Why don’t you use Firefox and can IT change that?
At the beginning of the quarter I set an ambitious goal of trying to look at ways IT/Ops can reduce Firefox adoption/retention barriers.
I kept rotating on the best way to frame this question and in the process probably kept over thinking it. Mentally I tied this goal into looking at under-served geographic regions like South America (Brazil) and Asia-Pacific. I’m trying to answer these questions:
- How can IT help drive adoption/retention in Asia-Pacific or South America?
- What sort of local IT resources does it take to move the needle?
- If I put a data center in Brazil or Singapore, does it help? Does it move the needle? Do users care? (What’s my ROI?)
In other words, is there some function of Mozilla’s network or server infrastructure that prevents users in far away geographies from using Firefox?
I’ve been working with Asa Dotzler, Ken Kovash, Seth Bindernagel and Staś Małolepszy to try to figure out the right questions to ask and how to get them asked.
We came up with several points to explore:
- Web page load time for Mozilla properties (
www.mozilla.com/www.mozilla.org) & other “participatory sites” (planet.mozilla.org,bugzilla.mozilla.org,developer.mozilla.org,labs.mozilla.com…) - How easily and quickly it is to get Firefox support using
support.mozilla.com(is it too slow to be useful?) - Interactive browsing on
addons.mozilla.org(is this site too slow to use?) - Performance of AMO updates to extensions inside the Addons manager.
I had come up with three survey methods:
- Installer survey
- Uninstall survey
- Community surveys
The Installer survey is interesting but I’m not convinced that helps answer my questions. At the point of that survey, the user hasn’t really used Firefox and any of those websites. So I scratched it.
My goal for Phase 1 is to target the Uninstall survey and relate the responses to the users geography.
Looking for suggestions – followups in comments or email.
Comments (4)
I think you should convert the four topics above into four questions with a scale of responses that can be tested with some commonly used statistical tests. Your users can assess their experiences by rating each topic that’s interesting to you.
Here is how I would phrase each, with a scale of 1-5 where 1 is very poor and 5 is very good
1. Please rate the Web page load time for the following Mozilla web properties/pages (if you’d like to click on the links now to test, please do so and give us your rating)
– http://www.mozilla.com/www.mozilla.org
–planet.mozilla.org
–bugzilla.mozilla.org
–developer.mozilla.org–labs.mozilla.com
2. Mozilla’s Firefox Support is located at http://support.mozill.com Please tell us about your experience by rating the following
– ease of finding support you need
– speed in receiving support
3. Mozilla’s addons.mozilla.org Website is an interactive Website that provides users thousands of browser add-ons to customize their browsing experience.
–Please rate the speed of the interactivity of the site.
4. Add-ons that customize your browsing experience are updated through the Add-ons manager in your Firefox (link to a screen shot for reference?)
– Please rate your experience with the performance (i.e speed) of the Add-on Manager
===================
With each of these questions created with a scale of responses, you will be able to get a sense of how your users rate performance of each element you are testing.
You can ask a demographic question at the end (or set of questions) to assess who is taking the survey.
Finally, it would be most helpful to make sure you have a way to collect where the surveys are being taken. It goes without saying, but this information gathered above is helpful, but for your purposes, it would be useless if you couldn’t map it directly to a geographic location so you can then propose what can be done to improved the experience.
another thing to consider: what is the target sample for this survey?
if you were to only serve this to those who went through the uninstall survey, you would likely have skewed results due to an unhappiness factor. on the flip side, if you sent this to our locales (even the smallest ones), you’d likely have a sample of more happy users.
i think you want to start in geographies where you think Mozilla infrastructure investment will help them. then, consider offering to a sample that gets all three sets of users: happy firefox users, unhappy users who uninstalled, and users surfing our sites for support (i.e. not unhappy, but also looking for help because something is going wrong).
maybe you can offer the survey in three ways:
1) specialized builds to locales of interest
2) uninstall survey
3) pop-up survey to those surfing SUMO or AMO
Here’s a suggestion…
It looks like most of your questions are page/site specific. Why don’t we just ask people directly on those pages/sites? There are unobtrusive survey tools, such as Kampyle (www.kampyle.com, we’re using their “software” service for our installer survey), and then there are more comprehensive site feedback services such as 4Q (http://4q.iperceptions.com/).
It seems like asking people for feedback at the very moment they’re on a page/site in question should be far more productive than leveraging the uninstall survey (or other methods casting a wide net).
I’m all for making our websites faster, I’ve been working on it for AMO and other sites. Big sites like Amazon and Google see purchases and searches (respectively) go down when latency or page load times increase.