The internal-comms platform with a press desk's discipline
15 April 2026 · By Astrid Bergvall · 7 min read
"Eighty-seven percent of employees viewed the announcement," the dashboard says. The leadership team nods. The internal-comms director quietly knows that "viewed" means the email was opened or the page was loaded — not that anyone read past the first line. The gap between viewed and landed is where most internal communication dies. Here are the five metrics we've found actually move with the thing you care about.
How long did the recipient have the announcement open, weighted by their typical reading speed? A 600-word announcement with a median time-on-page of 18 seconds did not land. Same announcement with a median of 90 seconds probably did. The metric is noisy at the individual level and tight at the cohort level — exactly the resolution you want for a piece of communication that's meant to reach a population.
One caveat: time-on-page over-counts when the user opens a tab and walks away. Combine it with scroll depth and you mostly cancel that out.
The fraction of the announcement the recipient actually scrolled through. We've found that the depth-distribution shape — not just the average — tells you the most. A bimodal distribution, with one cluster at 10% and another at 90%, says half the audience bounced and half engaged. A unimodal distribution at 60% says everyone read most of it. Both are useful; the bimodal pattern is a signal that your audience is heterogeneous and you should consider segmenting.
Most all-hands announcements link to something — a deck, a doc, a video. The fraction of recipients who actually clicked the link, and the fraction of those who then dwelled on the linked artefact, tell you whether the announcement created the action it was supposed to create. A high view rate with a 4% click-through rate means the message was received but not acted upon. The two metrics are usually decoupled, and the decoupling is itself diagnostic.
For announcements where you require the recipient to acknowledge — policy changes, code-of-conduct updates, security advisories — the median time between send and acknowledgement is a strong proxy for how prominently the message landed. Acknowledgements that come in within an hour mean the announcement reached the front of the queue. Acknowledgements that drift in over four days mean the announcement got buried and only resurfaced via reminder cycles.
This is the one nobody thinks to measure, and it's the most useful. After every major announcement, monitor the volume of follow-up questions to your HR business partners, your help desk, and the team-level manager queues. A high question volume relative to the announcement's complexity means the message wasn't clear enough — the recipients understood it landed but didn't understand what to do with it. A low volume could mean the message was clear, or it could mean nobody cared. Combine with the engagement metrics and you can tell which.
This metric is harder to capture automatically — it usually requires manual sampling — but it's the metric that most reliably predicts whether the next quarter's similar announcement will land better or worse.
If you want one composite metric to put on the comms dashboard, we'd suggest: weighted engaged-read rate — the fraction of intended recipients who scrolled past 60% of the announcement and spent at least 30 seconds on it. That single number is a better default than view count and a fraction of the complexity to track. Combined with click-through where applicable and acknowledgement-completion-time for required messages, it's enough to know whether you're getting better or worse over the months.
If you'd like to see how Internal Announcement renders these metrics inside the platform, we're happy to walk through it.
Further reading: the BBC's editorial measurement frameworks (publicly available) and Poynter's archive on newsroom analytics translate well to internal comms.