Guide
Why change programs fail, and what to do this week
Most large-company change programs miss their goals, and the plan is rarely the reason. The plan is usually sound. What stalls is people: they do not see why, they do not believe it will last, or they slip back into old habits by the third week.
The hard part is that this shows up late and vaguely. Surveys come back months later, dashboards say a team is red, and nobody can say what to do on Monday. This guide names the five human reasons a program stalls, and gives one concrete move for each that you can run this week.
Understanding
People hear the decision but never the reason. A town hall says what is changing, not why the old way stops working.
Try this week: Write the one-paragraph reason yourself and ask three managers to say it back in their own words. Fix the paragraph until they can.
Belief
People have seen programs announced before and quietly outlive them. The doubt is reasonable, not stubborn.
Try this week: Find one small thing already promised and delivered, and point at it. Belief starts with one kept promise, not a campaign.
Emotion
A change that looks rational on paper still takes something away from someone: status, routine, a skill they were proud of.
Try this week: Name the loss out loud in the next team meeting. People move once the cost of the change is said plainly, not hidden.
Habit
The new way asks for effort every day, while the old way runs on autopilot. Willpower loses that race by Thursday.
Try this week: Shrink the change to the first 2 minutes and attach it to something the team already does. In Switch, Chip Heath and Dan Heath call this shrinking the change.
Support
The manager closest to the team was handed the change, not bought into it. The team reads that in minutes.
Try this week: Sit with that one manager before the next announcement. One converted manager beats ten broadcast emails.
From a hunch to a reading
Guessing which of the five is weak in which team is where most programs lose time. A short anonymous pulse per team turns the hunch into a reading: which driver is low, where, by how much. cmi scores each team on understanding, belief, emotion, habit and support, then writes a ready-to-run play for the weakest one, each built on a named idea from the books on how people change.
You can see a full sample play for a return-to-office program on the cmi landing page. It takes about a minute to read and is labelled as a sample.
cmi is built and operated end to end by AI agents on NanoCorp, which is how a guide like this stays plain and current.