In case you missed last week’s announcement, the AMO team recently re-launched its stats dashboard. If you want to know how the universe of Firefox add-ons has been growing over time (e.g., via downloads or usage), you can easily adjust the chart to see the visualization you’re interested in. You can also easily export the full data set. Check it out!
When we left off, we showed that the #1 reason why people refused the Firefox Major Update offer was their frustration/preference/confusion related to the awesome bar (or location bar). There are a few things interesting with this piece of feedback and with all the feedback more generally:
Returning to our cohort of users not fully satisfied with or fully comfortable with the awesome bar, we’ve released a series of enhancements in Fx3.5 that should resolve their concerns. Alex Faaborg provided an extensive walk-through, and here’s one key section that should help this group of users:
Ability to Control What Appears in the Location Bar Search Results (Proactive Privacy)
When we expanded the capabilities of the location bar to search against all history and bookmarks in Firefox 3, a lot of people contacted us to say that they had certain bookmarks they didn’t really want to have displayed. In some cases users had intentionally hidden these bookmarks in deep hierarchies of folders, somewhat similar to how one might hide a physical object. Having something from your previous browsing displayed to someone else who is using your computer (or even worse) to a large audience of people as you are giving a presentation, is really one of the most embarrassing things that Firefox can do to you. So now in Firefox 3.5, users have complete control over what types of information are displayed in the location bar (or suggestions can be turned off entirely):
So, what actions items can come of this user feedback?
If we think this concern surrounding the location bar is solely limited to existing Fx2 users, we could consider some special messaging for them. However, I would venture to guess that some small fraction of the Firefox user base currently on either Fx3 or Fx3.5 (93% of all users) share similar thoughts about the location bar. If so, then it would likely be worthwhile for us consider a couple options:
• Modify Firefox itself.
Below is a simplified mock-up of just one idea I came up with (please note that I don’t know anything about UI). I’m not sure if the average user knows about Tools->Options, so the concept here is that a person could easily understand how the location bar works and adjust it (according to their preferences) within a fraction of a second.
• Up-level messaging about the location bar and its latest enhancements.
Utilizing key touch points, such as the firstrun page, whatsnew page, or firefox.com page, could help users feel more comfortable with the location bar and its awesome functionality. For example, is it more important for us to be using these critical touch points to promote open video formats and Fx add-ons, or to highlight Firefox’s most utilized feature and how it can best help a user in his/her everyday life (and not frustrate them)? At the very least, this seems like a question we should be asking.
What are your thoughts?
Back in May, we launched one final push to migrate Fx2 users to Fx3 via a major update (MU). At the time, 10% of Firefox users were still on Fx2. If you’re curious about the results of that push, that 10% was down to 8% a couple weeks later. Aside from these results, there was one interesting twist integrated into this MU offer – people seeing the prompt below were also asked if they wanted to provide some feedback if they were choosing not to upgrade
(Please ignore the Fx version numbers in this screenshot. Unfortunately, we don’t have an exact copy on hand.)
For people interested in leaving their feedback, here is the survey they saw:
In total, nearly 5,000 users provided feedback. Similar to our other recent feedback efforts, we decided to manually read through comments from the second question above in order to extract the most insight and best understand the true pain points and thoughts of these Firefox users. Sorting through the comments, I noticed about fourteen different types of comments/issues that were cited multiple times… a distribution breakdown and example comments are below.
Location Bar (clearing history) = “There are several issues with Firefox 3, but the browser is not very secure if browsing history cannot be deleted from the location bar. The Tools->Clear Private Data command was no help.”
Speed/Memory Leak Concerns (real or perceived) = “I had heard it made your computer run slowly when it first came out….but after a few months i had forgotten about this tid bit and installed ’3′. It did indeed bring my computer to a near standstill. My friend did this as well. Another friend then jarred my memory as to why and so…i then un-installed it and re-installed ’2′ and since then my computer has been running great….so….would you take the chance again?”
Add-ons Compatibility = “There are a number of add-ons which will not work with Firefox 3, Getmail, Silverscreen, Feebe and others.”
Look & Feel (general UI) = “Compared to v2, v3 has an interface that sucks, sucks, sucks big time. Graphically, it’s much worse: harder to read (too dark, type too small), icons lack color, etc.”
Bookmarks (UI) = “Getting to alphabetized bookmarks with the latest version is a serious pain in the fundament and my wife and I continue to be amazed as to why you decided to take a huge step backwards with what was supposed to be an advance. Cure this problem and we’ll move up. Thank you.”
Crashes = “Firefox 3 Crashes every time I install it. Even with a clean upgrade. It won’t load at all it just crashes restarts and crashes again.”
Lost Stuff when Upgrading = “Last time I took a Firefox upgrade, my entire Favorites list was wiped out. It would take me more than a year of frustration and work if that happened again. So, I’m afraid to take another upgrade.”
Developer for Testing = “I am Web developer and have to test my page in all browser. So I am keeping ff2 alongwith ff3 and other browser .”
Don’t Fix What’s not Broken = “I need more time to figure out what to do. I am happy with all the stuff I have and the status quo. My computer works well as it is. Do I really need this update? ”
Fonts (size) = “The window is huge. the writing is huge. when I go searching only 2 or 3 websites fit in the window.”
U3 Compatibility = “Firefox 3 does not work on U3 devices in fact it makes the device unusable and so does the latest version of Thunderbird. The U3 device only accepts the mobile versions although I keep getting upgrade versions popping up.”
Cost? = “If you say this is free. I have always heard there is really nothing free in this world. Times are tough all over, especially for us seniors living on social security.”
Printing Problems = “When printing from FF3, text comes out as scrambled ASCII. doesn’t work on Win(XP/Vista), Linux, etc. Big pain in the butt. I love FF3 otherwise, but as a network admin I just can’t make the switch.”
Enterprise Support = “We can use only software expressly approved and that works together with our network and other older software. We can’t download or install anything on our own.”
In our next post, we’ll provide some interpretation around these findings and discuss possible implications from a product standpoint. Stay tuned for Part II!
Lastly, thanks to Sam Sidler for implementing this survey and to Eric Hergenrader for providing much of the analysis.
Let’s start with the punch line – roughly 50% of people arriving at Firefox Support (support.mozilla.com) do so accidentally. And yes, cats are a common culprit! Below is the story of how we arrived at this insight.
The metrics and SUMO teams have been working hard to better understand the traffic at support.mozilla.com and users’ experiences there. One critical item we’ve had trouble wrapping our heads around has been the unusually high bounce rate on the SUMO home page. Looking strictly at the web analytic numbers, roughly 86% of visitors immediately exit the site.
As a multi-pronged approach to figure out what the heck is going on here, we considered a couple options:
For approach #1, we initially implemented a button in the bottom corner of the page (button and feedback form are pictured below). This page sees nearly 10,000,000 visitors monthly, so we were fairly sure that we would be able to understand some key insights rather quickly. Unfortunately, events did not play out as expected. Over the course of nearly a month, we had fewer than 1,000 total responses, which translates to a click-through rate of roughly 0.0001% (ouch!). To boot, there didn’t appear to be any revealing patterns or insights within the responses.
At this point, we felt bewildered and didn’t know where else to turn. So, what did we do? We implemented a pop-up.
Let me stress that I hate pop-ups, hate them. Given our circumstances, we decided to run the feedback pop-up question for just a 24-hour period and show it to 60% of visitors to the SUMO home page (they saw it when trying to leave the page).
The findings were amazing. Firstly, out of 700 total comments, only one or two people said something negative about the pop-up itself. Many comments were along the lines of, “Firefox is great” (someone even pointed out that we should have included a positive category within the survey form). Here’s one example:
“i admire this site and this browser. it is faster, and safer,…….. so many add ons that can use and its fun…… i think this is the best internet browser that i have…. i using ubuntu 9.04 and its work fine…tnx more power to you…. and to your team….”
Secondly, we were able to confirm a hypothesis – nearly half of the traffic to the SUMO home page appears to be accidental (e.g., a user accidentally hitting F1 on his/her keyboard). The screenshot below comes from Kampyle’s dashboard (they’re the feedback analytics firm who helped us with this project) and it shows a breakdown of people’s feedback by category. That blue slice of the pie is “Arrived by Mistake”… that’s huge! Even though it shows as 40%, we’re guessing that the true percentage is closer to 50%, as some folks arriving by mistake likely closed their tab/window within a fraction of a second and missed our feedback question.
[As a side note, this is a perfect example of why pure web analytics – alone – can fail us (analysts, marketers, business managers, etc.). Without hearing real comments from real people, we could have easily been led down the wrong path by relying solely on the data.]
Lastly, we wanted to look at the exact feedback left by people visiting support.mozilla.com by mistake. The vast majority indicated that they indeed realized that they had accidentally hit the F1 key. Diving in a little further, we wanted to share a small sample of actual comments. Enjoy!
“actually the cat walked on the keyboard”
“bent over to pick something up and accidentally leaned on the f1 key”
“Cat jumped onto laptop”
“CHILD GOT A HOLD OF THE KEYBOARD AND STARTED PUSHING BUTTONS.”
“finger too big. finger hit f1 by mistake. stupid bigfinger.”
“i was dusting, its nice to know this site exists.”
“I was stretching in my chair, and next thing I know, I’m on the FF help site. OOPS!!”
“I was using a netbook in the dark and thought I was pressing 1 but I hit F1 by mistake. Aside from that I’ve usually found the firefox support site quite helpful on the rare occasion that I’ve needed it.”
“my 2 year old was playing with the keyboard”
“My 9 month old was playing with my keyboard.”
“My baby likes to grab my computer and press buttons. She grabbed my keyboard and held the F1 button down so the Support site opened in about nine tabs.”
“My cat stepped on my laptop.
”
“My daughter is always pressing it. I never knew which button it was before now.”
“my fat finger syndrome is acting up again”
“My nails are long and I constantly hit the F1 button.”
“now I know which key my baby hit when he was banging on the keyboard.”
“Oops. Accidentally hit F1, but wow! The support site is great! Lots of info. I may have to come back intentionally! Thanks for a great product!”
“the key was hit by the telephone”
“Yes, I arrived at this site by pressing F1 key. No matter how I came here but would like to let you know – I love Firefox and thanks for provinding us this great product.”
“cat hit keys”
“Cat on keyboard”
“Dropped food on keyboard”
“my cat hit the help key….lol”
“My cat pressed F1, but it’s cool”
“my pet rat walked over my keyboard and hit the F1 Key LOL – Sorry!”
“No clue–my cat was wandering around and probably stepped on something but it was nice ending up here because I’m seeing all kinds of things that look very interesting.”
“my son was hitting some random keys on my laptop and suddenly I get this.”
There’s a relatively simple question we’ve been asking for quite some time… of the total Firefox user base, how many people (or what percentage) use at least one add-on?
Thanks to some recent work by Simon Krueger, our metrics intern, we’ve been able to arrive at a rough approximation. Here’s my thought process for the important data points Simon’s been able to extract:
I want to conclude with a few words of caution. This methodology is far from perfect. It’s conservative and there are also potential holes in it (some of which we haven’t yet thought of). It also remains unclear as to how useable this slice of data will be as we see it trended in the future. Moreover, the data above is wide open to interpretation. The add-ons team and other folks across the community may arrive at a completely different interpretation (read here for Justin Scott’s point of view). Regardless, a fruitful discussion around this topic should hopefully ensue.
(photo from mynameisideal.blogspot.com)
As mentioned a few days ago, we saw a massive impact to the Firefox new user experience with a recently implemented change to the Firefox installer. A problem previously encountered daily by thousands of new Firefox users completely vanished with a single change. We now want to share a few more details and suggest one or two additional changes moving forward.
First up, let’s look at the distribution of users’ comments — by step — in the Firefox installation process, i.e., on what step were they on when hitting the “cancel” button (a visualization of these steps is shown here). The screenshots below come courtesy of Kampyle, the start-up who helped us with this initiative.
THEN (late March):
NOW (last week):
The difference between these two pictures is night and day. A couple things stand out:
Looking at the actual comments from users canceling out of the final step, about 75% are related to the zombie process problem (a workaround solution is in the works), about 20% are people expressing their unhappiness over Firefox becoming their default browser, and about 5% fall into an “other” category. You’ll recall that those two issues were the top issues highlighted by users overall.
What’s up with people commenting on Firefox becoming their default?
It’s a simple option to check or uncheck as desired, right?
Diving into the comments reveals something critical. All of the people making this comment are doing so on the “Summary” step (or final step, below), not on the earlier “Setup” step (above). Clearly, many new users are overlooking the checkbox, and just seconds later, are unsure as to how to install Firefox as their non-default browser. And the outcome here is poor – the person is frustrated and refuses to install Firefox.
What fix should we consider here? Should we make the checkbox more prominent on the “Setup Type” step? Should we include instructions (e.g., “simply hit the back button to change”) on the “Summary” step?
Thanks to Rob Strong, the Funnelcake team, and the folks at Kampyle, among many others, for making this ongoing initiative a success.
It has now been a full month since the launch of Firefox 3.5 – what impact has the new browser had?
In just over 30 days, Firefox 3.5 has been downloaded 47.6 million times! This does not include downloads from users who go to (in Firefox) Help->Check for Updates. Looking at the conversion rate from user download to capturing that download as a daily user, we seem to have almost regained the conversion of Firefox 2.0:
This is especially impressive when noting that we are in the middle of July in which ADU numbers are most likely depressed from events like schools on summer vacation. Firefox 2.0 was released in October, traditionally a stronger month than July for Firefox.
