Yesterday I had a chance to talk with a bunch of other web data-analysts about how we think about web traffic (thanks for hosting us, Google!). We talked about a bunch of different topics, including how we think about 'mobile' traffic.
I hadn't thought about it before, but there was a wide difference of opinion, especially about whether tablets should be considered mobile devices. So like a good analyst, I decided to dig in on the numbers. Here's what I learned.
Here's a daily breakdown of our audience by device type for the last year:
- Our tablet traffic has been remarkably flat: 12% of sessions on the weekend, 8% on weekdays. That's a similar usage pattern to mobile phones, which are a larger part of our traffic on weekends.
- The big disruption you can see in the chart is during the holidays: people aren't at their desks, so mobile & tablet traffic (as a percentage, not absolutely) spike.
- While tablet usage is flat, phones have grown steadily, from about 40% to about 45% of traffic. We now get 50% of our weekend sessions on phones.
- Post Christmas, we saw a meaningful jump in mobile traffic, but not tablet sales. This is consistent with declining tablet sales data.
Tablet usage peaks at the same time phone usage peaks, but other parts of tablet data looks like desktop data. About 60% of our pageviews from tablets are in landscape orientation, compared to about 3% of mobile (and almost all desktop hits).
Even more notable, when readers aren't chained to their desks during the work-week, desktop users and tablet users spend the same amount of time reading.
The numbers are broadly similar for other session engagement metrics (pageviews-per-session, etc).
While in general, tablet activity is somewhere between desktops and phones, browsing and traffic patterns are unique. Social traffic is less important for tablets than either of the other device types, and direct traffic is much more important. Below is each device type's direct traffic over the last 5 weeks.
I suspect that this is due to an 'app-gap': Twitter and Facebook have been really successful at driving app usage on phones, and website usage on desktops. However, I'm guessing that app installs on tablets haven't taken off, and the web experience on tablets isn't quite as strong as on desktop. The end result is that our tablet traffic underweights social, and overweights users from direct and search channels.
- Our readers use tablets like desktops, but they use them at times when they use phones.
- Tablets aren't used for social discovery as frequently: instead, readers check out their favorite pages or answer email.
- We're probably better off having our tablet experience look more like our desktop experience than our phone experience, and thinking about those readers like desktop readers, not mobile readers.