Measurement framework

Internal communication metrics beyond channel performance

Opens, clicks, attendance and listens are useful. They answer different questions, for different measurable audiences. A complete measurement system starts one step earlier: who was the message intended for, and did each group have a realistic way to receive it?

Use this framework to separate workforce coverage, channel delivery, attention, understanding and action without turning one metric into a claim it cannot support.

The measurement question comes before the dashboard

A metric is useful only when its denominator matches the decision. An email open rate describes recipients included in an email send. Meeting attendance describes invited people whose presence was recorded. An audio completion rate describes listeners who started an episode.

None of these numbers automatically represents the whole workforce or proves that a message was understood. The first step is to define the intended audience, then map who could realistically access each channel.

Six layers

Move from intended audience to action

1. Intended audience

Name the people who should receive the message. Segment by role, location, language, shift or team when their access conditions differ.

2. Realistic access

Record whether each group can use the channel in its actual working context, including device, login, inbox, app, shift overlap and manager dependency.

3. Delivery

Confirm that the channel accepted or delivered the message. Delivery is a technical event, not evidence of attention.

4. Attention or consumption

Use opens, attentive reading, attendance, starts, completion or time spent according to the channel. Keep the denominator visible.

5. Understanding

Ask a short question, pulse survey or manager check. A click, open or completed episode does not prove comprehension.

6. Action or outcome

Measure the requested behaviour when it is observable. Treat wider engagement or business outcomes as separate measures with many possible causes.

Metric definitions

What each channel metric can and cannot show

Delivery rate

Shows whether a system accepted the message for the addresses or devices in scope. It does not show attention or whether people outside that addressable audience had another route.

Open rate and attentive reading

Show activity among email recipients using the measurement rules of the email platform. Privacy controls, image loading and client behaviour can affect the event.

Clicks

Show interaction with a tracked link. A message with no link can still be useful, and a click does not prove understanding.

Attendance

Shows presence at a live or recorded event when attendance is captured. It does not establish attention throughout the session.

Audio starts and completion

Show playback activity for invited listeners. They provide format-specific evidence, not proof that the listener understood or acted.

Email evidence

Internal email performance is not the problem statement

PoliteMail's 2026 benchmark aggregates more than two billion internal emails sent to roughly 11 million employees. It reports a 66.2% average open rate, and 83.7% of opens met its attention rate threshold under the provider's methodology.

Those figures contradict the claim that internal email is broadly unread. The legitimate question is narrower: which intended workforce groups were represented in the sendable and measurable audience?

Read the source context and limitations in chapter 2 of the report.

Coverage layer

Map access before calculating a rate

Start with a count, not a percentage. For one recurring message, list every intended group and document whether at least one asynchronous channel is realistically available. Mark the answer as verified, partial or unknown.

Define

Which people should receive this message, and why are they in scope?

Verify

Which channel can each group access in practice, and what evidence supports that answer?

Count

How many intended people have verified access, partial access or an unknown access path?

Investigate

Treat unknowns as discovery work. Do not label them disengaged or unreachable without evidence.

Monthly scorecard

Keep the scorecard small and decision-oriented

For each critical recurring message, record:

  • Intended audience and segment sizes.
  • Verified, partial and unknown access paths.
  • Delivery and consumption metrics by channel.
  • One understanding or feedback signal.
  • The requested action, when measurable.
  • Owner, follow-up and next review date.

Do not collapse the layers into one engagement score. Separate numbers make it easier to see whether the problem is access, delivery, attention, understanding or action.

Common mistakes

Claims the dashboard does not support

  • Calling opens workforce reach.
  • Calling completion understanding.
  • Calling attendance attention.
  • Comparing percentages with different denominators.
  • Using consumer podcast benchmarks for employees.
  • Assuming an unknown group is disengaged.
  • Attributing a business outcome to one message or channel.

Frequently asked questions about internal communication metrics

What is the most important internal communication metric?

There is no single universal metric. Start with the intended audience and the decision you need to make, then choose separate measures for access, delivery, attention, understanding and action.

Is email open rate a workforce reach metric?

It is a channel-performance metric for the measurable email audience. It does not automatically describe employees who were not in that audience or whose practical access conditions differ.

Does audio completion prove understanding?

No. Completion shows playback activity under the platform's measurement rules. Use a question, survey, manager check or observable action when understanding matters.

How should desk and non-desk groups be compared?

Define the same intended message, then compare practical access and evidence separately for each group. Avoid ranking groups by a metric that only one group can generate reliably.

When is a communication pilot justified?

A pilot is justified when one recurring message, one defined group, an owner, a safe context, a written equivalent and a decision date are all clear. It should also be acceptable for the result to be no rollout.

Audit one message before adding a channel

Use seven questions to map the intended audience, realistic access and the evidence available by segment.

Related resources