Archive for February, 2010

Understanding what chartbeat is measuring

February 5th, 2010 by Tony

When people initially think about "real-time analytics" they often think of it as "like traditional analytics, only faster." But what's often missed is the transformative way chartbeat measures visitor activity: it gives you the data you need faster, but it also makes that data much richer. With a traditional service like Google Analytics, a user sends a "ping" to the service when they first arrive on the page, and, sometimes, whey they click on links. From this data, it's possible to count the number of page views in a period of time. But measures of how long users remain on the page are just guesses. In contrast, with chartbeat, one user sends repeated pings, as frequently as every 15 seconds, to say "I'm still here, and this is what I'm doing." Are they actively reading your site, or do they have it open in a browser tab for later? Are they writing something, maybe a comment or search? Are they scrolling down and engaging with the content or not getting much further than the headline? The chartbeat method of counting provides a richer sense of how many qualified and retained views you're getting, rather just raw visits to your web site. Our data gives a more accurate accounting of time on page and a better sense of how much of that time was actually spent looking at and interacting with your page. This is information that that can be valuable in optimizing your website, improving your marketing, or selling more advertising. Why your chartbeat numbers don't match your traditional analytics numbers All analytics services measure data in different ways and reconciling them can be difficult, but doing so can be even more difficult when trying to reconcile real-time analytics with traditional analytics. Knowing how many users are on your site at any given time is qualitatively different from knowing how many page loads occurred during a given period of time. The same 5 page views in an hour could be 1 concurrent view if users do not remain engaged and leave your site, or 5 concurrent views if those users remain engaged and stick around. This is illustrated in the diagram above. While traditional measurement only knows that there were 5 visits between 10:00 and 11:00, chartbeat knows which sessions are active at any given moment and what they're doing. When you load the dashboard or view your history, you see slices in time, some with 2 users present at once (10:05), some with 4 users present (10:00). So what's better, page views or concurrent views? We think they both have their place. But it's important to understand that these are two very different perspectives on visitor behavior, and there can be valuable insight in the differences. If chartbeat shows a larger percent of traffic on some page than Google does, that page may be holding users' attention for longer. If chartbeat shows more traffic coming from direct than search, it may mean that users coming via Twitter are more likely to linger than hits from web search. Viewing this behavior from the angles of both traditional and real-time analytics will give you far greater insight into what's really happening on your site and how to improve it than any one tool alone.