User: Password:
Subscribe / Log in / New account

How would it help?

How would it help?

Posted Feb 9, 2012 17:55 UTC (Thu) by scripter (subscriber, #2654)
In reply to: How would it help? by NAR
Parent article: Tracking users

I'd like to know how it well help as well. Measure the wrong things, and people will come to the wrong conclusions.

I don't see how they can measure the conversations people have, where one person recommends chrome to another.

Mozilla needs to reverse the now-outdated perception that Firefox is slow. Firefox has improved significantly since Chrome made its splash.

On the Linux desktop side, the perception problem may stem from using enterprise distributions where a Firefox 3.6 package is still the standard.

(Log in to post comments)

How would it help?

Posted Feb 12, 2012 16:28 UTC (Sun) by deinspanjer (guest, #82864) [Link]

If we have this data from installations and we can detect when data has stopped coming in from one of those installations, we can review the configuration, usage, performance, and stability characteristics of those abandoned installations. This is called retention analytics. You can use it to answer questions such as the following:

What was the breakdown of abandoned installations by age of installation? Were they installed, used, and abandoned or were they used for a long time before being abandoned?
What was the breakdown of abandoned installations by usage intensity?
Were they used once every week or every day?
Were they actively used throughout the day or started up once, used for 5 minutes (maybe to test a webdev bug) and then shut down?

Part of the problem right now is that we have a very limited understanding of the composition of installations that contributes to our active install-base. We don't know what portions are regularly and actively used, and what portion might be installed by webdevs for occasional compatibility testing. There are a lot more possibilities than those two, but they help demonstrate how we could be trying to solve the wrong problem if we don't understand the situation.

Do certain add-ons or groups of add-on contribute to abandoned installations?
Do certain add-ons or groups of add-on "prevent" the abandonment of installations?

It is likely that there are certain add-ons that provide such a practical value to large portions of our user-base that those installations are less likely to be abandoned.
It is also likely that there are certain add-ons that either cause a bad user experience or otherwise cause slowdowns or stability issues. We are constantly hunting these issues down, but we can really only get data from people who take the time to visit our support site or otherwise reach out to us. There are likely lots of users who just give up and go away. If we can learn something about those installations then we can better deal with those issues.

What are the performance and stability characteristics of abandoned vs active installations?
What are the performance and stability characteristics of installations by install age?

Users who abandon installations likely have some good reason. There are some reasons we could never hope to understand such as a friend encouraging them to try out an alternative, but the more we do understand, the better we can focus on improving Firefox in those areas.

How would it help?

Posted Feb 13, 2012 21:31 UTC (Mon) by robbe (subscriber, #16131) [Link]

According to version 10 will be the next release with a longer┬╣ support period. So the more stable distros will very probably pick up that.

┬╣ One year still seems pretty short to me.

Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds