Friday, June 21, 2013

A #GoogleReader Puzzle From 2010

A Google Reader Puzzle From 2007: Back in 2007, Google accidentally made public an internal video about Google Reader. The video included a lot of useful information about Google Reader, the kind of details you'll never find in a Google post or presentation.

I'm trying to solve the puzzle and find the number of Google Reader users from 2007. Here are the hints:

1. At that time, Google Reader crawled 8 million feeds.

2. Two thirds of the feeds had only one subscriber, one third of the feeds had more than one subscriber.

3. Google Reader used 10TB for storing all the raw data.

4. The rate of user growth = the rate of growth for the number of feeds.

5. The index size grew 4% every week.

6. 70% of the Google Reader traffic came from Firefox (at that time, Firefox's market share was about 15%).

7. Gmail and orkut were the only Google applications that had a bigger number of pageviews/user than Google Reader.

8. The main Google Blog had 100,000 subscribers and this number includes all the subscribers from iGoogle, Reader and orkut.

So, at that time, only 2.6 million feeds indexed by Google Reader had more than one subscriber. Probably many of the feeds that had only one subscriber were used by other Google services powered by Google Reader's backend (orkut, Blogger widgets, Google Spreadsheets, Ajax API) or they were Google News/Blog Search feeds. It makes sense to assume that the number of Google Reader users was lower than the number of feeds with more than one subscriber.

Here's a chart from 2010 that shows Google Reader's user growth:

And something else from 2010: "the average Reader user reads about 105 items a day".

It's probably worth distinguishing Reader as a service from Reader as a platform.