Essay
There is a familiar sequence inside large organizations. A transformation programme gets a name, a steering committee and a budget line. Vendors are shortlisted. A platform is chosen after months of evaluation. Training is scheduled. And then, eighteen months later, people are quietly running the new system exactly like the old one, entering data into the same fields in the same order, because that is how they were taught to think about the work long before anyone bought them a licence.
Nobody in that story did anything obviously wrong. The technology usually works. The rollout usually happens on schedule. What fails is the assumption underneath the whole programme: that installing a capability is the same thing as changing a behaviour.
The technology is the easy part
This is uncomfortable for anyone whose job title includes the word “transformation”, but it is worth saying plainly: the hardest part of a digital transformation has almost nothing to do with digital. Selecting a platform is a procurement exercise. Configuring it is an engineering exercise. Getting three hundred people to work differently, on purpose, every day, without a supervisor watching, is an entirely different kind of problem, and it is the one that determines whether the investment was worth making.
Organizations consistently underspend on this part because it is harder to plan, harder to schedule and much harder to put a number on in a business case. You can quantify a licence fee. You cannot easily quantify the cost of a team quietly deciding the new way of working is not worth the discomfort of learning it, so they will look compliant while doing the minimum.
Fear does not show up in the project plan
Ask people privately why they resist a new system and you rarely hear a technical objection. You hear something closer to a fear of becoming visibly incompetent in front of colleagues who used to see them as capable. A new tool strips away years of accumulated shortcuts and workarounds that made someone look fast and confident. For a period, everyone using it looks like a beginner again, and very few adults enjoy looking like a beginner in front of people who report to them.
That fear is entirely rational. It is also almost never named in the change management deck, which tends to talk instead about “building enthusiasm” and “communicating the vision”. Enthusiasm does not solve a fear of looking incompetent. Psychological safety to be visibly bad at something for a few weeks does, and almost nobody builds that deliberately into a rollout.
People do not resist change. They resist the version of themselves the change requires them to become in public.
Incentives are still pointing at the old behaviour
The second failure point is structural, and in a sense more embarrassing, because it sits entirely within the company’s own control. Bonuses, promotions and performance reviews frequently keep rewarding the metrics the old process optimized for, even after the new process has gone live. A sales team told to adopt a consultative, data-informed selling motion but still measured and paid purely on monthly volume will use the new tool to do the old thing faster. Call it what it is: a rational response to the actual incentive structure. No amount of training fixes an incentive.
- Does the new way of working get someone promoted, or does the old shortcut still get there faster?
- Who is visibly rewarded, this quarter, for behaving the new way while it is still slower and more uncomfortable than the old habit?
- Does a manager’s own bonus depend on their team learning something that makes the manager temporarily look less essential?
- What happens to someone who tries the new process, does it imperfectly, and it costs them a result?
That last question is usually the real one. If the honest answer is that they get quietly penalised, everyone in the building has already understood the actual policy, regardless of what the town hall said.
Unlearning is slower than learning
Most transformation timelines budget for learning a new system and never budget for unlearning the old one. These are not symmetrical tasks. Learning a new interface takes an afternoon. Unlearning an instinct that a particular field means a particular thing takes months, because instinct does not live in a manual and does not respond to being told it is wrong once in a workshop.
Leaders who succeeded under the old model are especially slow to unlearn it, because the new model implicitly suggests that some of what got them promoted no longer matters. Asking someone to unlearn the exact judgement that made them senior is one of the harder requests in organizational life, and it deserves to be treated as hard rather than scheduled as a two-hour session on a Tuesday.
Culture decides what the technology is allowed to become
The same platform, deployed in two divisions of the same company, will often produce two entirely different outcomes, and the difference is rarely explained by the configuration. One division treats raising a problem early as safe and normal, adapts the tool, and fixes friction as it appears. The other lets small problems accumulate silently, because admitting one reads as admitting weakness, until the rollout is quietly declared a failure in a meeting nobody records.
This is why transformation programmes that begin with a technology audit and end with a culture problem are not unusual. They are close to the default outcome, because the technology audit is the part everyone knows how to run, and the culture work is the part almost nobody in the organization has been trained, incentivised or given the authority to lead.
What actually has to be led
None of this argues against ambitious technology investment. It argues for sequencing. Before a single licence is purchased, a serious leadership team should be able to answer what specific behaviour is expected to change, who currently benefits from that behaviour not changing, what will be visibly different about how success is measured, and who is accountable for the uncomfortable, unglamorous work of helping people be bad at something in public for a while.
Transformation gets sold as a technology decision that happens to involve people. Run it that way and you get an impressive system quietly running an unchanged organization. Run it as the leadership decision it actually is, one that happens to involve technology, and the system itself becomes almost incidental, because by then you’ve already done the part that was genuinely difficult.