We've noticed that device type and distribution channel percentages in our Real-Time Dashboard (and/or the Heads Up Display) do not equate with total concurrents viewing from device/distribution category, and/or that these percentages do not align with our historical pageviews or uniques by device-type/distribution-channel data. How are these percentages calculated?
Device and distribution percentages in Real-Time and HUD are calculated by comparing total engaged sessions from each device type and distribution channel to total engaged sessions site-wide. An engaged session in Chartbeat represents an active user session where the visitor has directly engaged (clicking, scrolling, typing, mousing) with your page in the last 15 seconds. Chartbeat measures both engaged and non-engaged sessions as concurrents, so comparing the ratio of total concurrents viewing from a particular device or platform to your site-wide total concurrent count will not match the share-of-traffic percentage displayed in our Real-Time Dashboard.
To view your engaged session totals by device type and site-wide in Real-Time, select any of the Device filters and hover over the most recently populated portion of the Concurrents by Traffic Source graph:
In the example pictured above, we would expect the unfiltered Real-Time view to display 46% traffic-share for the Mobile device type (mobile engaged sessions / total engaged sessions).
Note that engaged session counts will always represent a portion of your total concurrents, both by device and site-wide. Concurrents are the total number of people on your site at any given moment, as measured by the number of open browsing sessions to your site. This number includes both engaged and non-engaged browsing sessions open to your site. A non-engaged session represents an open browser tab to your site where the visitor has not shown any signs of engagement within the last 15 seconds. If an open browser session to your site shows no signs of engagement for two continuous hours, we will stop tracking the session as a concurrent at that point.
Additionally, the ratio of active engaged sessions by device on your site at any given time should not be expected to match your sites' average pageviews or uniques by device type ratios in your historical Chartbeat data. Historical pageview and unique session counts are very different metrics and should not be directly compared to your engaged session totals in Real-Time.
Desktop readers tend to return to backgrounded browser tabs with much more frequency than mobile readers do, so we track sessions differently based on device type to accurately capture that engagement. By comparing only engaged concurrents for each platform, you get a more accurate ratio for the share of traffic by device type and distribution-channel on your site.