A wrong name rarely stays a naming problem. Within a quarter it has become a plan, a governance model and a budget — all built for work that nobody is actually doing.
Every experienced manager has watched somebody stop a meeting over a word. The room usually reads it as pedantry: we all know what we mean, let us move on. Sometimes that reading is correct. More often it is expensive.
In transformation and project work, a label is not a description of something that already exists. It is a configuration instruction. Call a piece of work a project and you have implied a defined scope, one accountable manager, a baseline and an end date. Call it a programme and you have implied outcomes, tranches and deliberate re-planning. Call it a transformation and you have implied that the organisation will be structurally different when it is over. Governance, reporting, resourcing and executive expectations are then all built on top of that noun. If the noun was wrong, everything built on it is wrong in the same direction — and nobody notices until the variance arrives at a steering committee dressed up as an execution problem.
What follows are the confusions I meet most often, and what each one actually costs.
A project you do not control is not your project
A supplier, a corporate centre or a joint-venture partner is running the work. Your organisation contributes requirements, receives the output and reports on progress. In the internal portfolio this appears as “our project”, with a project manager assigned to it. In reality the role is that of a stakeholder and a beneficiary, and the work being performed is delivery controlling or assurance.
The consequence is a commitment made without the levers to honour it. Dates are reported upward as though they can be moved internally; they cannot. When the schedule slips, the receiving organisation absorbs the accountability, while the remedies that would actually matter — acceptance criteria, contractual escalation, payment milestones — were never anyone’s job to negotiate, because the work was filed as delivery rather than as oversight. The mirror-image cost is just as real: assurance is genuine, demanding work, and when it is disguised as project management it gets neither the seniority nor the independence it needs.
A programme managed as a project will behave like neither
A project is a commitment to produce a defined output. A programme is a commitment to an outcome under conditions where the scope will legitimately change as the organisation learns. The two are governed differently on purpose. Treating the second as the first is the most common category error in transformation, and it is usually made at the moment of funding — because a project is easier to approve than a programme.
The consequences arrive on a predictable schedule. Re-baselining, which is a normal instrument of programme control, is read by the steering committee as failure. The recurring question becomes “when will it be finished” rather than “is the benefit still available and still worth the remaining spend”. Components optimise locally, because each project manager owns their scope and no one owns the interdependencies between them. Then, six or nine months in, the organisation is surprised that its large project has started behaving like a portfolio — which it always was. That distinction, and what it demands of the person leading it, is the subject of a separate piece on programme management.
If everything is a project, the portfolio stops telling you anything
Open almost any project register and you will find a replacement forklift fleet listed beside entry into a new market, and both listed beside an item that consists of a title, no scope, no owner and no estimate. Uniqueness, complexity and the degree to which the work is already proceduralised have all been ignored. Everything received the same noun because the noun is how work becomes visible.
Two costs follow. First, prioritisation is asked to compare things that are not comparable, so capacity gets counted in items rather than in the organisation’s actual ability to absorb change, and the list quietly stops being a decision instrument and becomes an inventory. Second, project governance gets applied to routine, repeatable work, which teaches the organisation that project discipline is bureaucracy — after which it is resisted precisely where it is needed. Most of this is decided before prioritisation ever begins, at the intake gate.
Change improves what exists; transformation replaces it
Rob Llewellyn draws the line bluntly. Organisations that pour their effort into improving what they already run are, in his phrase, “making faster caterpillars instead of creating butterflies” — and that, he argues, is exactly the difference between change and transformation. Change maintains and modernises the organisation. Transformation creates what does not yet exist. Both are legitimate activities. Only one of them justifies rebuilding the operating model.
Mislabelling costs in both directions. An improvement initiative granted transformation-scale budget, patience and executive attention will deliver modernisation and then be judged against a promise that was never within its scope — and the sponsor who approved it carries that verdict. In the other direction, genuine transformation funded and governed as an improvement project gets a nine-month horizon, a single functional sponsor and a cost-reduction business case, and dies at the first point where it needs authority over the operating model. There is a slower cost as well: when every initiative is called a transformation, the word stops mobilising anybody, and the one occasion when the organisation truly needs to be moved arrives with the vocabulary already spent.
A PMO that projects can bypass is not covering anything
“We have a PMO” is a statement about coverage. It tells the board that initiatives are registered, reported against a standard and visible in one place. Then, in practice, passing through the PMO turns out to be optional — a matter of goodwill, seniority or how much friction the sponsor is willing to tolerate.
The board is now steering by a subset of the portfolio, and it is not a random subset. The initiatives that route around the PMO are reliably the politically sponsored ones, the fast-tracked ones and the ones whose numbers would not survive a standard review. The blind spot therefore correlates with the risk. Compliance that is opt-in binds only the teams that were already compliant, so the reporting burden falls on the well-behaved while the exposure sits somewhere else entirely. And the PMO is held accountable for the completeness of a picture it has no right to complete. When the board eventually asks why nobody saw a problem coming, the honest answer is that seeing it was never mandatory.
The same three letters can mean a reporting desk or a stop button
The second confusion around those letters concerns authority rather than coverage. A function that consolidates status reports and a function that can halt, re-sequence or refuse to onboard a project are different organs, with different costs, different seniority and different relationships to the executive. They share an acronym.
So the board approves control and receives administration. The role is scoped and paid against the label rather than the mandate, which shapes who accepts it. Escalations arrive at a function whose only available action is to forward them, and delivery latency is then attributed to the escalating projects. Eventually somebody concludes that the PMO adds no value — when what added no value was a name issued without a decision right attached. This is why the charter comes before the manager: it is the document in which the organisation is forced to say which of the two things it is actually buying.
A sponsor who cannot decide is a contact
The name appears on the charter. The person is senior, engaged and genuinely supportive. They also cannot release budget beyond a threshold, and they cannot settle a priority conflict between two functions that do not report to them. That is not a sponsor. That is a well-intentioned point of contact.
Escalation then reaches somebody who must escalate again, and decision latency compounds where nobody is measuring it. Weeks later it surfaces as delivery variance, and the delivery lead is assessed on the velocity of a system whose decision cycle they do not control. There is a second, quieter effect: nominal sponsorship insulates the real decision-maker from the consequences of not deciding, because on paper the decision already has an owner.
Delivered and worth it are two different states
Closure is declared at go-live. The system is in production, the process is documented, the reorganisation has been announced. The initiative moves to green and then off the portfolio entirely. What has been established is that the money was spent as intended, not that the money was worth spending.
The visible consequence is a portfolio reporting near-total success against a profit and loss statement that has not moved, which corrodes executive confidence in reporting generally — and confidence, once lost, is not recovered by better dashboards. The structural consequence is that closure happens at precisely the moment when cost is fully incurred and value is entirely unrealised, so no one is ever asked to collect the benefits that justified the investment. The next business case is then written by people who were never held to the last one.
“What do you mean by that?” is a control, not an interruption
All of the above is preventable with two questions that cost less than a minute of meeting time. The first is: what do you mean by that? It asks for the definition of a word, not for the competence of the person using it, and it is remarkable how often the room contains three incompatible answers. The second is: did you really mean X, because then Y follows. That one makes the downstream consequence explicit while it is still cheap to change, and it offers the other person a graceful exit if X was not what they meant.
Managers hesitate to ask because asking looks like not knowing. In practice the reverse is true. The person willing to interrogate a word is almost always the person who has seen what happens when nobody did — who has watched an assurance role reported as delivery ownership, or a programme re-baselined into a credibility crisis, or a benefit quietly abandoned at go-live. Asking about the noun is the earliest, cheapest and least political intervention available in transformation governance. Every later instrument — a gate, a re-baseline, a steering escalation — costs more and arrives after the commitment has already been made.
An apple called a banana is still an apple. The plan built around it is not.