Flawed methodology: The approach considers only services Google has closed. There are fairly obviously flaws in using this to predict how long you can expect a newly-introduced solution to be around.
These kinds of posts (Google Reader being shutdown, how long will XYZ Google application run for) are starting to get old and redundant...