ePress Analytics has a new home in the platform, and with it a set of features built around a single question our customers kept asking: which titles actually get read, and by whom? The answer used to take a support request and a spreadsheet. Now it takes a dashboard.
Reading sessions, not just downloads
Download counts were never a good proxy for engagement. A patron who opens a magazine, reads two articles and closes it looks identical to one who taps the cover by accident. The new session model separates the two. Every reading session records its duration, the titles opened within it, and the point at which the reader left, which means a library can finally distinguish a popular cover from a popular publication.
Sessions are aggregated hourly and retained at daily granularity for two years. That is long enough to compare this January against the last two, which is the comparison most acquisition decisions actually rest on. Nothing in the session record identifies an individual reader; the model was designed with that constraint first rather than bolted on afterwards.
Collections that explain themselves
The second addition is collection-level reporting. Instead of reading a list of six hundred titles sorted by popularity, an administrator can now group titles into collections that reflect how the institution actually thinks about its catalogue: by faculty, by language, by audience, by supplier agreement. Each collection carries its own trend line, its own cost-per-session figure where licensing data is available, and its own share of total reading time.
This turned out to matter more than we expected during the pilot. Several universities discovered that a small number of trade publications accounted for a disproportionate share of reading in departments that had never requested them. Two of those institutions restructured their renewals as a direct result. One reduced its overall spend while increasing measured engagement, which is the outcome the feature was built to make visible.
Exports that fit existing reporting
Analytics is only useful if it reaches the people who write the annual report. Every view in the dashboard can be exported as CSV, and the same data is available through a scheduled export that drops a file into a location you nominate. Institutions already running a reporting warehouse can skip the dashboard entirely and consume the export directly, with a stable column contract that we version rather than change underneath you.
For teams that report to a consortium, there is a shared view: participating members see their own figures in full and the consortium aggregate alongside, without seeing any other member’s individual numbers. Getting that boundary right took longer than the rest of the work combined, and we would rather have shipped it late than approximately.
What comes next
Three things are already in progress. Saved views, so a report you build once can be reopened by a colleague. Alerting on unusual movement, so a title that suddenly stops being read surfaces without anyone going looking. And a public-facing summary widget for institutions that want to show their community what it is reading. All three are targeted for the next two releases.
Analytics is available now to every customer on a current plan, with no additional configuration required. Historical data is backfilled where we hold it, which for most accounts means the dashboard is useful on the first day rather than the ninetieth. If you would like a walkthrough with your own figures on screen, your account contact can arrange one.