It does not have to be. The organizations that move fastest are rarely the ones with the least governance. They are the ones whose governance tells people clearly what they may decide on their own, and then leaves them to decide it. What follows is how that kind of governance is built, what the evidence says about it, and how to tell whether yours is the other kind.

The false trade-off

The usual picture is a dial with control at one end and agility at the other: turn it toward control and you are safe but slow, turn it toward agility and you are fast but exposed. Most governance debates are arguments about where to set the dial.

The picture is wrong. Governance becomes slow when it is designed as a series of gates, where every decision of a certain kind has to stop and wait for a signature. A gate is the simplest control to design and the most expensive to run, because its cost grows with every decision that passes through it, and in a growing business that is all of them.

The cost is rarely counted, and it is larger than it looks. When Bain & Company studied the time budgets of seventeen large corporations, it found one where employees spent 300,000 hours a year supporting a single weekly executive committee meeting. The committee itself spent 7,000 hours in the room; its eleven members then held meetings with their unit heads that took another 20,000, and the cascade went on from there. Bain is careful to say there is no evidence that all of that time was wasted. The point is that nobody had seen it.

The drift toward gates is natural, which is why it has to be resisted on purpose. Jeff Bezos described it to Amazon’s shareholders in his 2015 letter: “As organizations get larger, there seems to be a tendency to use the heavy-weight Type 1 decision-making process on most decisions, including many Type 2 decisions. The end result of this is slowness, unthoughtful risk aversion, failure to experiment sufficiently, and consequently diminished invention.”

There is another way to design the same control. Instead of checking each decision, you decide once where the limits are, who may act inside them, and what must happen when someone wants to go outside them. That is not less governance. It is governance that does its work up front, once, rather than at every step.

Govern the boundaries, not every decision

Seneca put the first condition simply, nearly two thousand years ago, in a letter to Lucilius: “Our plans miscarry because they have no aim. When a man does not know what harbour he is making for, no wind is the right wind.” A boundary is the governance version of a harbor: a decision taken in advance on behalf of many decisions to come. Three kinds do most of the work.

The first is how much risk the organization is willing to carry: which risks it accepts, which it insures or shares, and which it refuses whatever the return. The UK’s Financial Reporting Council, in its guidance to the Corporate Governance Code, puts this squarely on the board: “The board is responsible for determining the nature and extent of the principal risks the company is willing to take in order to achieve its long-term strategic objectives.” Written down in plain language, on a page, that answer settles a large share of the questions that otherwise travel up the hierarchy one at a time.

The second is money and authority. A team with a budget envelope and the authority to spend inside it does not need a signature for each purchase. It needs a clear envelope, a regular look at how it is being spent, and a known route for asking for more. The same FRC guidance recommends a formal schedule of the matters the board keeps for itself, because it will “provide clarity to all over where responsibility for decision-making lies.” Everything not on that schedule is, by implication, someone else’s to decide.

The third is the short list of things that are never negotiable: the legal and regulatory obligations, the commitments made to clients, the few principles the organization will not trade. This list should be short. A long one is usually a list of preferences dressed up as principles, and it brings the gates back by another door.

Inside those boundaries, people decide. Outside them, they ask. The governance work is in drawing the lines well and keeping them current, not in reviewing what happens inside them.

Sort decisions by how hard they are to undo

Not every decision deserves the same care, and treating them alike is how governance becomes slow. The most useful single question about a decision is not its size but how hard it is to reverse.

Bezos’s 2015 letter gives the clearest modern statement of the idea. “Some decisions are consequential and irreversible or nearly irreversible – one-way doors – and these decisions must be made methodically, carefully, slowly, with great deliberation and consultation.” But most decisions, he wrote, “are changeable, reversible – they’re two-way doors,” and “Type 2 decisions can and should be made quickly by high judgment individuals or small groups.”

A two-way door can be undone next week at little cost. It should be taken quickly, close to the work, by the person who knows most about it. If it turns out wrong, it is corrected. Slowing it down buys almost no safety and costs a great deal of time.

A one-way door deserves the full treatment: the right people in the room, the options written down, the reasoning recorded. Signing a long contract, entering a new market, changing the structure of the organization, committing to a system that everything else will be built on.

Most organizations apply the full treatment to far too many decisions, because nobody ever sorted them. Sorting them, once, with the leadership team, is often the quickest gain available: a fast lane for what can be undone, and real attention for what cannot.

Short cycles of review instead of long cycles of approval

Speed does not come from removing oversight. It comes from changing its rhythm.

An approval sits in front of the work and stops it. A review sits behind the work and learns from it. An organization that reviews its portfolio every quarter and its main risks every month can afford to let work start without waiting, because it will see within weeks whether something is going wrong.

Bezos made the same trade explicit in his 2016 letter: “most decisions should probably be made with somewhere around 70% of the information you wish you had. If you wait for 90%, in most cases, you’re probably being slow. Plus, either way, you need to be good at quickly recognizing and correcting bad decisions. If you’re good at course correcting, being wrong may be less costly than you think, whereas being slow is going to be expensive for sure.” Course correcting is exactly what a short review cycle provides.

The same applies to the rules themselves. Every policy, every committee and every threshold should carry a date on which it is looked at again and kept, changed or retired. The FRC asks the same of the risk appetite, which it says should be “regularly reviewed and evaluated.” Rules written for the organization of three years ago are one of the main sources of the slowness people blame on governance. A review date is the cheapest protection against them.

Visibility replaces control

A board or a leadership team that cannot see what is happening has only one tool left, which is to require that everything come to it for approval. A board that can see does not need that tool.

Some of the fastest-moving companies write this principle down. The Netflix culture memo asks managers to practice “context not control — giving their teams the context and clarity needed to make good decisions instead of trying to control everything themselves,” and states plainly: “We avoid decision-making by committee, which tends to slow companies down and undermine accountability.” The memo is equally clear that this is not a license for managers to look away; context only works when it is kept true.

Visibility here means something modest: decisions recorded in one place, with who took them and why; the state of the portfolio on a page that can be read in a few minutes; the main risks and the actions against them, kept current. None of it needs to be elaborate. It needs to be true and up to date.

When that picture exists, the board steers by looking rather than by signing. It asks better questions, it intervenes earlier and less often, and the people doing the work stop waiting for permissions that were only ever a substitute for information.

Sized to the organization

“Never use a one-size-fits-all decision-making process,” Bezos wrote in the 2016 letter, and the same is true of governance as a whole. What it looks like depends on where the organization is, which is why we measure before we design.

A young or small organization needs very little: decision rights on a single page, one meeting with an agenda and a record, and review dates on what it writes down. We have written separately about governance for a forty-person company, and most of what it says applies here.

A larger or more mature organization needs the same ideas with more structure: delegated authorities by level, a risk appetite the board has actually discussed, a portfolio review that leads to decisions, and committees whose terms say what they decide and what they do not. The principle does not change with size. The amount of paper does.

Signs it is working, and signs it is not

Governance that keeps its speed tends to show the same signs. People can say, without checking, what they may decide on their own. Routine decisions are taken in days, not weeks. Meetings of the leadership team spend their time on the few decisions that cannot be undone. Policies have dates on them, and some have been retired.

Governance that has become a brake shows the opposite. The same kinds of decisions keep climbing to the top. Approvals are chased by email. Committees review things they cannot change. Nobody can remember why a rule exists, and nobody has the authority to remove it.

If the second list sounds familiar, the remedy is rarely a new framework. It is a short, honest look at where decisions are actually taken, which of them can be undone, and which boundaries are missing. That is the Map stage of how we work, and it is useful on its own.

Sources

• Jeffrey P. Bezos, 2015 Letter to Shareholders, Amazon.com, Inc., April 2016, filed with the US Securities and Exchange Commission.
• Jeffrey P. Bezos, 2016 Letter to Shareholders, Amazon.com, Inc., April 2017.
• James Allen, The 300,000-hour meeting, Bain & Company, May 2014, on research by Michael Mankins, Chris Brahm and Greg Caimi published in Harvard Business Review as “Your Scarcest Resource,” May 2014.
• Financial Reporting Council, Corporate Governance Code Guidance, accompanying the UK Corporate Governance Code 2024.
• Netflix, Netflix Culture Memo, section “People over process,” current edition.
• Seneca, Moral Letters to Lucilius, Letter 71, translated by Richard M. Gummere, Loeb Classical Library.

The research says the same. Monique Aubry, who with Brian Hobbs led a long research program on PMOs that included a survey of 500 of them, summed it up for practitioners in 2016: “Research just shows that 80% of organizations changed their PMO after 3 years.” The reasons, she added, “are various and not necessarily associated with something wrong.”

Both sides of the argument are right about something. A certain kind of PMO is disappearing, and it deserves to. The work underneath it is not disappearing at all; there is more of it than ever. What follows is the case against the PMO, taken seriously, and what is left once that case has been heard.

The case against

The argument against the PMO is not made by people who dislike structure. It is made by people who have watched one at work.

They have seen status reports produced every two weeks and read by nobody. They have seen templates that take longer to fill in than the work they describe. They have seen an office that asked for information and gave nothing back, that measured compliance with its own method rather than whether projects delivered anything, and that slowed every team down in the name of consistency. Peter Drucker named the problem in Harvard Business Review in 1963, long before anyone had a PMO to complain about: “There is surely nothing quite so useless as doing with great efficiency what should not be done at all.” The wording has been traced to that article by Quote Investigator.

They have also seen teams that organize themselves well without any of it: short planning cycles, visible boards, regular demonstrations of working results. Next to that, a central office collecting reports looks like an overhead from another era.

All of this is fair. We have written separately about why PMOs are abandoned, and the design decision that prevents it, and the pattern is real.

What actually disappeared

Look closely at what the critics object to, and most of it is clerical: chasing people for status, consolidating spreadsheets into a monthly pack, checking that documents exist and follow a format.

That work is disappearing, and it should. Modern tools collect much of it as a side effect of the work itself. A team that plans and tracks its work in a shared system does not need someone to ask it what it did last week. The time an office used to spend gathering information can now go somewhere more useful, provided the office knows where.

The Project Management Institute saw the shift coming in its 2018 Pulse of the Profession, a survey of more than 5,400 practitioners, executives and PMO directors. Organizations, it reported, were empowering the PMO “to shift the focus from an administrative function to one that manages value delivery,” and it added: “The strategic role of the PMO and enterprise project management office (EPMO) will continue to be vital, no matter what they are called.”

An office whose whole value was in the clerical work has nothing left to offer. That is the PMO that is disappearing.

What did not disappear

Now look at the questions no team can answer for itself.

Which of the forty things the organization wants to do should it actually do this year, given the people it has? Which can wait, and which should stop? When two projects need the same five specialists in the same quarter, who decides, and on what basis? When one project’s delay will break three others, who sees it coming? When a project finishes, did the benefit it promised arrive, or did the organization simply move on to the next one?

These are portfolio questions, and most organizations answer them badly by not answering them at all. A McKinsey study of more than 1,600 US companies over fifteen years found that “most companies allocate the same resources to the same business units year after year”: for one business in three, the capital it received in a given year was almost exactly what it had received the year before. The companies that moved resources most actively earned, on average, 30 percent higher total returns to shareholders each year than those that moved them least. The study is about capital across business units, not projects, but the lesson carries: the choosing is where the value is, and the default is not to choose.

Michael Porter put it in one sentence in Harvard Business Review’s “What Is Strategy?” in 1996: “The essence of strategy is choosing what not to do.”

These questions do not go away when teams become more autonomous. They become sharper. An organization running more work in parallel, with more of it done by small self-directed teams, has more dependencies, more competition for the same scarce people, and more opportunities to start things it never finishes. Autonomy at the level of the team makes coordination at the level of the portfolio more necessary, not less.

PMI’s newest edition of the Pulse, published in 2026, names the same conditions: “The factors behind organizational complexity — unclear decision authority, competing priorities and misaligned objectives — are precisely the conditions that a well-functioning PMO is designed to address by ensuring the organization has the governance, visibility and relational infrastructure to navigate it.” In its survey of more than 2,500 professionals and senior leaders, those working in organizations with a PMO were more likely to say they managed complex projects very or extremely well (63% against 57%). The gap is modest, and it is self-reported, but it points the same way.

Somebody has to help the leadership team answer those questions with real information. That is the work that remains.

The PMO that stays

The office that survives looks different from the one that is disappearing, in four ways.

It supports decisions instead of collecting reports. Its main product is the information a leadership team needs to choose: what is underway, what it costs, what it needs, what is at risk, and what would happen if something stopped. Its main event is the portfolio review where those choices are made.

It manages capacity, not just projects. It knows how many people with which skills the organization actually has, and it says plainly when the list of commitments is longer than the people available to deliver it. That one sentence, said early and with evidence, is worth more than a year of status reports.

It follows value, not activity. It asks, after delivery, whether the benefit arrived, and it brings the answer back to the people who will choose the next round of work. Few organizations do this well: in PMI’s 2018 survey, only one in three reported high maturity in realizing benefits, and PMI described the common failure as managing projects on time, scope and budget “without consistently tracking whether they help the organization achieve its larger strategic goals.”

It fits the method to the work. Some projects are best run in short iterative cycles, some need a careful plan because the cost of getting them wrong is high, and many are a mix of both. PMI’s 2024 Pulse found the use of hybrid approaches rising from 20% of respondents in 2020 to 31.5% in 2023, and reported that “the project management approach adopted alone has no significant impact on higher project performance rates.” The method is not what decides success; fitting it to the work is. The office that stays helps each team choose a way of working that fits, rather than imposing one method on everything. It offers services people ask for, and it earns its authority by being useful.

Its shape depends on size

In a small organization, the PMO may be one person, or part of one person’s role: keeping the list of everything underway, preparing a monthly review with the leadership team, and helping project leads who are in difficulty. A spreadsheet is enough.

In a larger organization, the same function becomes an office close to the executive team, sometimes called a strategy or enterprise office, which links the portfolio to the strategy, manages capacity across departments and reports on benefits. The tools become more serious because there is a practice for them to support.

The size is set by the maturity the organization actually has, not by what a larger organization has. We measure first and design to the level.

Is yours the old kind?

Five questions tell you quickly.

When the PMO’s report arrives, does a decision follow it? If not, the report is serving the office, not the organization.

Can the PMO say whether the organization has more work committed than people to do it? If not, it is missing its most valuable piece of information.

Has the PMO ever recommended stopping a project? An office that never recommends stopping anything is counting projects, not managing a portfolio. Drucker’s test for every activity, set out in his writing on what he called planned abandonment, is the right one for every project on the list: “If we did not do this already, would we, knowing what we now know, go into it?”

Do teams come to the PMO for help, or only when they are told to? Requests for help are the best measure of an office’s usefulness.

Does anyone check, after delivery, whether the promised benefit arrived? If nobody does, the organization is measuring effort, not results.

Why it stays

The PMO stays because organizations still have to choose, and choosing well across many competing pieces of work is hard. It changes because what they need from it is help with choosing, not control over doing. An office built around that difference is not an overhead. It is how a leadership team sees its own organization clearly enough to steer it.

If your organization has an office that is being worked around, or is wondering whether it needs one at all, the starting point is the same: a short look at the list of work, the people available to do it and the forum where choices are made. That is the Map stage of how we work, and it is useful on its own.

Sources

• Monique Aubry, Organizing for Projects: The Integrative Role of PMO, PMO Conference, London, 2016, drawing on Hobbs and Aubry, The Project Management Office: A Quest for Understanding (PMI, 2010), and Aubry, Hobbs, Müller and Blomquist, Identifying the Forces Driving Frequent Change in PMOs (PMI, 2011).
• Peter F. Drucker, “Managing for Business Effectiveness,” Harvard Business Review, May 1963; wording and attribution checked by Quote Investigator, 2017.
• Peter F. Drucker, Drucker on Planned Abandonment, Drucker Institute, adapted from Management Challenges for the 21st Century (1999) and earlier works.
• Stephen Hall, Dan Lovallo and Reinier Musters, How to put your money where your strategy is, McKinsey Quarterly, March 2012.
• Michael E. Porter, What Is Strategy?, Harvard Business Review, November–December 1996.
• Project Management Institute, Pulse of the Profession 2018: Success in Disruptive Times, 2018.
• Project Management Institute, Pulse of the Profession 2024: The Future of Project Work, February 2024.
• Project Management Institute, Pulse of the Profession 2026: Driving Success in Complex Projects, May 2026.

That is the problem in one story. Each step made sense to the team that owned it. Nobody owned the whole, and the whole is the only thing the customer experienced. Customer journey analysis exists to make that whole visible, and it has moved from a good idea to something most organizations can no longer do without.

The principle is old. Adam Smith wrote in The Wealth of Nations in 1776 that “Consumption is the sole end and purpose of all production; and the interest of the producer ought to be attended to only so far as it may be necessary for promoting that of the consumer.” He was arguing about trade policy, not service design, but the sentence describes the journey view exactly: the organization’s parts matter only as far as they serve the person at the end of them.

What changed

Three things changed, and together they make the analysis hard to avoid.

Customers now move between channels as a matter of course. They read on the website, ask on a messaging app, call when something goes wrong and sometimes still come to the counter. They expect each channel to know what happened in the others, and they notice immediately when it does not.

Customers also compare you with the best experience they have had anywhere, not with your competitors. Someone who can rebook a flight in two taps on their phone does not understand why changing a delivery address takes three days, and they are not interested in the reasons. Governments have recognized the same comparison: as long ago as 1993, a US executive order required federal agencies to “benchmark customer service performance against the best customer experience provided in the private sector,” as the 2021 order that renewed the effort recalls.

And switching is easy. In most markets the cost of leaving is a few minutes, so the cost of friction is paid quietly, in customers who simply stop coming back. That is expensive. Harvard Business Review, summarizing research by Frederick Reichheld of Bain & Company, reports that “increasing customer retention rates by 5% increases profits by 25% to 95%.” The range is wide because businesses differ, but the direction is not in doubt.

Public services are not exempt. Citizens bring the same expectations to a government office that they bring to a shop, and the gap is felt in trust rather than in revenue. The 2021 US order put it in so many words: lost time “operates as a kind of tax—a ‘time tax’—and it imposes a serious burden on our people as they interact with the Government.”

One journey, many departments

The customer sees one path from the first question to the last follow-up. Inside the organization, that path crosses sales, operations, finance, customer service and often an outside partner or two. Each owns its part, measures its part and improves its part.

The places where it breaks are almost never inside a department. They are in the hand-offs between them: the information that does not travel, the case that waits in a queue nobody watches, the rule that one team applies and another team has never heard of. No department can see those breaks from where it sits, because each one sits on a boundary.

Measuring one touchpoint at a time does not reveal them; it hides them. Three McKinsey partners, writing in Harvard Business Review in 2013, warned that a narrow focus on satisfaction at individual moments “can create a distorted picture,” suggesting that customers are happier with the company than they actually are, and drawing attention away from the end-to-end journey.

A journey analysis is the one view that crosses all the boundaries at once, in the order the customer crosses them.

What a journey analysis is

At its simplest, it is a careful description of what a particular kind of customer does to get a particular thing done: the stages they go through, the points where they touch the organization, what they are trying to achieve at each step, and how it feels to them.

The useful version adds the inside view. Behind each step the customer sees, it shows the process that serves it: who does what, which systems are involved, where the work waits, where it is checked and where it is handed on. That second layer is what turns a description of the experience into a diagnosis of its causes.

It is built from evidence, not from opinion. Complaints and their reasons, call records, the time cases actually take, what customers say when they are asked directly, and what front-line staff know and are rarely asked. A journey drawn in a workshop from what managers believe happens is a picture of the organization’s assumptions, and it is usually more flattering than the truth.

How much more flattering was measured long ago. When Bain & Company surveyed 362 firms, 80% believed they delivered a “superior experience” to their customers. Asked the same question, customers said only 8% did. The study dates from 2005, but the gap it named, between what management believes and what customers live, is exactly the gap a journey analysis closes.

What it reveals

When the two layers are laid side by side, the same findings appear again and again.

Steps that are repeated because information collected once is not shared. Long waits that no one notices, because each team’s own part is fast. Dead ends where a customer is sent elsewhere with no one following up. Rules that make sense inside the organization and none at all from outside it.

And a cost most organizations never measure: the work created by failure. The second call because the first one did not resolve the issue, the complaint, the correction, the refund, the escalation. The systems thinker John Seddon gave it a name, failure demand, which the Open University’s course on service management defines as “demand caused by a failure to do something or do something right for the customer.” In one study of 1,200 calls to a council housing service, 80% were failure demand.

That was one service. A larger and more recent figure comes from the UK’s National Audit Office, reporting in 2024 on the tax authority, HMRC: “HMRC has estimated that 72% of calls in 2023-24 were caused by ‘failure demand’, which includes calls caused by HMRC’s process failures or delays, customers chasing progress and customers’ errors. This proportion has increased from 65% in 2018-19.” Nearly three calls in four, in other words, existed because something earlier in the journey had gone wrong.

That work is expensive, and much of it disappears once the step that caused it is fixed. A journey analysis is often the first time an organization sees how much of its effort goes into repairing its own friction.

From map to change

A map changes nothing by itself. The value is in what follows, and what follows is a piece of change management.

Choose a few moments, not all of them: the points where the most customers struggle, or where the failure costs the most. Trying to fix the whole journey at once spreads the effort too thin to fix anything.

Give each chosen moment an owner who can act across departments. A problem that sits on a boundary between two teams will not be solved by either team alone; it needs someone with the authority to change how both of them work, and a governance forum where that person can bring decisions.

Fix the process first and the technology second. Many journey problems are solved by changing who does what, what information travels with a case, or a rule nobody needed. Where technology is part of the answer, a new form, a connected system, a better app, it is built by a technology partner, Plasis or another, to fit a process that has already been redesigned rather than to automate one that is broken.

Then measure from the customer’s side: how long the whole thing took, how many contacts it needed, whether it was done right the first time. The share of contacts that are failure demand is one of the most honest measures there is, because it falls only when the journey really improves. Measures taken one department at a time will keep telling you everything is fine.

Common mistakes

The most common is the beautiful poster. The journey is mapped with care, printed large, hung on a wall and admired, and nobody is responsible for changing anything on it. Within a year it describes a journey that no longer exists.

The second is mapping the ideal journey instead of the real one. It is tempting to draw what should happen; the value is entirely in seeing what does.

The third is starting with the app. A new digital channel laid over a broken process gives customers a faster way to reach the same dead end.

How a strategy reaches the customer

Most strategies say something about customers. Journey analysis is how that sentence becomes specific: which customers, which moments, which changes, owned by whom. It connects what the leadership team intends with what a customer actually experiences on an ordinary Tuesday.

If your organization suspects that its customers find it harder to deal with than it should be, the first step is modest: one important journey, analyzed honestly from both sides, with the evidence that already exists. That is the Map stage of how we work, and it is useful on its own.

Sources

• Adam Smith, An Inquiry into the Nature and Causes of the Wealth of Nations, 1776, Book IV, Chapter VIII, Library of Economics and Liberty (Cannan edition).
• James Allen, Frederick F. Reichheld, Barney Hamilton and Rob Markey, Closing the delivery gap, Bain & Company, October 2005.
• Amy Gallo, The Value of Keeping the Right Customers, Harvard Business Review, October 2014.
• Alex Rawson, Ewan Duncan and Conor Jones, The Truth About Customer Experience, Harvard Business Review, September 2013.
• The Open University, Capacity and demand management, OpenLearn, on John Seddon’s concept of failure demand.
• National Audit Office (UK), Taxpayers let down by poor HMRC customer service, May 2024, on the report HMRC customer service (HC 726, 2023-24).
• Executive Order 14058, Transforming Federal Customer Experience and Service Delivery to Rebuild Trust in Government, Federal Register, December 2021.

It also has more change than ever to absorb. Gartner research, reported in Harvard Business Review in 2023, found that the average employee experienced 10 planned enterprise changes in 2022, up from two in 2016. Nobody has time for change management partly because everybody is already in the middle of several changes.

The good news is that the department-sized program is not what makes change last. In our experience most of it is volume, and volume is not what is missing. What is missing is a handful of specific moves, made at the right time, by the people who already have the authority to make them.

A statistic worth retiring

Almost every article on this subject opens with the claim that 70 percent of change programs fail. We do not, because the number does not hold up. When Mark Hughes of the University of Brighton traced it back through five published instances in the Journal of Change Management in 2011, he concluded that “there is no valid and reliable empirical evidence to support such a narrative.”

Change does fail, often enough. But it does not fail at a known rate, and it does not fail where most programs put their effort.

Where change actually fails

A change is announced, explained and launched. The plan covers all three. Then the project team moves on, the vendor’s support window closes, and the people who have to live with the change go back to work.

That is where it fails: in the weeks after go-live, when the new way is still awkward and the old way is still available. The spreadsheet is still there. The old approval route still works if you know whom to ask. The manager who was never quite convinced says nothing, and the team reads the silence. Six months later the system is used for the minimum the organization can get away with, and everyone agrees it was a disappointing system.

None of this is new. Machiavelli wrote in The Prince in 1513 that “there is nothing more difficult to take in hand, more perilous to conduct, or more uncertain in its success, than to take the lead in the introduction of a new order of things, because the innovator has for enemies all those who have done well under the old conditions, and lukewarm defenders in those who may do well under the new.” He also gave the reason the weeks after launch matter so much: people “do not readily believe in new things until they have had a long experience of them.”

It was not the system. It was the gap between the announcement and the habit, and nobody was watching it.

Five moves that fit in the time you have

Name who is affected, and how

An impact map is one page: down the side, each group whose work changes; across, what they do today, what they will do instead, and what they lose. It takes an afternoon with the people who know the work. Its value is not the document but the conversation, which surfaces the group everyone forgot and the loss nobody mentioned. Machiavelli’s “all those who have done well under the old conditions” are exactly the rows this map exists to find. A change that costs a group something they value, without acknowledgment, will be resisted quietly and for a long time.

Remove the reasons the old way is easier

For each row of the map, ask what makes the old way easier than the new one, and remove it. If the spreadsheet is still accepted, stop accepting it on a named date. If the old approval route still works, close it. If the new system takes longer because the data is wrong, fix the data before go-live rather than asking people to be patient after it.

Behavioral science has made this its first rule. The Behavioural Insights Team, which began in the UK government, opens its EAST framework with it: “If you want to encourage a behaviour, make it easy, attractive, social and timely.” It warns that “Even apparently small features that make a task more effortful can have a disproportionately large effect on behaviour,” and that these “friction costs” are often ignored when a policy, product or service is designed. People do not resist change out of character. They choose the easier path, and the easier path is a design decision.

Give managers the words and the time

The line manager is the change program. Gallup, from its research on manager-led teams, reports that “The manager or team leader alone accounts for 70% of the variance in team engagement.” That is a finding about engagement, not about change, but it describes the same lever: what a team believes about its work runs mostly through one person. If a manager can explain, in their own words, what is changing, why, what it means for their team and what they personally expect, the team will follow. If the manager cannot, no amount of corporate communication will substitute.

So the manager gets a one-page briefing, not a deck: the three sentences that explain the change, the two questions the team will ask and the honest answers, and the one thing the manager should do this week. And the manager gets fifteen minutes in an existing team meeting, not a town hall. Reaching all of the managers this way takes one person a few days.

Measure the habit, not the attendance

Training attendance tells you who sat in a room. Adoption is whether the work is now done the new way. The difference is well documented: the same EAST report cautions that “Changing behaviour is different from changing people’s intentions, beliefs or attitudes,” and that people often intend something they then do not do, a tendency it calls the “intention-action gap.”

Pick one measure per affected group that shows the habit: the proportion of orders entered in the new system, the number of approvals still arriving by email, the days between a request and its first action. Read it weekly, out loud, in the leadership meeting. A measure that is read is a measure that moves, because the manager whose number is lagging will act before being asked.

Stay past go-live

The project ends at go-live. The change does not. Habits take longer to form than a project plan allows. In a study published in the European Journal of Social Psychology in 2010, Phillippa Lally and her colleagues at University College London followed people adopting a new daily behavior; the time it took them to reach their peak of automaticity “ranged from 18 to 254 days.” The often-quoted figure from that study is about 66 days, and Lally herself, asked in 2026 whether 66 days is a rule, answered: “No. I mean for someone somewhere for one habit yes, but for most people most of the time, no.” Either way, it is longer than the vendor’s support window.

And what is not sustained tends to fade. The EAST report describes a megastudy of 54 four-week programs to encourage exercise: 45% increased gym visits during the programs, but “Only 8% of these programmes showed significant changes in behaviour following the four-week intervention period.” Gym visits are not order entry, but the shape is the same: support that stops at launch buys a launch.

Someone has to keep reading the measure and keep closing the easier paths for the first quarter, which is exactly the period when the project team has been reassigned. Naming who does that, and giving them the standing to act, is the single most neglected decision in a small organization’s change.

What to drop

Drop the newsletter. Drop the roadshow. Drop the generic training course that teaches every screen to everyone. Drop the readiness survey nobody will act on. Each of these consumes the little time there is and produces the appearance of a program without the substance of one. If any of them earns its place later, it will be because a measure said so.

Who does this when nobody has time

The line managers do most of it, inside the jobs they already have, because the moves above are mostly management: deciding, explaining, closing paths, reading a number. What they need is a coordinator for a few hours a week whose job is to keep the list of moves honest, keep the impact map current, and bring the measures to the meeting. In a small organization that is often someone from the project team, kept on for a quarter after go-live at a fraction of their time. It costs far less than the program nobody had time for, and it is the part that works.

The honest limit

None of this rescues a change that is wrong. If the system does not fit the work, if the reorganization solves a problem the organization does not have, no sequence of moves will make people adopt it, and it would be wrong to try. The impact map usually reveals this early, in the row where the loss is large and the gain is nobody’s. When we see that, we say so, before go-live rather than after.

Change management when nobody has time is not a lighter version of the program. It is the part of the program that was doing the work all along, done by the people who were always going to have to do it, with someone keeping count. That fits in six months and a Tuesday afternoon.

Sources

• Cian Ó Morain and Peter Aykens, Employees Are Losing Patience with Change Initiatives, Harvard Business Review, May 2023, reporting Gartner research.
• Mark Hughes, Do 70 per cent of all organizational change initiatives really fail?, Journal of Change Management 11(4), 2011.
• Niccolò Machiavelli, The Prince, chapter VI, 1513, translated by W. K. Marriott, Project Gutenberg.
• The Behavioural Insights Team, EAST: Four simple ways to apply behavioural insights, revised edition, 2024 (first edition 2014).
• Gallup, How to Improve Employee Engagement in the Workplace, drawing on State of the American Manager, 2015.
• Phillippa Lally, Cornelia H. M. van Jaarsveld, Henry W. W. Potts and Jane Wardle, How are habits formed: Modelling habit formation in the real world, European Journal of Social Psychology 40(6), 2010.
• University of Surrey, Does it really take 66 days to form a habit? We asked expert Dr Pippa Lally, January 2026.

A forty-person company is not a small case. In Canada, according to Innovation, Science and Economic Development Canada, 98.2% of the country’s 1.10 million employer businesses are small, with 1 to 99 employees, and 95.6% have fewer than 50. Small businesses employ 5.8 million people, 46.6% of the private labor force. Most of the economy is run by companies for which the governance literature was not written.

At some point someone in such a company opens a governance framework, and the trouble starts.

What the frameworks are for

The best-known frameworks were written for a particular kind of organization, and most of them say so.

The G20/OECD Principles of Corporate Governance, revised in 2023, state plainly: “The Principles focus on publicly traded companies, both financial and non-financial.” They allow that “some of the Principles may be more appropriate for larger companies than for smaller ones.” The national corporate governance codes address the same listed companies with dispersed shareholders. Even the United Kingdom’s code for private companies, the Wates Principles, serves a reporting duty that applies to companies with more than 2,000 employees, or a turnover above £200 million and a balance sheet above £2 billion. The employee threshold alone is fifty times the size of the company in this article.

COBIT addresses the governance of information and technology in enterprises large enough to have a board, an executive and an audit function as separate bodies, and ISACA, which publishes it, is candid about the fit: “experience and research have shown there is no such thing as a one-size-fits-all governance system for enterprise I&T,” and “Any COBIT 2019 implementation will, thus, require tailoring before it can be fit-for-purpose,” as one of its own experts wrote in 2019. ISACA later published a guide for enterprises of up to 250 employees, which says much about how far the full framework sits from them.

ISO 37000 describes the governance of organizations in general terms, and the committee behind it says it “is applicable to all organizations regardless of type, size, location, structure or purpose.” That is true of its principles. But its vocabulary assumes a governing body distinct from management, and in a forty-person company the two are the same people. The public-sector codes assume a legislature, an accounting officer and an audit institution.

None of this is wrong, and much of it is wise. But it is written for organizations where the people who own, the people who govern and the people who manage are different people, often hundreds of them. In a forty-person company the same five people own, govern and manage, and they do it before lunch.

We treat the frameworks as libraries, not laws. The question is never whether to adopt one. It is which three or four ideas to borrow, how to write them small enough to be kept, and what to leave on the shelf with a note saying when to fetch it.

Borrow: decision rights on one page

The single most useful idea in every framework is that it should be known, in advance, who decides what. It is also one of the oldest ideas in management. In the Book of Exodus, Moses is judging every dispute of his people himself, from morning to evening, until his father-in-law Jethro tells him: “this thing is too heavy for thee; thou art not able to perform it thyself alone.” The remedy is a delegation by the weight of the matter: “every great matter they shall bring unto thee, but every small matter they shall judge: so shall it be easier for thyself, and they shall bear the burden with thee.” Every founder who has become the bottleneck will recognize the conversation.

In a forty-person company this fits on one page. Down the side, the kinds of decisions that recur: hiring and dismissal, spending above a threshold, signing a contract, taking on a new client or dropping one, changing a price, starting or stopping a project. Across the top, who decides: the founder or managing director, the leadership team together, a department head, a manager. In each cell, one mark, and where useful a threshold: a manager spends up to this amount alone, the leadership team above it, the founder above that.

Writing the page takes an afternoon. The hard part is the conversation it forces, which is about what the founder is prepared to let go of. That conversation is the governance work; the page is only its record.

Borrow: one meeting with an agenda and a record

Every framework has a committee structure. A forty-person company needs one meeting, and it needs that meeting to be different from the Friday catch-up.

The leadership meeting has a standing agenda: decisions needed, decisions taken since last time, the few numbers that matter, and the risks anyone is worried about. It has a record, which is a decision log, not minutes. One line per decision: what was decided, by whom, when it will be reviewed. The log is the organization’s memory, and it is what lets a new manager, a bank or an auditor see how the company decides without anyone having to remember.

Once a quarter, the same meeting looks at the decision-rights page and the log together and asks what should change.

Borrow: a review date on everything

The frameworks talk about continuous improvement. The usable version of the idea is simpler: nothing is permanent by default. Every rule, threshold and meeting carries a date on which it will be kept, changed or retired. That one habit prevents the slow accumulation of rules nobody remembers agreeing to, which is how governance becomes something people route around.

Leave on the shelf: the committee structure

An audit committee, a risk committee, a nomination and remuneration committee, each with its own charter and cycle: in a company of forty these are the same five people wearing different hats in consecutive meetings. The separation of duties the committees exist to provide cannot be provided by the same people, so the structure adds meetings without adding independence. Even the Wates Principles, written for large private companies, say that “The size of a board should be guided by the scale and complexity of the company.”

The need is real and arrives later. The note on the shelf says: fetch when there is an external investor who wants a seat, when a regulator requires it, or when the company has grown to the point where owning, governing and managing are done by different people.

Leave on the shelf: the full policy set

A framework-led governance project tends to produce forty policies, because the framework lists forty topics. Nobody reads forty policies. The company needs three, each on a page: how spending is authorized, how conflicts of interest are declared and handled, and how a complaint or a whistle is heard. Perhaps a fourth on data, depending on the business. Everything else is a decision right or a habit, and belongs on the one page or in the log.

Leave on the shelf, for now: the assurance layers

Formal risk registers, internal audit, management assertions, independent assurance. These exist to let people who cannot see the work satisfy themselves that it is being done properly. In a forty-person company the people who need to know can walk across the room. A risk conversation on the standing agenda does the job.

The trigger to fetch these from the shelf is usually external: a regulated sector, public money, a large client’s supplier requirements, an acquisition. When one of those arrives, the company that has a decision log and a decision-rights page builds the assurance layer in weeks, because the practice already exists and only needs to be evidenced.

The test

A governance practice fits a forty-person company if a new manager can be taught the whole of it in an hour and can find every rule they need on a few pages. If it takes a training course, it was written for a different organization.

When the company grows

The shelf is not a rejection. It is a plan. Larry Greiner’s classic Harvard Business Review article of 1972, “Evolution and Revolution as Organizations Grow”, described growth as a series of phases, each ending in a period of turmoil, and in HBR’s own summary, “Managers therefore experience the irony of seeing a major solution in one period become a major problem in a later period.” The founder’s direct control that serves a company of forty well becomes its constraint at a hundred.

There is even a rough sense of where informal coordination runs out. Robin Dunbar of the University of Oxford and a colleague, in a 2003 study of personal social networks, found an average maximum network size of 153.5 people, “remarkably close to the group size of 150 predicted for humans on the basis of the size of their neocortex.” The research concerns social networks, not companies, and no one should treat it as a rule. But many founders notice the same thing somewhere around that size: they no longer know everyone, and what used to be settled in the corridor now needs to be written down.

A company that borrows three ideas at forty people, writes them small, and reviews them each quarter has the habits that the larger frameworks assume. At eighty people the thresholds move. At a hundred and fifty, the first committee may earn its place. At the first external investor, the assurance layer comes down from the shelf, already knowing where it fits.

That is the whole of our approach to governance at this size: measure where the company actually is, borrow what that level can carry, name what was left on the shelf, and put a review date on all of it. The frameworks were written by careful people. The care they ask for is in choosing from them, not in adopting them whole.

Sources

• Innovation, Science and Economic Development Canada, Key Small Business Statistics 2025, data as of December 2024.
• OECD, G20/OECD Principles of Corporate Governance 2023, September 2023; the wording quoted was checked in the Financial Stability Board’s compendium copy.
• Financial Reporting Council, The Wates Corporate Governance Principles for Large Private Companies, December 2018.
• Stefanie Grijp, Designing Your Organization’s Custom COBIT, ISACA, March 2019.
• ISACA, Small and Medium Enterprises Seeking to Start a Governance Program in New COBIT Resource, press release, November 2021.
• ISO/TC 309, ISO 37000 – the first ever international benchmark for good governance, on ISO 37000:2021, Governance of organizations — Guidance.
• The Bible, King James Version, Exodus 18, Wikisource.
• Larry E. Greiner, Evolution and Revolution as Organizations Grow, Harvard Business Review, July–August 1972, reprinted 1998.
• R. A. Hill and R. I. M. Dunbar, Social network size in humans, Human Nature 14(1), 2003.

The research agrees that PMOs rarely stay still. A peer-reviewed study in the Project Management Journal opens with the observation that PMOs are “dynamic organizational entities, frequently in transition from one charter and structure to the next.” A later review in the International Journal of Project Management, summarized by the Association for Project Management, notes that “PMOs are often short-lived and their value is hard to quantify.” Not every change is a failure; some offices are reshaped because the organization around them has changed. But many are reshaped because they never earned their place. What follows is that pattern as we see it, why it happens, and the one design decision that prevents it.

The pattern

The first quarter is busy and hopeful. A charter is approved. Templates for a project brief, a plan and a status report are issued. Project managers are asked to report every two weeks, and they do, because the office is new and someone senior asked for it.

By the second and third quarters the reports are arriving on time and nobody is reading them. The leadership team gets a summary it did not ask for, in a format it did not choose, about projects it already knows about from the corridor. The project managers notice that no decision ever follows a report, and the reports get thinner.

In the fourth and fifth quarters the exceptions begin. A project is too small for the template, or too urgent, or run by someone who was never asked to comply. The office asks for more authority to close the gaps. The leadership team asks, usually for the first time out loud, what the office gives them for its cost.

Somewhere in the second year the office is disbanded, or kept as a reporting function with one person and no influence, or renamed. The templates survive in a shared folder. The next office, two or three years later, starts with a new charter and the same assumptions.

Gartner describes the same end in its 2024 strategic roadmap for enterprise PMOs: offices “that do not have a clear identity and purpose are often relegated to an operational role with a narrow focus on execution, or even disbanded.”

Why control-first fails

The office in this pattern was designed as a control function before the organization had anything to control, and before the office had earned any authority to control it.

Authority in an organization of this size is not granted by a charter. It is earned by being useful, repeatedly, to people who can decline your help. A new office has given nothing yet. When its first act is to ask for status, it is making a withdrawal from an account with no deposits. Gartner’s 2026 research on PMO services states the consequence: “When PMO services go underutilized, PMOs risk becoming irrelevant, losing credibility, influence, and their ability to shape consistent ways of working across the enterprise.”

The status it collects then has nowhere to go. A report is only worth producing if a decision depends on it: to stop a project, to move people, to change a date, to say no to a new request. In most organizations at this stage there is no forum where those decisions are taken, so the report informs nothing. A Gartner analyst made the point a decade ago, when the firm set out its best practices for PMOs: “It’s important to check that reporting provides organizational leadership with status information that supports effective decision-making.” People are quick to notice when their effort informs nothing, and they stop making it. The decay of the data is not a discipline problem; it is a design problem.

The link to strategy is weaker than most charters assume. In the Project Management Institute’s 2018 Pulse of the Profession, “Only 41% of organizations with an enterprise-wide project management office (EPMO) report that it is highly aligned to the organization’s strategy.” An office that is not connected to the decisions that matter most will be judged on the paperwork it produces, and paperwork rarely survives that judgment.

Tools make this worse. An office that buys project software in its first quarter has committed the organization to a way of working before the way of working exists. The tool becomes the practice, and it is rarely the right one.

The design decision

The decision that prevents the pattern is taken before the charter is written, and it is this: what will the office give on its first day, to people who are free to ignore it?

An office designed around that question starts differently. On day one it produces a visible list of every project actually underway, with an owner and a next decision against each. Nobody has seen that list before, and everyone wants it. It offers a clear way to get a project onto the list and off it, which is a service to the people who are tired of being handed work sideways. It offers help to the people delivering: a planning session, a risk conversation, a hand with a difficult stakeholder.

Control arrives later, as a consequence. Once the list exists, the question of whether it is longer than the capacity to deliver it answers itself, and the leadership team starts asking for the decision forum it never had. Once projects enter through a known door, prioritization becomes possible. Once the office has helped a few project managers out of trouble, a request for a status update is answered honestly, because the answer might bring help.

We call this services before control. It is not a softer version of the same office. It is a different sequence, and the sequence is the whole point. The idea is far older than project management. Robert K. Greenleaf, who coined the term servant leadership in 1970, offered a test for it that fits a PMO well: “Do those served grow as persons?” For an office, the equivalent question is whether the people it serves deliver better because it exists.

What it looks like at different levels

The right size of the office depends on where the organization is, which is why we measure before we design.

At the first and second maturity levels, where outcomes still depend on individuals and the shared way is written but unevenly applied, the office is often one person. That person keeps the register, runs a monthly portfolio review with the leadership team, and spends the rest of the time helping. There is no software beyond a spreadsheet, and the status report is the register itself, read aloud.

At the third level, where practice is consistent and reviewed on a cadence, the office adds a capacity view the leadership team can read in a minute, two or three measures that matter, and intake criteria the leadership will actually apply. A tool may now be worth buying, because there is a practice for it to support.

Reaching the fourth level, where the practice adjusts itself on evidence, is not everyone’s goal, and for many organizations the third level is the right place to stop.

Three questions before setting one up

Before an office is chartered, three questions are worth an honest hour with the leadership team.

Is the list of projects shorter than the capacity to deliver it? If nobody knows, that is the office’s first job, and it is a service, not a control.

Who decides to stop a project? If the answer is nobody, or everybody, the office will produce reports that lead nowhere until that forum exists. Designing the forum is part of designing the office.

What will the office give on its first day? If the only answer is a template, the pattern above has already started.

What we do not do

We do not set up an office by issuing a tool. We do not write a two-hundred-page manual for it. We do not treat its first year as a compliance exercise. We design the smallest office that gives something on its first day, run it on real projects for two or three cycles, adjust what turns out wrong, and leave when the organization’s own people are running it.

The Tao Te Ching, in James Legge’s 1891 translation, describes the best of the ancient rulers in a line we would gladly apply to our own work: “Their work was done and their undertakings were successful, while the people all said, ‘We are as we are, of ourselves!’”

If your organization has a PMO that is being worked around, or is about to set one up, the first step is the same: a short look at the list, the capacity and the decision forum, honestly scored. That is the Map stage of how we work, and it is useful on its own.

Sources

• Monique Aubry, Brian Hobbs, Ralf Müller and Tomas Blomquist, Identifying Forces Driving PMO Changes, Project Management Journal 41(4), 2010.
• Bredillet, Tywoniak and Tootoonchy, “Exploring the dynamics of project management office and portfolio management co-evolution: A routine lens,” International Journal of Project Management 36(1), 2018, summarized by the Association for Project Management.
• Gartner, Gartner Identifies Seven Best Practices for an Effective Project Management Office, press release, April 2016, quoting analyst Mbula Schoen.
• Kevin Rose and Woojin Choi, 2024 Strategic Roadmap for the EPMO, Gartner, January 2024.
• Shivica Mathur, Enhancing PMO Service Adoption Through Targeted Stakeholder Engagement Strategies, Gartner, April 2026.
• Project Management Institute, Pulse of the Profession 2018: Success in Disruptive Times, 2018.
• Robert K. Greenleaf, The Servant as Leader, 1970, as quoted by the Greenleaf Center for Servant Leadership.
• Laozi, The Tao Teh King, chapter 17, translated by James Legge, 1891, Project Gutenberg.