so prevalent that Apple had to change their popularity algorithm.
Separate Paid and Free Top 100 Category Lists
In December of 2008, Apple reorganized the App Store and created separate top paid and top free lists for each category. This reduced the incentive to use free app sales to increase popularity. However, it did not eliminate it. During searches for apps, the popularity meter would still show popularity based on free app sales. This is a good time to mention that the top-100 popularity algorithm is different than the algorithm for the popularity meter that shows up when searching with iTunes. Without going into details, the top-100 popularity algorithm gives a much higher weighting to recent sales. This allows new applications to more easily break into the top-100. However, it is not entirely based on recent sales, and that allows apps that had a boost of popularity to remain on the list longer than their current sales would otherwise allow. The popularity meter does give some additional weighting to recent sales, but much, much less.
Tweaking the Popularity Meter to Favor Paid Apps
Although the top-100 sales lists were now fixed, there was still a problem. Any search of the App Store would show free apps with a much higher popularity than paid apps (using the popularity bar). Around March of 2009, Apple addressed this by knocking the popularity of free apps down by about 90%. A developer with a paid app and a free app would see both apps with the same popularity even though the free app had 10x the number of downloads.
The Latest Change – Fixing the Popularity Meter to Favor Paid SALES!
Some time in the last 36 hours, Apple has changed the popularity meter once again. The change affects applications that accumulated free sales and then switched to paid. These applications previously enjoyed an increased popularity score due to the fact that their sales, while free, were not discounted since they were now considered a paid app. With the new change, now only an application’s paid sales will significantly contribute to the app’s popularity meter. This mirrors the change made to the algorithm used to determine the top-100 lists. Now, it is no longer important whether an application is free or paid. Instead, it is important whether sales were made while the app was free or paid. Popularity of an app can be summarized (ignoring the weighting that favors more recent sales) as:
Popularity = (number of free downloads)*10% + (number of paid downloads)
The Next Change – Weighting to Higher Priced Apps
Without getting myself into any trouble, let me just say that the next, and likely final, change to the App Store ranking system will be to incorporate pricing into the popularity algorithm. Weighting paid app sales differently than free app sales is only a preliminary step in implementing full priced-based weighting of app store sales. This will show up both in the popularity meter and in the top-100 lists (although the effect on the free top-100 lists will be minimal, for obvious reasons). With this change, the average application price will triple and the days of the $0.99 app will be numbered. $1.99 will become the new low price for developers wanting to find the ‘sweet spot’ (or is it the ‘sweat spot’ – check out our upcoming app WhichWord to find out) that optimizes popularity and sales (and revenue).
References
To see an example of the changes in the App Store popularity ranking, observe the popularity of all of the applications by IMAK Creations for May 26 and May 27, 2009. On May 26, Hold On! and Brain Blaze Divide, were both high in the rankings. On May 27, Hold On! dropped by half and Brain Blaze Divide dropped to the bottom of the list. Both of these apps sold most of their sales as free applications. ColorTilt, which has never been free, held steady at the top. This caused recently released That Ain’t It! to make it into the top 4 and be featured with an icon shot. Nice!
Shows how free app sales of Hold On! and Brain Blaze Divide boost pop
Shows popularity change of Hold On! and Brain Blaze Divide
IMAK Creations 