Ask someone who has lived through a company-wide transformation how it went, and you'll usually get a pause before the answer. Most of these programs aren't disasters. They just end quietly. The steering committee meets less often, the program office gets absorbed into another team, and a year and a half later someone notices that the old way of working never really left.
You've probably seen the claim that around 70 percent of transformations fail. The figure is shakier than its popularity suggests, and "fail" means different things to different people. Still, anyone who has been close to these efforts knows the basic point holds up. Big change is hard to finish.
Below are the patterns that show up again and again, along with what tends to work against them.
The goal was a slogan
"Become customer-centric." "Go digital." Phrases like these look good on a banner and make terrible plans. When the end state is vague, every team reads it in whatever way matches what they already wanted to do. Six months in, you have forty initiatives under one umbrella and no way to tell which of them matter.
Try this test. Could a frontline manager explain what will be different for their team a year from now? If the answer leans on the word "journey," you're not there yet. Good goals are concrete enough to argue about: cut order-to-cash time from 30 days to 12, or move most routine service requests to self-serve. Those targets might turn out to be wrong, but people can at least disagree with them in useful ways.
Sponsorship fades after the kickoff
Launch events get executive attention. Month nine doesn't. A familiar pattern is the CEO giving a rousing speech and then handing the work to a program office that has no authority to settle disputes between business units.
Those disputes are where transformations live or die. When two departments disagree about who owns a process, someone senior has to decide, and quickly. If that person keeps postponing the call, the whole program slows to the speed of its slowest argument.
Sponsors don't need to attend every meeting. They need a small, recurring commitment, such as a 30-minute decision forum every two weeks, plus a few visible acts that show they mean it. Using the new system themselves helps. So does publicly retiring an old report that everyone knows has been replaced.
Change gets treated as a messaging problem
Plenty of programs spend heavily on newsletters, town halls and intranet pages. The trouble is that people rarely resist change because they haven't heard about it. They resist because it costs them something: status, expertise, time, or a workaround they quietly depend on.
Think about a senior analyst who has spent ten years mastering a complicated spreadsheet model. A new platform makes that knowledge less valuable overnight. Another town hall won't fix that. What helps is bringing people like her into the redesign early, being honest about what they're giving up, and giving them a meaningful role in the new setup.
Training matters here too. A two-day classroom session three months before go-live is mostly forgotten by launch day. Short, hands-on support in the first weeks of real use does far more.
Everything changes at once
Ask a team to adopt a new ERP system, a new reporting structure and new performance metrics in the same quarter, and you'll get exhaustion rather than progress. Each change might be sensible on its own. Stacked together, they overwhelm the people who still have to hit this month's numbers.
Sequencing is unglamorous but powerful. Work out which change has to land first for the others to make sense, and protect teams from the rest until it does.
The dashboard is green and nothing has improved
Program reports often track activity: milestones met, sessions delivered, modules switched on. You can hit every one of those and leave the business exactly where it started.
Pair the activity tracking with a handful of outcome measures that the business already cares about, such as cycle time, cost per transaction or customer retention. If the new system has been live for six months and cycle time hasn't moved, you've installed software. That's not the same thing as transforming anything.
The old way stays switched on
Running the old and new processes side by side feels prudent. Sometimes it is, for a short, defined period. More often the parallel run becomes permanent, because the old way is familiar and nobody is told to stop. Set decommissioning dates, announce them, and keep them unless there's a serious reason not to.
Giving your transformation a fighting chance
None of this calls for a new methodology or an expensive framework. It mostly comes down to discipline over a long stretch of time. Define the end state in terms people can measure. Keep senior leaders involved well past the launch. Treat the people affected as participants rather than an audience. Pace the change so teams can absorb it, and judge progress by results rather than activity.
The programs that succeed are seldom the ones with the boldest vision. They're the ones still getting attention in year two.
