Why change initiatives fail — and what it actually costs

A few years ago I went to Vegas on a two-night work trip and learned an expensive lesson about why change fails. I'd never really gambled before. I set myself one rule: I was only bringing $200 into the casino. My goal? Double my money.

I spent most of my time at the blackjack tables. First night I ended up with about $350. Second night I lost some, gained some back, and walked away with $235. A total gain of $35 over two nights. I was proud of myself. I'd played smart. I'd been disciplined.

The next year my company went back to the same place. This time I felt prepared. I'd learned from the first trip, I had coworkers coaching me, and I was hopeful.

I lost it all in about 45 minutes on the first night.

I was shocked. I was mad. So I took another $200 out of an ATM and ended that night with about $140. The second night? Lost it all. In total I was down $435 over the two years.

Here's what hit me on the flight home: it didn't matter that I'd learned from the first trip. It didn't matter that I had good coaching or came back better prepared. I was playing a game designed for me to lose. The structure of the game itself was rigged against me.

And then I started asking myself a question that changed how I think about leading change: how often do we initiate changes that are structured for people to fail?

The problem everybody can see

Most change initiatives are reactive, which is to say that plans get built around structures, systems and numbers on a spreadsheet while ignoring the relational chaos, broken trust and emotional fallout that derail even well-designed transformations. That is the external problem, and it is the one everybody can see.

Sitting underneath it there is an internal problem that gets discussed a good deal less. Most leaders already know that change is messy for people, and many of them carry the memory of an overwhelmingly stressful change process they lived through personally. They want to minimize the human cost while still hitting ambitious targets, and nobody has ever shown them how.

So the failure here is rarely a failure of intent, which is what makes it so frustrating to watch. It is a failure of method. What follows is what that failure actually costs, the six patterns that produce it, and a way of working out which one you are currently in.

What it costs when you get it wrong

Three things tend to happen, and in my experience they happen in roughly this order.

The objective slips

It rarely slips dramatically, and almost never in a way that anyone can point at directly. Adoption runs slower than the plan assumed, workarounds start to proliferate, and the old process survives in a spreadsheet that somebody maintains privately because it still works and the new one is unfamiliar. Reporting gets softer. Eventually the initiative is quietly rescoped and declared a success at a considerably smaller size than it started.

What makes this so difficult to catch is that no individual week ever looks like failure. The slippage is distributed across dozens of small accommodations, not one of which is worth escalating on its own, and by the time the aggregate becomes visible the narrative has usually already settled: the plan was too ambitious. In my experience the plan was generally fine. What was never resourced was the adoption.

Your best people disengage first

This is the part that surprises people, because it runs against the intuition that your strongest performers are also your most committed. The people with the most options leave first, and they leave quietly. The Work Institute puts the replacement cost at 1.5 to 2 times annual salary, but the real cost is that the people who understood how the place actually worked are now gone in the middle of you changing how it works.

They rarely leave over the change itself, which is what makes the exit interviews so misleading. They leave because the change revealed something about how decisions get made in that organization, and they did not like the answer. Exit interviews collect the polite version of this. The honest version is usually that they stopped believing their judgment counted for anything, and once a strong performer reaches that conclusion the notice period is a formality.

The next change is harder

Trust spent is not trust replenished. A transformation that hits its financial targets while breaking relationships leaves you with an organization that meets the following change with cynicism and resistance, and that relational damage takes years to repair when it repairs at all.

This is the compounding cost, and it is the one that never appears in anybody's business case. Every subsequent initiative now opens at a deficit, because people have a prior, and the prior is that leadership says one thing at kickoff and something rather different by month four. You are no longer asking them to adopt a new process. You are asking them to be wrong about you, which is a much larger request than most leaders realize they are making.

You spend ten times more fixing the new problems than it would have cost to train people how to lead through the change in the first place.

The structure matters six times more than the coaching

Research on team effectiveness gives us a number for this. Roughly 60% of team performance is determined by structure — how teams are set up for success before they ever meet. Only about 10% comes from the day-to-day coaching and management that most of us spend our energy on.

Let me say that again. The structure matters roughly six times more than the coaching does.

Yet most leaders spend 80 to 90% of their time on that 10%. We coach our way through a transformation while the structure of how we work quietly undermines it every single day.

Ruth Wageman's studies of Xerox service technician teams showed how lopsided this gets. Well-structured teams with good coaching thrived. Poorly structured teams with good coaching? Almost no improvement at all. But well-structured teams with genuinely bad coaching still did well, because the structure protected them.

That is the finding worth sitting with, and it reframes everything below. The six patterns that follow are not communication mistakes. They are structural ones.

The six failure patterns

Here's what rigged games look like in a transformation. Across dozens of organizations I keep seeing the same mistakes recur, and none of them are exotic. Every one is the reasonable-looking shortcut of a leader working under real time pressure, which is precisely why they are so hard to avoid.

A rigged game is when we ask managers to carry the emotional weight of a restructure and never give them a single hour of preparation to do it.

A rigged game is when we measure the transformation on milestones hit and never once on whether trust survived it.

A rigged game is when the people closest to the work find out about the change from a town hall, and we call their reaction resistance.

1. Announcing the what without the why

People can absorb an enormous amount of disruption when they understand its purpose, and they will resist relatively trivial disruption when they don't.

The shortcut is understandable enough, because the why is often commercially sensitive, only half-formed, or genuinely uncertain, so leaders announce the decision and hold the reasoning back until it firms up. The difficulty is that people do not wait around for an explanation. They manufacture one, and in the absence of a stated purpose the assumed purpose is almost always cost-cutting, and almost always personal.

Instead: say the why before you can say it perfectly. Something along the lines of "here is what we know, here is what we don't, and here is when we will know more" is a complete and credible message. Certainty was never a prerequisite for honesty.

2. Treating communication as a broadcast

A cascade of slides is not communication. Communication is when you know what landed.

Broadcast feels like progress largely because it is measurable — the deck went out, the town hall happened, attendance was ninety percent — and none of those numbers tell you whether a single person changed their mind. What they tell you is that the message was transmitted, which is genuinely the cheapest part of the job.

Instead: build yourself a return path and then actually use it. Ask your managers what their teams are saying rather than whether they delivered the deck. The useful question is never "does anyone have questions?" asked of a room of forty people. It is "what is the version of this that's going around?" asked of six people individually.

3. Mistaking compliance for commitment

Silence in the town hall is not agreement, and it is usually the sound of people deciding something privately.

Compliance and commitment produce identical behavior for about a quarter, which is what makes this so easy to miss. The two diverge the first time the new way turns out to be inconvenient — when the deadline is tight and the old process is simply faster. Committed people absorb that friction. Compliant people revert, having correctly judged that nobody is going to notice.

Instead: treat disagreement as a resource rather than an obstacle. The person arguing with you in month one is doing unpaid quality assurance on your plan, and if you make it safe to be the one who says there is a hole in it, you get to find the holes while they are still cheap to fix.

4. Resourcing the systems and not the leaders

The middle managers carrying your change out to their teams are typically given the least preparation and the most exposure of anyone involved.

This is the most consequential and least examined line item in most transformation budgets. The technology gets a program manager, a training plan and a support desk, while the manager who has to explain all of it to eleven anxious people gets a slide deck and a talking-points document, and then absorbs every hard question personally — usually without knowing the answers, and frequently without permission to admit that.

Instead: resource the relational load in the same way you resource the technical one. That means preparing managers before the announcement rather than after it, giving them the honest version including the parts that remain unresolved, and building the capability to hold a difficult conversation before they are standing in the middle of one. This is the work executive coaching does best, and it is why our leadership workshops are built for intact teams rather than open enrollment.

5. Ignoring the fundamental human needs at play

Every change threatens some combination of security, belonging, significance, autonomy and justice, and naming which one is under threat turns out to be most of the work.

These are not interchangeable, and offering the wrong reassurance tends to make matters worse rather than better. If you answer a security fear — "nobody is losing their job" — when the actual threat is to someone's significance, because their expertise no longer seems to matter here, it reads as evasion. It reads that way because it is evasion, in the sense that you have answered a question nobody asked. Most failed change communication is a well-intentioned answer to the wrong need.

Instead: diagnose before you reassure. Resistance is information about which need has been exposed, and it is usually legible enough if you are listening for it rather than listening through it. Our guide to the five fundamental human needs during change works through each one and what it tends to sound like in the room.

6. Declaring victory at go-live

Go-live is the beginning of the behavioral change rather than the end of it, though almost nothing about how these programs are run reflects that.

The program has a launch date because programs need launch dates, but that date is an operational milestone which has quietly been borrowed as a psychological one — the point at which attention, budget and leadership presence all withdraw together. That is precisely the moment when people are first doing the new work under real conditions, discovering everything the pilot never surfaced, and deciding privately whether the new way is survivable.

Instead: fund the ninety days after go-live as deliberately as you funded the ninety before it. Keep a visible escalation path for problems, and respond to the first few loudly, because everyone in the organization is watching to see what happens to the person who put their hand up.

How to tell which one you're in

Most struggling transformations are showing more than one of these patterns at once, but there is usually a primary. Four questions will get you close:

• Can three people at three different levels state the why in their own words? If they hand you the slide language verbatim, or if you get three genuinely different answers, you are looking at pattern one or two.
• What happened the last time somebody raised a serious objection? If nobody can recall an instance of that happening, it does not mean there have been no objections. It means you are in pattern three.
• What did the managers get, and when did they get it? If the answer is a deck, on the morning of the announcement, that is pattern four.
• What have people stopped saying out loud? Whichever topic has gone quiet is where the threatened need is sitting, and that is pattern five.

Answer those four honestly and the intervention is usually obvious, which is why I say the difficulty here is rarely diagnosis. The difficulty is that the honest answers are uncomfortable, and the people best positioned to give them report to the person doing the asking.

What actually prevents it

The through-line running through every one of those patterns is the same: the plan accounted for the operational change and not the relational one. The correction is not softer management, which is how this argument usually gets caricatured. It is treating the human side as a workstream with an owner, a budget and a measure, exactly like any other part of the program.

That phrasing is doing real work, so let me be specific about it. A workstream has a named owner who is accountable for it, a budget that cannot be quietly raided when the technical build overruns, and a measure that gets reported alongside all the others. Anything less than that becomes the thing everyone agrees is important and nobody is responsible for, which is how it ends up as a values slide nobody reads.

Concretely, that comes down to three things:

• Humanize the vision. Build the transformation strategy so that it achieves the business goal and creates conditions in which people can actually function. These are not competing objectives, whatever the pressure of the moment suggests, but they do both have to be written down, because unwritten objectives lose.
• Develop the relational skill before it is needed. The capacity to hold a hard conversation is built in advance or it is not built at all, since nobody has ever acquired it during the week they needed it.
• Implement in a way that reduces fear. People can feel whether they are being valued or processed, and they calibrate their discretionary effort accordingly, usually without ever saying so.

Leaders shouldn't have to choose between business success and treating people with compassion during change. There are proven ways to humanize a transformation so that both survive it, and that is what our change leadership programs are built to do.

But before you look at any program, sit with one question.

If you're honest with yourself, what percentage of your energy on this transformation has gone into coaching people through the game — and what percentage has gone into changing the game itself?

Because you can have the best change plan in the world. You can execute it flawlessly. But if you've handed your people a rigged game, the house still wins.

Sources

• Wageman, R. Studies on self-managing service technician teams at Xerox.
• Hackman, J. R. & Wageman, R. When and How Team Leaders Matter.
• Work Institute. Retention Report — employee replacement cost.