Most communication training is bought as a topic. A manager asks for communication skills for the front office, someone sources a two-day workshop, and the workshop covers listening, assertiveness, email etiquette, presentation and body language. Attendance is good. The feedback form is good. Three weeks later, nothing in the building sounds different.
The delivery is rarely the problem. The unit of design is. A topic cannot be practised, observed or scored. A behaviour can, and that difference decides whether anything survives the month.
Why the category is the problem
Communication is a container word. It holds at least six skills with almost nothing in common, and a facilitator given two days has to spread them thin, roughly ninety minutes each including theory and a group exercise. Ninety minutes creates recognition. It does not create fluency, and it does not survive a Monday morning inbox.
The second problem is quieter. The examples in a generic programme belong to nobody. Participants analyse a sample email written by the facilitator, then return to the email they actually have to send to a client who is threatening to escalate. The distance between the two is where transfer dies.
The fix is not a longer programme. It is a narrower one, designed against what a specific group actually writes and says, which is the difference between buying training solutions and buying a topic.
Twelve elements that hold
This is the checklist we use when scoping a communication programme, and it is the same set of conditions we write into every programme design. It is a set of conditions rather than a curriculum. A programme missing four or five of them will still collect good ratings, which is one reason ratings are a poor test of anything.
| Element | What it looks like in practice | Owner |
|---|---|---|
| 1. Diagnosis on real artefacts | Ten to fifteen actual emails, call notes or meeting recordings read before the design is written | L&D with the business |
| 2. One named behaviour per module | “Gives the decision and a date in the first three lines” instead of “writes clearly” | Facilitator |
| 3. The participant’s own material | Each person works on a message they have to send that week | Participant |
| 4. A credible demonstration | Someone the group respects performs the behaviour live, not an actor in a video | Facilitator and line manager |
| 5. Practice in the room | Every input followed by an attempt, not a discussion about the input | Facilitator |
| 6. Feedback against anchors | Three or four observable markers, written down and shared before the session starts | Facilitator |
| 7. A record of attempt one | The first draft is kept, or filmed, so the learner can see their own change | Participant |
| 8. Manager briefing before | The reporting manager receives the behaviour and the checklist in writing, two weeks ahead | L&D |
| 9. A cue in the workflow | The email template, agenda or call script is edited so the situation prompts the behaviour | Business owner |
| 10. Spaced re-practice | Two sessions of forty-five minutes at week three and week six | Facilitator or internal champion |
| 11. Measurement on artefacts | Twenty post-programme emails scored blind against the same markers | L&D |
| 12. A named owner for follow-up | One person whose job description includes chasing the week-three session | Programme sponsor |
Four of these are usually missing, and they are the four that decide the outcome:
- The manager briefing, because nobody wants to add work to a line manager’s week before a programme has proved itself.
- The workflow cue, because it belongs to a system owner in another department.
- The spaced sessions, because they are not in the fee.
- The blind scoring, because it takes two people half a day and produces an answer nobody is sure they want.
None of the four is expensive. Each one requires a decision before the programme is booked, which is why each one gets skipped.
What practice with feedback looks like
Forty-five minutes with twelve people. Everyone writes for eight minutes on a real message. They read it aloud to one partner, who scores it against four markers in front of them. They rewrite for ten minutes. A second partner scores the rewrite. The facilitator collects the two scores and names the pattern for the room.
The markers are agreed before anyone writes, and they are visible: the decision appears in the first line, one reason is given rather than five, a date is named, and the message closes with one action. A participant who scores two out of four in the first attempt and four out of four in the second has learned something they can repeat. That is a different experience from being told they communicated well.
Compare this with the standard role play: two volunteers, an audience, a facilitator comment about empathy, applause. Everyone in the room learns something about the facilitator’s taste. Nobody learns anything about their own writing.
Listening is a useful test case because it looks unteachable until it is broken into behaviours: summarising before responding, asking one clarifying question rather than three, not stacking a second question on top of an answer. These are describable, so they can be practised and observed, as this note on listening well sets out.
Measurement that is not a smile sheet
Take twenty artefacts from before the programme and twenty from six weeks after, strip the names, and ask two people to score them against the same markers. Report the movement per marker, not a single average, because an average hides the fact that dates improved and reasons did not.
Self-report is fine as a supplement. It is not a substitute. People are poor at judging whether their own email now leads with a decision, and they are generous with themselves when a form arrives from the L&D team that organised the course.
If a programme cannot be measured on artefacts, choose a different first programme.
What to put in the brief
Write five lines into the brief and make the supplier answer each one in writing. Name the behaviour in observable terms, with example markers. Name the artefacts that will be sampled before design begins. Name who briefs the managers, and on what date. Put the week-three and week-six sessions on the calendar with an owner. State how the post-programme sample will be scored and who sees the result.
A supplier who cannot answer the artefact question is telling you what kind of programme they sell. That answer is worth more than the proposal document, and it costs one email to get.
A topic can be covered in two days. A behaviour has to be practised, prompted and scored, and that takes six weeks of somebody’s attention.