Eighteen months after the policy went live, the CHRO asks the obvious question: has anything changed? The intranet has a page explaining the right to disconnect. Managers went through a briefing. Everyone signed something. And nobody in the room can say whether the organisation actually communicates differently at nine in the evening than it did two years ago.
France has had a legal right to disconnect since 2017, negotiated at company level for organisations above a certain size. Belgium, Italy, Spain and Portugal have their own versions, and the European Parliament has pushed for a directive. What almost none of these frameworks include is a way to tell whether the policy has had any effect. That evidence exists, in your Microsoft 365 tenant, and looking at it raises a question that deserves a straight answer before anyone opens a dashboard.
What after-hours data actually shows
When internal communication teams look at publication and reading patterns by hour for the first time, three findings come up repeatedly, and none of them is the one they expected.
The first is that total after-hours volume is rarely the problem. In most organisations the share of internal communications published outside working hours is smaller than people assume. What matters is not the volume but who is sending. A director posting to Viva Engage at 21:40 does more to set the norm than two hundred routine messages sent at the same hour by nobody in particular. Everyone below calibrates to what leadership visibly does, and a policy document does not compete with an observed habit.
The second is the Friday and Sunday shape. Campaign launches scheduled for Monday morning generate a wave of preparation traffic across the weekend, concentrated in the communications function itself. Teams that have written a right to disconnect policy are often, on the data, among its most consistent breachers. That is not hypocrisy, it is a workflow problem, and it is fixable with scheduling rather than with training.
The third is that “after hours” is not one thing in a multinational. A message published at 17:00 in Paris arrives at 08:00 in Los Angeles and at 20:30 in Mumbai. Without a country or time zone attribute joined to the activity data, an hourly chart of a global tenant is close to meaningless. This is the same attribute problem that shows up everywhere else in internal communications measurement, and our note on audience segmentation covers how to structure it.
The line you have to draw before you look
Here is the uncomfortable part. The dataset that answers “are we sending less at night” is the same dataset that answers “who in this department works at night”, and the second question is not one an internal communications team has any business asking.
The distinction is between channel health and individual behaviour. Channel health is a property of the communication system: when things get published, when they get read, how quickly a reply is expected. Individual behaviour is a property of a person. The first is legitimate internal communications analysis. The second is workforce monitoring, and in France and across the EU it carries obligations, consultation requirements and a real risk of destroying trust in the communications function permanently.
| Legitimate question | Question that crosses the line |
|---|---|
| What share of internal news is published outside working hours, by channel? | Which people publish outside working hours most often? |
| Has the evening reading tail shortened since the policy? | Who is still reading at 23:00? |
| How fast is a reply expected on the leadership community, by hour of posting? | Which team members reply fastest at night? |
| Does the pattern differ between countries or functions above a minimum size? | Does it differ between the six people on this team? |
Practically, this means three rules built into the tooling rather than into a policy nobody reads. No named individuals in any communications report. A minimum segment size below which a visual refuses to render, so a small team cannot be reverse-engineered into a person. And no manager-facing drilldown into their own team’s timing. Tryane measures aggregate patterns and does not report individual behaviour, which is a product decision rather than a configuration option, and it is the reason works councils tend to approve deployments quickly. The wider governance picture is in our piece on security and compliance for internal communications analytics.
Where Viva Insights fits, and where it does not
Microsoft has a product for the wellbeing question. Viva Insights gives individuals personal insights about their own working patterns and gives leaders aggregated, de-identified organisational views with privacy protections built in[1]. If the question is “is our working culture healthy”, that is the right instrument and it belongs with HR.
Internal communications analytics answers a narrower question: whether the communications channel is itself contributing to the problem. Keep the two separate, with different owners and different audiences. Conflating them is how a communications dashboard ends up in a disciplinary conversation, which is a bad outcome for everyone including the dashboard.
Measures that stay on the right side of the line
Four measures cover most of what an internal communications team can usefully say about the right to disconnect, and all four are aggregate rates rather than counts of people.
- Out-of-hours publication share. The proportion of internal news, Viva Engage posts and internal newsletters published outside 08:00 to 18:00 local recipient time, by channel and by sending function. Track it monthly. It is the one number the IC team directly controls.
- Evening reading tail. The share of total reads occurring after 19:00 local time, as a trend. A shortening tail after a policy launch is weak evidence but it is evidence. A lengthening one is worth investigating.
- Expected response latency. Median time to first reply on community posts, split by the hour the post was published. If a 22:00 post gets a median first reply in eleven minutes, you have measured an expectation, not an accident.
- Weekend campaign share. The proportion of campaigns whose preparation traffic falls on Saturday or Sunday. This one usually points straight back at the communications team’s own planning calendar, and our note on content calendar performance covers how to fix it upstream.
The fix is mostly scheduling
The most useful thing to come out of this analysis is usually operational rather than cultural. Internal communications teams control publication timing, and both SharePoint news and Viva Engage support scheduled publishing. Writing at 22:00 because that is when you had a clear hour is fine. Publishing at 22:00 is a choice, and it is the choice that sets the norm.
Setting a team standard, that nothing scheduled goes out before 08:00 or after 18:00 recipient local time unless it is genuinely urgent, moves the out-of-hours publication share more in one quarter than any awareness campaign. It also improves reach, because the best time to post on Viva Engage is almost never late evening. The right to disconnect and effective communication happen to want the same thing here.
The harder half is leadership. If the CEO posts at 22:00, no team standard survives it. This is a conversation, not a dashboard, but the dashboard makes the conversation possible: showing that leadership accounts for a disproportionate share of after-hours publication, in aggregate, is a factual observation that is difficult to argue with and does not name anyone.
What the data will not tell you
Be honest about the limits, because the temptation to over-read this data is strong. Reading a post at 21:00 is not evidence of harm. Some people work compressed hours by choice, some are parents catching up after bedtime, some are in a different time zone from the sender and are simply at work. Communication data sees behaviour, not intent, and the difference between choosing to read at night and feeling obliged to is exactly the variable that matters.
That variable comes from asking people. Gallup’s long-running research on burnout points to unmanageable workload and unclear expectations as the dominant drivers, not raw hours[2], and Gallagher’s State of the Sector consistently finds measurement to be the capability internal communicators most want to strengthen[3]. Pair the pattern data with one or two survey items and you have a defensible read. Use either alone and you have half an argument. Our piece on engagement surveys versus communications data makes the case for running both.
Tryane reads SharePoint, Viva Engage, Teams and newsletter platforms together, joins that activity to organisational structure through Entra ID or an HR file so that local time and function are available, and keeps history long enough to show a before and after around a policy change. It reports on channels and aggregate segments, never on individuals. SOC 2 Type 2 certified, GDPR compliant by design, EU hosting by default with US data residency on request, single sign-on through Entra ID, deploys in a couple of hours. If your context is distributed teams specifically, the companion piece on measuring hybrid work communications goes further.
Frequently asked questions
Does the right to disconnect apply outside France?
Several European countries have their own statutory or negotiated versions, including Belgium, Italy, Spain and Portugal, and the European Parliament has called for an EU-level instrument. Elsewhere it is usually a company policy rather than a legal obligation. The measurement question is the same in both cases: a policy without evidence of effect is a statement of intent.
Can we report on which departments send the most after-hours messages?
Yes, above a minimum segment size, and framed as channel behaviour rather than as individual performance. A department of two hundred is a legitimate unit of analysis. A team of four is a person with extra steps. Set the floor in the tool, agree it with the works council, and publish it so nobody has to wonder.
Is Viva Insights the same thing as internal communications analytics?
No. Viva Insights looks at working patterns for wellbeing and productivity purposes, with personal insights for individuals and de-identified aggregates for leaders. Internal communications analytics looks at whether messages reach and land. They answer different questions for different owners, and keeping them separate protects both.
What is a realistic target for out-of-hours publication?
There is no benchmark worth quoting, because it depends on shift patterns, time zones and what counts as urgent in your sector. Set your own baseline in the first month, then aim for direction rather than a number. A team that halves its after-hours publication share in two quarters has done something real, regardless of where it started.
Will our works council accept this kind of measurement?
In our experience yes, provided you go in with the boundaries already defined: aggregate only, minimum segment sizes enforced in the tool, no individual reporting, a stated purpose limited to communications channel management, and a defined retention period. The objections arise when a team arrives with a general-purpose analytics tool and no answer to what it will not be used for.
Sources
• Microsoft Learn, Viva Insights introduction
• Gallup, preventing and dealing with employee burnout
• Gallagher, State of the Sector
Further reading
• Internal communication and hybrid work
• Security and compliance for internal communications analytics
• Engagement surveys versus communications data
• Measuring internal communications on Microsoft Teams
• Audience segmentation for internal communications
Tryane runs a 15-minute working session with Heads of Internal Communication, walking through what your own tenant shows about publication and reading patterns, and where the aggregation boundaries need to sit. Book a slot with Jérémy.