Things I Wish I Could Tell Everyone About Analytics

Data is sexy, but data is dumb
I've noticed two recurring problems with website measurement. The first is easier to deal with than the second, and it's that data is seductive (and yes, I realize that data is plural but I just don't want to say "data are sexy" because I mean data as a concept not as a particular set). Because we have so many tools with which to collect data about our websites, our purchases, our tools, our bodies, ourselves, we are left with the problem of making sense of it all. Enter data-visualization, that shiny, sexy stuff that we see on news broadcasts and in magazines, that somehow validate the existence of the data even if the point itself is obscure. Nothing a little handwaving and photoshop can't fix. But seriously, the sexiness of dataviz is a stumbling blog to actually taking measurement seriously. Producing data isn't the point, nor is crunching it and spitting out reports. The point is asking the right questions and finding answers that help you to actually do productive things.

Ok, so the second problem is harder. The second problem is that many people believe that measurement is far more complex than it need be, and that they simply don't have the time to do it well. I believe this is false, but that it can only be corrected through education, and because there is so much information on the web (both good and bad), it only introduces another filtration problem for which few have the time to sort out. So, to start, let me provide a mantra of sorts...

We must simplify our quantitative measurement in order to enrich our qualitative measurement.

What I mean by this is instead of dwelling on arbitrary metrics—How many visits did my site get? How many conversions? What's my bounce rate?—we need to identify a series of questions that are relevant to our pursuits and assemble the tools and extract the metrics that will help us answer them for the sole purpose of doing better.

It's not that keeping the pulse on various basic metrics is bad. On the contrary, I recommend it, provided it's not the only measurement you're doing. But it should be the easiest part of your measurement regimen, exceeded in emphasis (and time spent) by deeper landing page analysis and user testing. Remember, keeping the pulse will tell you if something is working or if something is not working, but it's not as good at delivering nuanced insights on how your site is being used. That's just as valuable for you to know.

Here are a few key points I want you to remember:

1. Measuring individual web metrics won't tell you enough. Combine them to ask specific questions, then take action on the answers.
2. Sometimes you need to pull data out of Google Analytics in order to figure out what questions you should be asking.
3. Bounce rate only applies to traffic that landed on a page. Google Analytics doesn't make this clear. More on that here and here.
4. One of the best things you can do with your time in Google Analytics is to get to know your referrers.
5. Where do your most conversion-rich referrers sit on the familiarity spectrum? Robot, friend, Obama or Grandma?