What is wrong with internal communications measurement in 2026?
Internal communications measurement is broken because the headline metric measures the wrong thing. Engagement scores capture how employees feel about answering a survey, not whether they understood the message, remembered it, or acted on it. In 2026, three separate studies made that gap impossible to ignore.
Why do engagement scores contradict each other?
Two credible 2026 studies produced numbers that cannot both describe the same workforce. Firstup research found that up to 89% of managers, 82% of corporate office workers and 75% of hourly employees describe themselves as engaged (Firstup State of Employee Engagement, 2026). Gallup, using a validated instrument, put global engagement at 20% for 2025, down from a 23% peak in 2022 and the first back-to-back annual decline it has ever recorded (Gallup State of the Global Workplace, 2026).
When self-reported engagement runs four times higher than measured engagement, the instrument is the problem, not the workforce.
The 2026 Integral Index, conducted with The Harris Poll, gives that gap a name. Its theme is the engagement fallacy, and its central finding is that one in three US employees looks engaged by every standard measure but stays primarily out of fear of losing benefits. The report also notes that AI anxiety is now reshaping workplace behaviour in ways engagement measurement was never designed to detect.
The Firstup data closes the loop: 40% to 46% of those self-declared engaged employees plan to leave within twelve months.
What are internal comms teams actually measuring?
Mostly traffic. The Gallagher Employee Communications Report, drawing on more than 1,300 comms and HR professionals across 40 countries, found that 70% of teams primarily measure output metrics such as page views and click-throughs. Only 16% measure outcomes such as employee sentiment, and only 12% track how their work ties to wider business activity (Gallagher Employee Communications Report, 2026).
That has a career cost. In the same study, 73% of internal comms teams said they want to operate as strategic consultancies. Only 18% do, while 49% still run as centralised broadcast functions (Gallagher, 2026). It is difficult to earn a strategic seat while reporting on page views.
The consequence shows up in the work itself. Between 61% and 67% of employees report missing an important policy update (Firstup, 2026), and roughly 30 percentage points separate what leadership believes it is communicating from what employees actually receive (Gallup). Every one of those misses happened inside a system that was reporting healthy numbers. If that pattern sounds familiar, our analysis of why internal comms are not landing in 2026 covers the reach side of the same problem.
Which metrics do communicators actually want?
They know the answer already. The Trendmonitor Interne Kommunikation 2026, based on 431 practitioners across the DACH region, asked which data internal communicators wish they had. The top two answers were behaviour change at 68.7% and actual comprehension of content at 67.7% (Trendmonitor Interne Kommunikation, 2026).
The same study found that 56.3% of DACH internal comms functions do not measure impact at all, and only 32.8% have any mechanism to check whether a communication met its goal.
So the wanted metrics and the collected metrics barely overlap. The usual explanation is time and budget, and the 2026 internal comms resourcing gap is real. There is also a more uncomfortable explanation.
Why comprehension is a format problem, not a survey problem
Most internal channels cannot produce comprehension data, no matter how well you design the dashboard. An email open tells you a window opened. A click tells you a link looked interesting. An intranet page view tells you nothing about whether paragraph nine was read, understood, or believed.
Video behaves differently, because attention is recorded continuously rather than once. A completion curve on a three-minute policy explainer shows the exact second attention broke. A replay spike shows which sentence people needed twice, which is as close to a comprehension signal as internal comms gets without stopping people to test them. Per-clip drop-off across a town hall recap shows which leadership message landed and which one did not.
None of that is a survey. It is behavioural, passive, and continuous, which is precisely what 68.7% of practitioners said they wanted. It also travels: the same signals work for frontline employees who never open a desktop intranet.
The catch is volume. One video a quarter produces an anecdote. A regular cadence produces a baseline you can read trends against, and that is where production capacity becomes a measurement issue rather than a creative one.
Key takeaways
Engagement is a mood reading, not a comms KPI. One in three employees who look engaged are staying out of fear (2026 Integral Index with The Harris Poll).
Output metrics dominate by default. 70% of teams measure page views and click-throughs, and 12% connect their work to business activity (Gallagher, 2026).
Practitioners already know what is missing. Comprehension and behaviour change top the wish list at 67.7% and 68.7% (Trendmonitor Interne Kommunikation, 2026).
Comprehension data is format-dependent. Completion curves, replay spikes and drop-off points exist in video and do not exist in an email open rate.
Cadence beats craft for measurement. A trend line needs regular output, which makes production capacity a measurement decision.
Where cofenster fits
cofenster is the AI video platform for enterprise communications. Our AI Video Agents let Comms, HR and Marketing teams produce on-brand video at scale, which is what turns comprehension data from a one-off into a trend line.
Theo (text-to-video) turns a policy update into a three-minute explainer without filming, so the completion curve tells you whether the message survived to the end.
Milo (highlight clips from existing recordings) cuts a town hall into clips whose individual drop-off points show which leadership message actually landed.
Ella (guided employee recording) makes executive messages repeatable enough to measure across a quarter rather than a single event.
Brand Templates keeps every output brand-compliant, so higher content velocity never costs brand consistency.
Teams at Continental, Commerzbank and Hugo Boss use this to publish faster and reach distributed and frontline employees without adding headcount. cofenster is ISO 27001:2022 certified, GDPR compliant and EU AI Act compliant.
Frequently Asked Questions
Is it worth running engagement surveys at all?
Yes, as one input among several. Surveys are good at surfacing themes and comparing sentiment across sites or functions over time. They are weak as a communications KPI because they measure feelings about the organisation rather than the effect of any specific message. Treat the survey as context and pair it with behavioural data from individual communications, so a falling comprehension trend can be spotted long before the next annual survey confirms it.
How often should internal communications impact be measured?
Continuously for individual communications, quarterly for trends. The Trendmonitor Interne Kommunikation 2026 found 56.3% of DACH teams do not measure impact at all, and a further group measures only sporadically, which makes trend reading impossible. A workable minimum is per-message data on every priority communication plus a quarterly review that compares comprehension and completion against the previous period.
What is a realistic first step if we currently measure nothing?
Pick one recurring communication and instrument it. A monthly leadership update or a policy explainer works well because the format repeats, which gives you a comparable baseline within a quarter. Record completion, drop-off and replay data, then add one comprehension question to the channel where the communication lives. Two data points repeated over three months will tell you more than a new dashboard covering everything at once.
Does measuring comprehension require a works council agreement in Germany?
It can. Measurement that is capable of monitoring individual behaviour or performance falls under co-determination, and German case law in 2026 has widened works council involvement in AI-supported systems. Aggregate, non-individual measurement is generally easier to agree, so involve the Betriebsrat early and define what is collected, at what level of aggregation, and for how long it is retained. This is a legal question for your own counsel rather than general guidance.
Photo by Vitaly Gariev on Unsplash.
