Mobile Works Nights and Weekends

People spend an increasing portion of their time online, so it’s not surprising that their daily rhythms (i.e. work schedule, commute, etc) are reflected in their online activity.

Our presence on 14 million websites worldwide gives us a unique perspective on this behavior, and the opportunity to use data science to help our publishers understand their users. Here’s an example comparing mobile vs. desktop users by hour over a typical week.

These are users in the eastern U.S. seen on AddThis sites during the week May 6 to May 12. Users are categorized as desktop/laptop or mobile (which includes phones and tablets).

How (And When) People Use Desktops

As you can see, there are wide swings in desktop/laptop user numbers from working to evening hours. On the weekends (the right end of the chart above), you’ll see a pronounced decrease which is what we’d expect when people aren’t at work. And looking more closely at the blue desktop curve, we notice a dip on weekdays around 5pm, commute time.

How People Use Mobile

The red mobile curve is more surprising. There’s no weekend dip as there is in desktop, and the high point of weekday use is relatively late, around 7-9pm, coinciding with the second evening peak of desktop/laptop use. Also note how steady mobile use is throughout the day Saturday.

So what’s going on? Is the evening mobile peak “second screen viewing” in action? Is mobile enabling web surfing on weekends, traditionally a slow period for online activity?

In future posts, we’ll dig into this data to uncover some possible explanations, and brainstorm how you can use this knowledge to optimize your visitors’ experience.

  • That is pretty interesting. I think a bar graph with times on y axis might be a better presentation of this information.

  • Jim Merino

    Thanks for the great feedback Michael. I went back and forth for a while
    before deciding which type of graph to use. My decision was tipped toward a line graph because I wanted readers to interpolate between the discrete hourly values.

  • Jim, what do the two graphs look like if you normalize the y axis to a common measure? It looks to me like there is a drop-off on Saturday for mobile which might be comparable to the desk-top drop-off.

  • Daniela Keil

    Love the visualization. Would like to see for other markets, esp. Japan and Germany.

  • Jim Merino

    The mobile and desktop data are on the same scale, but the absolute numbers aren’t shown. Saturday mobile doesn’t have the same evening peak as weekdays, but the daytime user numbers are comparable with the rest of the week. Here’s a cliffhanger: the lack of an evening peak for mobile on Saturday and to some extent Friday is interesting – that will be the subject of my next posting.

  • Great, look forward to it.

  • This is strange behaviour of users Interesting too. How can we measure use of internet during office times and what are the relax timings. But Drop of every three hour is very strange.

  • Nazia Rehman

    I’ve a wordpress site. I want to know from where should I take a start for addthis. My site is HowBees- find how to do anything. Thanks

  • susan

    Hi Nazia – great to hear you’re ready to get started. Check out how to work specifically with WordPress at When you’re ready, it’s as simple as downloading the code straight from our site. Email if you have trouble. Good luck!

  • Nazia Rehman

    Thank you, Susan. I’m thinking to give it a try very soon as soon as I get out of my current busy schedule. Hope it’ll be easy just like wordpress.
    Again Thanks a lot for your help!