If you don’t want to read the rest of the post, suffice to say that Tapefailure rocks my socks off!
Tapefailure, like any other web analytics service, lets you understand your site’s traffic. What makes it special is the fact that it records user sessions and replays them to you!
So instead of showing you aggregated statistics about where your visitors came from, what keywords they searched for to reach you etc., TapeFailure shows you what users did when they landed on your site. Not every single user - right now, TF can record at frequencies of 1 in 2, 3, 4 or 10 visitors (i.e. 50%, 33%, 25% or 10% of visitors are recorded).
These recorded sessions are called “tapes” (scroll down for a sample tape). Apart from the tapes, TF also gives you a variety of charts and visualizations to help you understand how visitors behave on your site (screenshots below).
Irrespective of whether or not this is useful, it is incredibly cool!
My experience - Screenshots & Sample tapes
I setup TF on this site almost immediately after I got my invite to try the private beta. I embedded a Javascript into my site and immediately, I had my first tape! Click below to check it out:
Funky isn’t it? The Tape player is a small piece of javascript that scrolls around a cached html version of the website. The script also shows a nifty mouse pointer that runs around duplicating the mouse movements of the website visitors. If you would like to take a look at the scripts,
here is a downloaded version of a tape.When you login, TF displays a summary of the tapes captured up until that point. This screen looks like this:
How useful would such data be? I mean, what do you do with hundreds of captures a day? Well, there are filters to classify users into different types so that their behavior can be studied. Here is a sample filter screen:
Filters
So now you can filter out all the users who spent less than a minute on your site and understand what they did before leaving (and perhaps, figure out why they left).
So those were the individual tapes. The next screen shows aggregated information about mouse movements, mouse pauses and clicks on each page that you are tracking (this is filterable too).
I didn’t understand my “heat map”. First, how is it generated? Is it based on mouse movements and scrolling behaviour? My heat map was uniform throughout… which is impossible, obviously most people would not have scrolled to bottom portions of this page. This might be a bug.
And a whole slew of other aggregate visualizations:
And finally, here is the settings screen that lets you set up the trackers. Very intuitive.
Settings screenIn summary… what I like and what I’d like to see
TapeFailure looks great. It is extremely intuitive (just look at the screenshots)! The aggregate user behaviour filtered by multiple criteria seems useful.
To-do:
1. As the FAQ mentions, it would be cool to have more filter criteria (perhaps search engine source, time of day etc.).
2. I’m not sure how the individual tapes are useful! Individual tapes are not representative, so why not just have the aggregate behaviour tracker?
3. The player does not have a timer (how much time has elapsed and how time is left on the replay).
4. The visual attention map maybe buggy. Not sure how it calculates attention information.
5. For now, TF is really fast. But this is with 150 users, hopefully it’ll scale nicely.
btw, for those who are interested, this is the prospective pricing plan:
Technorati: TapeFailure, review
.