Kotter's 8-Step Change Model — The Sequence Matters
John Kotter found that most change efforts fail because steps are skipped. Here's the sequence, where most leaders break it, and the diagnostic to find which step you're stuck on.
- Kotter's research at HBS: 70% of change efforts fail — almost always due to skipped or rushed steps.
- The 8 steps create urgency, mobilize a coalition, and lock the change into culture.
- Most leaders skip steps 1-3 and start at step 5 (act).
- Sequence is non-negotiable; speed inside a step is fine, skipping is fatal.
- Use as both a planning template and a forensic tool for stalled change.
A 400-person company tried to shift to product-led growth. Year one: nothing changed. The CEO had announced the strategy in a townhall and assumed momentum. He'd skipped the first four Kotter steps — the unsexy work of urgency, coalition, vision, and 10x communication. The next year, done right, took 9 months and stuck. The pattern is depressingly consistent across orgs: leaders start at step 5 because it feels like progress, and the change quietly dies.
Why it matters
Kotter's 1995 HBR article 'Leading Change: Why Transformation Efforts Fail' (later expanded into the book) remains the most-used playbook for organizational change. Every step exists because skipping it breaks something downstream. Step 6 (short-term wins) doesn't work without step 2 (a coalition to engineer them). Step 8 (anchor in culture) doesn't work without step 4 (10x communication) creating the shared language to anchor to.
It also matters as a forensic tool. When change has stalled, you can walk backwards through the 8 steps and find the skipped one — and that's usually the one to repair before you do anything else. Most stalled-change diagnostics conclude with: 'we need to go back to step 1 (urgency) because we never actually built it'.
The 8 steps
- 11. Create urgencyShow why the status quo is more dangerous than the change. Data, not slogans.
- 22. Build a guiding coalitionGet 6-12 influential people with the credibility, expertise, and authority to lead.
- 33. Form a vision and strategyOne paragraph anyone can repeat. The 5-minute test.
- 44. Communicate the vision — 10xUnderestimate the comms budget. Then 10x it.
- 55. Empower actionRemove barriers — structures, managers, systems that block the new behavior.
- 66. Generate short-term winsEngineer visible wins within 90 days. No wins = momentum dies.
- 77. Sustain accelerationDon't declare victory early. Use wins to attack bigger barriers.
- 88. Anchor in cultureBake into hiring, promotion, rituals, and stories.
The 8 steps as a timeline
- →Phase 1: Set up (steps 1-3)urgency, coalition, vision
- →Phase 2: Mobilize (steps 4-5)communicate, empower
- Phase 3: Prove & embed (steps 6-8)wins, sustain, anchor
Where did you skip?
| Symptom | Likely skipped step | Repair |
|---|---|---|
| 'No one is taking this seriously.' | Step 1: Urgency | Show data on cost of status quo; eliminate complacency. |
| 'Initiative is one person's project.' | Step 2: Coalition | Recruit 6-12 cross-functional power-brokers. |
| 'I can't repeat our vision in 1 sentence.' | Step 3: Vision | Rewrite. Test on 5 people. Iterate. |
| 'People say they didn't know about it.' | Step 4: Communication | Re-launch across 7 channels, repeat 7x. |
| 'They want to but the system blocks them.' | Step 5: Empower | Remove process / manager / tool barriers. |
| 'It's been a year, no momentum.' | Step 6: Short-term wins | Engineer a visible win in next 90 days. |
| 'It worked for 6 months, then died.' | Step 8: Anchor | Bake into hiring/promotion/rituals. |
Example
Satya Nadella's Microsoft cultural shift to 'growth mindset' followed Kotter precisely. Urgency: the mobile and cloud miss made the cost of standing still impossible to ignore. Coalition: a renewed senior team aligned on the shift. Vision: 'mobile-first, cloud-first' — repeatable in one sentence by anyone. 10x communication: every all-hands for years. Empowerment: re-orgs that removed silos. Short-term wins: Azure deals visible internally. Sustain: acquisitions (LinkedIn, GitHub) signaling commitment. Anchor: review and hiring rubrics rewritten to make the new behavior the way you got promoted.
Apply on Monday
- Diagnose your current change against the 8 steps. Where did you start?
- Walk back to the lowest skipped step and do it now.
- Identify your guiding coalition by name — if you can't, you don't have one.
- Schedule a 90-day short-term win deliverable now.
- Write your vision in one paragraph. Have 5 people read it back to you in their own words.
Common mistakes
- Starting at step 5 — acting before urgency, coalition, vision.
- Communicating the vision once, in one channel.
- Declaring victory at the first quick win.
- Skipping step 8 — and watching the change unwind in 18 months.
- Treating the coalition as 'the exec team' — it should cross levels and functions.
- Building urgency through fear alone (works short-term, corrodes long-term).
Reflection prompts
- Which step did I skip on my last failed change?
- Where is urgency theoretical rather than felt?
- What 90-day win could I engineer to prove momentum?
- Which system (hiring, promotion, review) am I not yet using to anchor?
Takeaways
- Sequence is the model. Skipping is the failure mode.
- Use it forward (planning) and backward (forensics).
- Coalition is people; vision is one paragraph; comms is 10x.
- Without step 8, every change has an 18-month half-life.
Urgency → Coalition → Vision → Communicate (10x) → Empower → Win (90 days) → Sustain → Anchor. Skip none. Speed up within steps, not by skipping them.
Read next
All playbooksThe simplest and oldest change model — and the one most useful for thinking about behavior change at the human level. The fundamental physics that every later model is built on.
Kotter and Lewin work at the org level. ADKAR is the model for the human level — what each person needs to actually change, and which letter is your current bottleneck.
Most leaders treat resistance as the problem. It's actually the most valuable data you have. Here's how to read it — and the 5 drivers that need 5 different interventions.