working on a "Google/AWS Cloud management console" style web-app where I need to "make UI better".

I recently noticed a pattern in zocdoc app where if you click the arrow on date selector 5 times, it will show a message "you can also open date picker".

I wonder if logging events and trying to recreate simulations might give me insight into what problems users are actually facing. I wonder if this has ever been done. I did not find anything on the internet.

Would this be effective? Is anyone aware of anything like this?


It sounds like what you're describing is some form of analytics. And yes, that is a thing.

Not only traditional analytics such as Google Analytics that show you the various journeys people take on websites (what pages they go to, how they arrived, how long they stay...) but also page-level analytics, such as heat-maps (showing you the areas of the page they tend to click on) and other various stats too.

There are many tracking / analytics systems you can embed that can do this for you. Far easier than trying to log everything that happens in a log file, because while you can do that you'd still need to analyse it all, and log files themselves aren't really that user-friendly. And they get big and heavy, very quickly, slowing things down for the user (and yourself).

Here's some examples, such as HotJar, Kissmetrics and CrazyEgg: https://blog.prototypr.io/10-analytics-tools-for-optimizing-ux-886e64a6bcdd

Now, something to bear in mind:

The main issue with analytics such as this is that they only show you what did happen. There's no coverage of what didn't (so, no way to log who didn't visit, or who didn't click on various items) and also doesn't give you any reasoning behind why events did or did not happen. So you'll have to interpret it all yourself. And while the data itself is objective, your interpretation of it certainly isn't. So you should pair this with more traditional usability reviews - namely, talking to people who use it.

  • 3
    There's no coverage of what didn't . Love that part, is kind of a struggle I'm having more and more every day, trying to find out the how and why things didn't happen – Devin Feb 19 '19 at 18:40
  • Google Analytics can also track events but you need to tell it what to track. If you can get an event to trigger a JavaScript snippet, you can track that event. developers.google.com/analytics/devguides/collection/… – locationunknown Feb 20 '19 at 6:05
  • 1
    @Devin that's the kind of thing that people use eye tracking for in usability studies. You need to know if people even saw the option, etc. – Natanael Feb 20 '19 at 14:40
  • well, I have used eye tracking sometimes, but sadly is not something I can use on every project unless the client pays for that. However, while I'm not sure what do you mean so I might be completely wrong, I feel like eye tracking may help on some cases, but would also add another dimension of uncertainty on the subject of "why?". Right now I'm writing a paper and have a cogSci team working on experiments trying to find what is failing on some common research methods and why a perfect research with amazing data still fails (sometimes) – Devin Feb 20 '19 at 16:54
  • @Devin I've never found much value to eye-tracking to be honest. Other than being able to demonstrate that you've done some testing to clients (heatmaps are a good visual thing to show off) but as far as actual useful data I think they're not a huge amount of use really. Not when event tracking and other analytics are available. And actual usability testing with real people. Eyetracking falls into the 'hmm, that's cool looking' category rather than the 'hmm, that's useful' one. – JonW Feb 20 '19 at 17:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.