Behavioural Change Toolkit · Sustain

Change Retrospective

90 minutes, at day 90. Two purposes: learn, and bank the credibility for next time.

⚠️ This is not a lessons-learned document for a shared drive. Those are written, filed and never read. This is a session that produces three things the organisation will actually do differently.

Who

The core team, 2–3 line managers, 2–3 champions, and at least two people who were sceptical. Not the steering group — a retro with the governance layer in the room becomes a status report.

Run sheet

0:00 — The data, no interpretation (10 min)

Behaviour metric over time. Outcome. Counter-measures. Put it up and be quiet.

0:10 — What did we predict? (10 min)

Read the original hypotheses and the original Change on a Page aloud, unedited.

“This is what we said would happen. What actually happened?”

⭐ Reading your own original assumptions out loud is uncomfortable and it’s the most valuable ten minutes in the session. It’s also how you build the habit of predicting rather than describing.

0:20 — Timeline (20 min)

Draw the timeline on the wall. Everyone adds:

  • 🟢 moments it went well
  • 🔴 moments it went badly
  • ❓ moments they were confused
  • 😮 moments that surprised them

The ❓ marks are the most useful. Confusion is cheap to fix next time and expensive to leave.

0:40 — Five questions (30 min)

  1. What did we get right that we should do again? (Start here. A retro that opens with failure gets defensive and stays there.)
  2. What did we believe at the start that turned out to be wrong?
  3. What did we find out late that we could have found out early?
  4. What did we do that made no difference? ⭐ (Almost never asked. Most change plans contain activity that changes nothing and gets repeated forever because nobody checked.)
  5. What did the organisation do to itself that blocked this?

1:10 — The three things (15 min)

Not a list of twenty. Three things done differently next time. Each with a named owner and a place they’ll actually land — a template, a checklist, a governance gate.

# What we’ll do differently Owner Where it’s captured
1      
2      
3      

1:25 — Thanks (5 min)

Specifically, by name, especially the people who told you things you didn’t want to hear. Say what changed because of them. Do this properly — it’s how you get honest feedback next time.

Capture

Retro record

Change: _______________ Date: _______________

Result, honestly (one paragraph, no spin): _______________

Behaviour metric: baseline ___ → day 90 ___ Embedded? ___/6 on the normalisation tests

What we were wrong about: _______________ What made the biggest difference: _______________ What made no difference: _______________ What the organisation did to itself: _______________

The three things: 1. ___ 2. ___ 3. ___

The version you give the sponsor

Different audience, different document. One page:

  1. What happened, in numbers, without spin
  2. The one thing that made the most difference — usually something they did, or didn’t
  3. The one thing that cost us most — usually an organisational contradiction nobody resolved
  4. What I’d ask for on the next one

Point 4 is the whole reason to write it

A retrospective that ends in learning is worth something. A retrospective that ends in a changed condition for the next change — an earlier sponsor conversation, a saturation check before approval, a rule that no rollout starts without a legacy switch-off date — is worth ten times as much. You have more influence in the fortnight after a delivered change than at any other time. Spend it.

Building the pattern

After three or four of these you’ll see the same causes recurring. That list is more valuable than any individual retro — it’s the diagnosis of how this organisation does change, and it’s the basis of the advice only you can give.

Keep a running note:

Recurring cause Seen in Structural fix
Legacy system never switched off   Mandatory switch-off date at approval gate
Sponsor engaged late   Sponsor briefing required before design starts
No baseline captured   Baseline in the standard gate checklist
     

→ Drift Check — 30 60 90 · Myths and Traps in Change Management · Change on a Page

← Back to the Behavioural Change Toolkit

Last updated: 11 August 2026