Most big change initiatives have an arc: kickoff energy, a few promising quarters, then a quiet slide back to the way things were. Six months after launch, the new operating model is hanging on by a thread, and the team is half-using the old process again because it's easier. Industry research consistently puts the failure rate of major change programmes around 60–70%, depending on how you define failure.
The pattern isn't mysterious. The initiatives that stick share six characteristics — almost all of them decided before anyone runs a single workshop. The initiatives that fizzle skip them. Here's the checklist.
1. A single, named owner who isn't the sponsor
Most change programmes have a sponsor (an executive who funded it) and somehow no owner. Or the sponsor is also the owner, which is a polite way of saying nobody owns it day-to-day.
The owner is the person whose name comes up in every status update, who is judged on whether the change sticks, and who shows up at the retrospective to defend or revise the approach. They should be one layer below the sponsor, with enough authority to make tradeoffs but not enough that the work becomes an exec project.
If you can't name the owner in one breath, you don't have a change initiative — you have an announcement.
2. A measurable outcome, written down before the launch
Vague outcomes invite vague results. “Improve cross-functional collaboration” can't be evaluated; “reduce handoff delays between sales and ops from 5 days to 2” can.
Write the metric down before the project starts and post it somewhere visible. Two practical rules:
- One headline metric, not five. Multiple metrics let you cherry-pick at review time.
- A baseline number with a date.“Currently 5 days, measured Q3 2025” protects you when someone six months later argues that 5 days was always optimistic.
3. A removal, not just an addition
Most change initiatives add something — a new process, a new tool, a new role. The ones that stick also remove something. New without removed becomes additional, and additional things get rationalised away under workload pressure.
Examples:
- New CRM rollout — kill the spreadsheet that sales ops uses to reconcile pipeline.
- New approval workflow — eliminate the email chain it replaces.
- New meeting cadence — drop the meetings whose purpose this one absorbs.
If you can't name what you're removing, the change will feel like additional work to the people doing it. They'll revert.
4. A pilot, not a big-bang rollout
Pilots aren't about reducing risk — they're about learning. The first attempt at any operating-model change will be wrong in ways nobody can predict. A pilot lets you find that out when the cost of being wrong is small.
A good pilot has three traits:
- One team, six weeks. Not the easiest team, not the hardest. The one with normal complexity.
- An explicit kill criterion.“If by week four we're not seeing X behaviour, we stop and rethink.” Most pilots don't have one, which is why they all succeed — by definition.
- A real debrief.Honest, written, shared. What's working, what isn't, what changes for wave 2.
5. Anchored in the calendar that already exists
Change initiatives that require new meetings, new rituals, and new tools all at once fail under the weight of their own infrastructure. The durable ones bolt onto existing rhythms.
If the team has a Monday standup, the new workflow update happens there — not in a new Tuesday meeting. If there's a monthly business review, the headline metric for this initiative appears on slide two — not in a separate quarterly readout.
The principle: every additional thing the team has to remember is a tax on adoption. Spend that tax sparingly.
6. A planned hand-off date
Change initiatives need to end. Not the change — theinitiative. There's a moment when the new way of working stops being a project with a steering committee and becomes the way the team works. That moment should be planned, not accidental.
Set a date — usually six to nine months in — when the project ends, the owner hands off to a line manager, the steering committee dissolves, and the headline metric becomes part of normal reporting. Plan how the work continues without the project structure.
Initiatives without a planned end-date stay in “project mode” forever, which means they keep consuming attention until something more urgent displaces them. Then they die quietly.
The signals at month six
Six months after launch, look for these:
- People describe the new way as “how we do X”rather than “the project we did.” Language matters; it signals normalisation.
- The headline metric has moved in the right directionand the team can attribute the movement to specific changes (not “we think it's working”).
- People outside the original pilot want in. Pull, not push.
- The old way feels awkward to revert to. Even if you suggested it, the team would push back.
Two or fewer of those, and the change is on the slide back. Time for an honest conversation about which of the six structural decisions you skipped.
One sentence that decides most of it
Before launching any change initiative, write one sentence: “In six months, X will be true for Y group of people, measured by Z.” If you can't write that sentence, the initiative is not ready. If you write it and then never refer to it again, the initiative will fail without you noticing for several months. The sentence is the entire spine of the work. Treat it accordingly.