Change Leadership: How a CEO Makes Organizational Change Hold

Most organizational change does not fail on the day it is announced. It fails in the months after, once the launch is over, the steering committee has moved on and the old way of working quietly comes back. Leading change that holds is mostly about those months: what the CEO keeps deciding, what the managers in the middle are given, and what the company measures once the slides are gone.

The research says the same thing. In BCG’s study of 825 senior executives, only 30 percent of digital transformations met their targets and left a change that lasted; another 44 percent created some value, missed their targets and faded. When economists went back to Indian textile plants nine years after a successful management experiment, about half of the new practices had been dropped. The two reasons cited most: managers who moved on, and senior people who stopped giving it time.

I coach CEOs and leadership teams through organizational change so that it holds after the launch. Before that I led it from the inside: a five-year cloud and DevOps transformation across more than 1,000 engineers at Cox Automotive, a 450-person product and engineering organization at New Relic, and a turnaround as COO of a venture-backed company. This guide is what I would tell a CEO about to announce a change: the part only you can do, what decides whether a new design holds, how to bring the middle of the company with you, and when outside help is worth paying for. If your change has already stalled, start with what I look for when a transformation stalls.

Key takeaways

• Change rarely fails at the announcement. It fails after the launch, when the work that keeps it alive has no owner.
• The CEO’s job is not the speech. It is naming the result and why it is a must, moving money, people and time to match, and staying visible after the launch.
• A new structure with an old scorecard drifts back. Decide who owns which decisions, then change what people are measured and paid on.
• Middle managers carry the change or stop it. Give them the why in their team’s terms, time to do it, and a say before the plan is final.
• Resistance tells you what people stand to lose. Ask what the current way protects for them.
• A change management consultant can design, facilitate, coach or add capacity. Sponsorship stays with you.

Why most organizational change fails after the launch

A change has two finish lines. The first is the launch: the announcement, the new org chart, the system going live. It has a date, a budget and a status report. The second is the point where the new way is simply how work gets done, with no one quietly holding it up. That one has no date, and it is where most change is lost.

Psychologists have a name for the pattern at a small scale. A postcompletion error is the step people skip once the main goal is reached, like leaving your card in the cash machine after the money comes out. Byrne and Bovair showed in 1997 that it happens most when people are carrying a heavy load. Companies do the same thing. Go-live is the cash. The exceptions no one owns, the analyst still checking each output by hand, the manager fixing the new process on Friday afternoons: that is the card, and the busiest people in the building are the ones who leave it behind.

Change also fails by being absorbed. The new language arrives and the incentives stay the same. The transformation becomes another reporting layer. A new committee is added on top of the old decisions instead of replacing them. From the outside it looks like progress. Inside, the company has kept what it actually rewards.

And people read what you move, not what you say. An announcement costs nothing to make, so employees discount it. What they watch is costly and hard to fake: where the budget went, who got the best people, which meeting the CEO stopped attending, what was switched off. If none of that changes, the announcement told them the change is optional.

The changes I have led ran through the same curve, the Kübler-Ross change curve: disbelief, then resistance, then a dip in productivity, then acceptance. Each person’s first question is how this affects them, and no one does their best work while they are worried about it. Leaders go through the same curve, usually weeks before the announcement, which is why the people who designed a change are often surprised by how hard it lands. The leader’s job is to make the dip shorter and shallower.

What the CEO has to do, and cannot hand off

Change management can be delegated. Change leadership cannot. A program office can run the plan, a consultant can design it and HR can build the training. Five things stay with the CEO.

1. Name the result, and whether it is a must

“Move to the cloud” is not a result. It is a tactic. I worked with a large healthcare company that had tried to move to the cloud three times, and each time success was defined as getting there. A result is a number and a date that would make the business better: ideas into production in a fraction of the time, a new market at a set margin, a cost line the CFO can audit.

Then ask why it matters, and whether it is a must or a should. When I ask a leadership team to describe where they will be in a year and why it is a must to get there, I usually hear as many answers as there are people in the room. That is the first thing to fix. If the honest answer is “should”, either the where or the why is wrong, or the team is not committed, and no amount of activity makes up for it. When the three line up, the how gets clearer and faster.

2. Move something that costs you

Commitment shows up in what you take away. Fund the change by stopping something, and say what. Put your strongest people on it, not the ones who happen to be free. Change your own calendar so the change gets your time each week, not just at the kickoff. A change added on top of a full plate is the first thing dropped when the quarter gets tight.

3. Stay visible after the launch

Prosci has asked change practitioners for more than two decades what made their changes work, and active, visible sponsorship has been the top contributor in each of its benchmarking reports since 1998. Its research also lists how sponsors go missing: present only at the start, too many changes at once, or handing the sponsor role to the project team or an outside consultant. Projects with an extremely effective sponsor met or exceeded their objectives more than twice as often as those with a very ineffective one. Visible means the people affected hear about the change from you, more than once, and see you working the new way yourself.

4. Settle who decides

Write down the decisions that will determine whether the change works, then ask your leaders, separately, who owns each one. You will find decisions two people think are theirs, decisions no one owns, decisions sitting three levels above the work, and decisions that drifted when the business changed. Two leaders who each believe a decision is theirs produce a busy program that does not move. Write the names down.

5. Keep deciding

Changes stall above the people running them more often than below. When the leadership team stops making the calls on outcome, funding and what gets switched off, and starts reading status reports instead, the change slows and no one ever says no. Your job after the launch is to keep taking the trade-offs as they arrive: which work stops, which team gives up headcount, which commitment slips.

A new structure with an old scorecard drifts back

The most reliable warning sign for a reorganization or a new operating model is that the metrics and the pay stayed the same. Move pricing authority to the regions while the sales plan still pays on volume, and pricing discipline erodes within a couple of quarters, with no one doing anything wrong. People do what the scorecard pays for.

So once the decisions have owners, change three things around them. The metrics, so each owner is measured on the decisions they now hold. The incentives, so the new behavior is the one that gets rewarded. And the review rhythm, so the new owners have a meeting where their decisions are discussed and stick; that is how authority becomes real to the people watching. One rule I use for targets: the team that sets a target owns it, even when it depends on another team. Otherwise a miss turns into a blame game.

At New Relic I took over a 450-person product and engineering organization after the previous leader was let go. The largest customers said there had been no meaningful innovation in nearly three years, and the sales team had stopped believing engineering could deliver. We changed how decisions were made before anything else: one owner for the business metrics across product and engineering, a written quarterly allocation of engineering capacity, and one escalation meeting for any change to the roadmap. About six months in, two of our largest customers told us we had delivered more of what they asked for in six months than in the previous three years, and the new products went on to add more than $50 million in incremental ARR.

Middle managers carry the change, or it stops with them

In the BCG study, three out of four executives believed their leadership was engaged in the change. Only one in three companies had committed engagement from middle managers. That gap is where many changes stall. Gallup estimates that managers account for at least 70 percent of the variance in engagement across teams. Strategy is set at the top; it is lived, or not, one manager at a time.

What middle managers need is specific:

• The why in their own team’s terms. An announcement says what is changing. A manager has to answer the question each team asks next: what does this mean for us? That takes context the announcement does not carry, including what the team heard the last three times leadership announced a new direction.
• A say before the plan is final. Hear from your top managers before you lock the plan, then show them where their concerns landed. A plan built at an offsite and pushed down meets resistance from the people who know the daily reality best.
• Time. Running the business and changing it are two jobs. If the change is added on top, it loses to the day job. Take something off their plate.
• Permission to carry it as their own. A manager who says “they told me to tell you” has become a messenger, and the team hears that the change is optional. If a manager disagrees, have that argument in private before they face their team.

Treat resistance as information

Resistance is rarely about the change itself. It is about what people stand to lose: status, the competence they built in the old way, control over their work, their place in the group. Telling people to embrace change answers none of those.

Better questions:

• What does the current way do for the people in it? Old processes protect something. Find it, and you know what the new way has to replace.
• Who is most disrupted, and who is running the change? If the answer is the same team, you have built in a conflict. In the healthcare company that tried the cloud three times, the operations team was both.
• What would have to be true to get to a yes? I ask it of teams who say no, and their answers become the plan.

Some resistance sits at the top. Each layer you remove has a person attached to it, and the processes being retired were often designed by people still in the building who delivered when it mattered. Leaders tend to negotiate rather than decide: the plan stays, the names stay, and a new committee is added on top. The cost is deferred, not avoided.

Build the rhythm that keeps it alive

The launch gets a project plan. The weeks after it need one too. The first danger point usually comes around 90 days in, when the novelty has worn off and the day job pulls people back to the old way. Give the work left after go-live what the launch had: a named owner, protected time, a review date and the authority to change things.

My frame for change has three parts: the right formula, the right tactics and repeatable actions. The formula is the big-picture design. The tactics are the blueprint for how it gets done here, with these people. The repeatable actions are the rhythm that keeps the change alive after the transformation team has moved on, and they are the part most plans leave out.

The rhythm includes four things:

• A standing forum where people can ask anything. A short update and a long open question period, with no screened questions, does more than a polished town hall. Say what you know, say what you do not know yet, and say what is not changing.
• A few leading signals, reviewed weekly. Decisions being made at the level you assigned, escalations falling, the old workarounds disappearing, the new process used for real work. Count adoption in work moved, not in logins or attendance.
• A plan for when the people carrying it leave. In the textile plants, the practices that disappeared were the ones tied to managers who moved on and to senior people who stopped giving them time. A change that depends on one person’s attention ends when that attention moves.
• Review dates set in advance, when the leadership team asks what is working, what is not, and what needs to change.

Watch for change fatigue while you do this. People are rarely reacting to one change alone. They are reacting to it plus the last few that were launched with the same energy, the ones that were quietly dropped and the one that is still half finished. A louder launch into a tired organization reads like the last one, so the fix is fewer changes at a time, not more energy. Stage them, finish or formally stop one before the next one starts, and say out loud which earlier initiatives are being dropped to make room.

AI adoption is the change most companies are leading right now

Most of what CEOs ask me about this year is AI, and it fails the same way. In IBM’s 2026 CEO study of more than 2,000 chief executives, 83 percent said AI success depends more on people’s adoption than on the technology. The same study found only 25 percent of workers use AI regularly in their jobs. The licenses were bought; the work did not change.

An AI transformation is a business transformation with a new tool in it, and the same rules apply: name the business result first, decide who owns each workload after go-live, and use the tools yourself where people can see it. One rule is specific to AI: say where the time saved will go before you automate someone’s work. In BCG’s 2026 AI at Work survey, 66 percent of regular frontline users said they get limited or no guidance on what to do with the time AI saves them. People share what they know more readily when they trust where the gain will land. For where to start, see my AI readiness assessment checklist.

When to bring in a change management consultant

Most changes do not need an army of consultants. They need a CEO who has decided, a leadership team that agrees on the must, and managers with time to carry it. Outside help earns its fee in a few situations:

• No one inside has led a change of this size, and learning on the job is expensive.
• The leadership team cannot agree on the result or the why, and the CEO is too close to it to run that conversation and take part in it at the same time.
• The people best placed to run the change are the people it disrupts most.
• You need capacity: the team running the business cannot also design and run the change.
• The board or the owners want an independent read on why it is not landing.

It is not worth it when what is missing is a decision only the leadership team can make, or groundwork your own people will have to own afterwards. Paying someone else to do that work leaves you dependent on them when they leave.

Prosci lists handing sponsorship to an outside consultant as one of the signs of a missing sponsor. A good advisor will tell you that in the first meeting.

The labels overlap, so decide what you are buying:

• A change management consultant does the work and owns a defined outcome for a period: the plan, the communication, the training, the readiness checks.
• A facilitator runs the process for your leadership team and stays out of the content.
• An executive coach works one to one on how you and your leaders lead the change.
• An operator-advisor has run changes from the inside and works with you on the decisions as well as the people.

How to choose one:

• Write the brief first. Two or three results specific enough to check, and what each is worth to the business. It sharpens each conversation and gets the budget approved.
• Listen to the questions they ask before they propose anything. A good one asks what has been tried before, which decisions are stuck and who stands to lose what. Be wary of a framework presented before they understand your business.
• Ask who will do the work. The person in the pitch should be the person in your meetings.
• Ask whether they have led a change from the inside, and what went wrong.
• Ask what will still be running 90 days after they leave, and how you will know. Attendance and satisfaction scores measure an event, not a change.
• Ask what capability they leave behind. The best ones work themselves out of the job.

That is the work I do: coaching CEOs and leadership teams through change (executive coaching), running the leadership conversations (facilitation), and advising on the operating model and the decisions underneath it (strategy consulting). If a change is about to launch, or has launched and is not landing, the first conversation is forty-five minutes. Book a discovery call.

Frequently asked questions

What is change leadership?

Change leadership is what the people at the top do to make a change happen and hold: deciding the result and why it is a must, moving resources to match, staying visible after the launch and making the trade-offs as they arrive.

What skills does change leadership take?

Mostly skills of judgment rather than of presentation: choosing the result that counts and holding to it, spending something visible to back it, staying present after the launch, settling who decides before it causes friction, and making trade-offs as they arrive. Two more sit around them: hearing resistance as information about what people stand to lose, and running a review rhythm that outlasts the launch.

What is the difference between change management and change leadership?

Change management is the process: the plan, the communication, the training and the readiness work, usually run by a team or a consultant. Change leadership is the decisions and the visible commitment of the leaders. You can delegate the first. You cannot delegate the second, and changes fail more often for lack of it.

Why do most organizational changes fail?

Most fail after the launch rather than at it. The work left after go-live has no owner, the scorecard still pays for the old behavior, middle managers were not brought in, and senior leaders stop giving it time. In BCG’s study, 30 percent of digital transformations met their targets and left a change that lasted.

What is the CEO’s role in leading change?

Five things: name the result and why it is a must, move budget, people and time to match it, stay visible after the launch, settle who owns which decisions, and keep making the trade-offs. A program office or a consultant can run the rest.

How do you get middle managers on board with change?

Involve them before the plan is final, give them the reasons in their own team’s terms, take something off their plate so they have time to carry it, and settle disagreements with them in private before they face their teams.

How do you handle resistance to change?

Treat it as information about what people stand to lose: status, competence, control or their place in the group. Ask what the current way protects for them and what would have to be true for them to say yes. Do not let the team most disrupted by the change run it alone.

When should you hire a change management consultant?

When no one inside has led a change of this size, the leadership team cannot agree on the result, the people who would run it are the ones it disrupts most, or the business needs extra capacity for a set period. Keep sponsorship and the trade-offs with the leadership team.

How long does it take for organizational change to stick?

Longer than the launch plan says. The first danger point is usually around 90 days, when the novelty wears off. People tend to believe a change after a few consistent months, and deeper culture change takes a year or more. At New Relic our largest customers felt the change in about six months; the cloud transformation at Cox took five years. Set the review dates and name the owner of the work left after go-live before you launch.