I guess privacy has been thrown out the window by Mozilla. I first noticed this when checking traffic log at work. It’s bad enough that Google has to make a random https requests (*-*-*-*-*-*.metric.gstatic.com:443) to gather info of every browser visiting their search page or related, now Mozilla is trying to be sneaky too.
Any time you visit your homepage / Firefox start page, even if it is about:blank, snippets-stats.mozilla.org aka the Snippets Impressions Metrics Service (http://snippets-stats.mozilla.org/) is called. I have found out how and why they are doing this on the mozilla.governance group. (https://groups.google.com/forum/?fromgroups=#!forum/mozilla.governance)
The proposal by Winston Bowden is linked here, all of the quotes below were created by him. (https://groups.google.com/forum/?fromgroups=#!topic/mozilla.governance/9RbsjecBaxM)
Also no optout, and the change homepage thing is bullshit.
Opt-out
It is not possible for a user to opt-out of this data collection
specifically. However, users can choose to use a different home page,
at which point snippets aren't shown and data isn't collected.
Why:
1. Comparing the effectiveness of our snippets (e.g. copy, pictures,
etc.) to industry standards (typically measured in CTR - click-through
rate). Today we can measure the clicks to our properties, but we
don't know how many impressions yield a click. This data will provide
us with the ability to analyze conversion rates and run accurate a/b
tests. For example, Mozilla runs two snippets on a similar topic. By
evaluating which snippet has the highest CTR, we can retire the low
CTR version and run the high CTR version thereby maximizing
engagement.
2. Measuring the effectiveness of the snippet delivery experience in
general -- is the snippet UX effective relative to other similar
systems?
3. Knowing how usage of about:home varies. It's likely that the
number of impressions of about:home varies over time in ways that is
not simply accounted for by changes in ADUs, changes in the product,
or the usage of the product (more/fewer new windows being opened, more/
fewer changes to the default home page, more/less usage of session
restore, etc..) Given the importance of this communication channel,
it's important to understand whether it's becoming more or less
effective over time.
4. Understanding user response to snippets over time -- do users click
on a snippet a day, or are they so excited and interested that they
would click on 20 snippets a day? Knowing the answer would dictate
how many snippets to ask Firefox to display to maximize the adoption
of new products/initiatives.
5. Understand at what point in time does a given snippet cause
fatigue. For example, do users ignore snippets in the evening? Do
they ignore once snippets have been displayed multiple times?
What is collected:
To answer these questions, the team proposes to record the following
information for each impression (meaning, every time a user sees the
Home tab):
- The snippet name (Mozilla authored, not user specific)
- User's preferred locale
- Date & time in the server time zone (precise to second - apache
standard example: 23/Apr/2012:13:15:12 -0700)
- User agent
User Data
The IP address of a user will be converted to a geo-hash. The geo-hash
will be used to understand how different regions correspond to a
snippet impression. We use MaxMind's geo-ip library (http://
www.maxmind.com/app/ip-location) that takes a given IP address and
converts it into a geo-hash.
You can review an example of what this collected data would look like
here:
http://dl.dropbox.com/u/37305222/log%20data%20-%20anonymized.pdf
You'll notice several columns of data. The first three of these
columns correspond to a user and will be used in aggregate form to
count the total number of impressions.