What should you measure?
Start with reach. Reach is the share of the intended audience that started an episode, expressed against the group you published to rather than against the whole company. A safety briefing published to eight hundred operations colleagues should be measured against those eight hundred, not against a headcount of twelve thousand. Without that denominator, every number looks either flattering or disappointing for the wrong reason.
Then listen-through, which is the share of an episode the average listener actually heard. This is the metric that separates audio from most other internal channels: you can see attention, not just arrival. Alongside it, completion tells you how many people reached the end, and the two together explain each other. High listen-through with low completion usually means the episode was too long rather than uninteresting.
Recurrence is the fourth number: how many listeners from the last episode came back for this one. A show with rising reach and falling recurrence is being discovered and abandoned, which is a content problem, not a distribution one. Finally, per-episode drop-off shows where in the running order attention fell away. A cliff three minutes in is almost always an intro that takes too long to get to the point.
What should you not measure?
Do not measure individuals. Named-listener tracking, per-employee listening records and manager-visible reports on who listened all answer questions nobody in internal communications actually needs. Whether a specific colleague finished the leadership update does not change a single editorial decision.
Aggregated measurement is enough because every useful decision is a group decision: publish shorter episodes, move the show to the employee app, target operations separately, drop the segment where attention falls. It is also safer. Individual-level analytics turn a communications tool into a monitoring tool, which is where works councils, data protection officers and, reasonably, employees start to object. Our approach to access and reporting is described on the security page.
There is a second thing not to measure: downloads as a proxy for listening. A download counts a file transfer, not a person, and it says nothing about whether anybody pressed play.
Also avoid measuring success by the number of episodes published. Output is easy to count and easy to grow, and a series that keeps its schedule while losing listeners looks healthy in a report and is not.
What is a good listen-through rate?
Internal shows average a 78 percent listen-through rate (Hypecast platform data, measured across customer programmes after their first six months). That number means the average listener heard roughly three quarters of an episode, which is far above what public podcasts and internal video typically achieve.
The reason is relevance rather than production quality. A show made for a known audience about a subject that affects their work does not have to compete for attention with the whole internet. Use 78 percent as an orientation, not a target: a fifteen-minute leadership episode and a forty-minute training deep-dive will not land in the same place, and both can be doing their job.
What matters more than the absolute figure is the trend across a series and the shape of the curve inside an episode. A show that moves from 60 to 75 percent over eight episodes is being edited well. A show flat at 78 percent with falling reach has a distribution problem instead.
How do you measure across distribution paths?
Most programmes run on several paths at once: an embedded player on the intranet, a post in a chat tool, a card in the employee app, and a private authenticated feed for the commute. Each path reports differently, so the only way to get one honest number is to aggregate them for the same episode and the same audience group.
Keep the audience groups identical across paths. If the intranet reports by department and the employee app reports by location, the totals cannot be compared and the resulting slide will be argued about instead of acted on. Where a path only offers page views, treat it as reach and take listen-through from the paths that report it, and say so in the footnote. A comparison of the paths themselves is in our internal podcast overview and in the platform description.
Set the reporting rhythm to match your production rhythm. A monthly view is enough for a fortnightly show, and a per-episode view is only worth building when you are actively changing format and want to see the effect of one variable at a time.
How do you report to leadership?
One slide is enough, and it should read top to bottom in twenty seconds. Put the headline first as a sentence rather than a chart title: reached 71 percent of the operations audience, 79 percent listen-through, 64 percent of last month's listeners returned. Three numbers, each with the group they refer to.
Underneath, one line of trend for the series so far, so the reader sees direction rather than a snapshot. Then one observation and one decision: attention falls after minute twelve, so the next three episodes are capped at ten minutes. Leadership rarely wants the dashboard; they want to know whether the channel is earning its place and what you are changing because of the data.
Close with the qualitative line, one sentence quoting a listener, which is what people remember from the slide. Keep the full report as an appendix for the people who ask for it, and keep it aggregated. If you want the wider framing for the channel, internal communications covers how the show sits next to your other formats.
How do you add qualitative feedback?
Numbers tell you where attention moved; people tell you why. Add a single question at the end of an episode description, rotated per episode: what should we cover next, what was missing, what would you tell a colleague about this one. One question gets answered, five get ignored.
Twice a year, talk to six listeners for fifteen minutes each, chosen across locations and roles rather than from the group that already replies to everything. Ask when and where they listen, because the answer usually changes your episode length more than any analytics view. Feed both back into the report as a short quote and a note on what you changed. If you are still defining the channel itself, what is an internal podcast is the place to start.
