The session obsession
Two users visit my site Data News Daily. Here’s how they compare:
Which is better?
On the basis of these numbers, Jenna seems clearly more valuable: 10x time spent per session; nearly three times as many pages per visit. She appears to be a highly engaged user.
But there’s more to the story…
Jenna is browsing Facebook and taps on a link a friend had shared about data visualization. That opens a long article that she reads in full. She then taps on a couple of “Read Next” links to other posts. In total she spends 25 minutes on the site. She doesn’t return. She couldn’t tell you what the publication was that she just spent 25 minutes reading.
Jason checks the homepage of Data News Daily every couple of days. He doesn’t have it bookmarked, but his browser autocompletes the web address easily enough. When he sees a story that interests him, he clicks on it and reads some or all of it. The length of his visits range between 30 seconds and 10 minutes, depending on whether he finds anything interesting to read.
But which is the better user? Jason, who makes repeated visits, or Jenna who has one (and only one) long reading session?
Most sites should prefer to have Jason as a user. He knows the site and keeps returning to it. It’s not clear that Jenna is ever coming back. And yet two of the most-loved metrics—Pages per visit and Average Session Length — are clearly much worse for Jason.
So, what’s the problem?
Session- and visit-specific metrics encourage short-term thinking and put the long-term value of the user at risk: platforms can’t build a sticky, dependable audience if they try to wring the most out of each visit, rather than designing for the longer term benefit of the user.
Here are some different numbers that better capture Jason’s value:
If these metrics are bad, why do we still measure them?
- Habits die hard (Google Analytics still has session-based metrics on its default dashboard). But there are better options now: back in the day there were no browser tabs (if a site was open, the user was probably paying attention) and it was hard to capture data across multiple sessions.
- Because these are shorter-/nearer-term measurements they feel more actionable. I’d argue that there’s not a lot of point in action-ing bad metrics. Better to wait.
In summary
If you’re designing for long term user experiences, you need to measure for them, too. This is vital to the success of much of the web and mobile today. E-commerce, publishing, social, gaming, education all benefit from building loyalty and habits in users. How do you measure that?
If you want help thinking about what your metrics should be, and how to measure them, get your name on the list for our beta program.
Building a new platform to measure and improve user engagement.
For more information and early beta access, visit us at runhop.com.
