What is portfolio governance?

Ask three people in the same strategy office what portfolio governance is and you will get an answer about committees, an answer about dashboards, and an answer about the budget cycle. All three describe parts of it. None is the thing.

What is portfolio governance?

Portfolio governance is the set of decision rights, records and review routines by which an organisation decides which work to fund, holds that work against the objectives it was funded to move, and confirms afterwards that the promised outcome arrived. Its unit of interest is the mandate and the money, not the task list. Project management asks whether a piece of work was delivered on schedule; portfolio governance asks whether it was the right work to fund at all, and whether it produced what it promised.

Put in terms of a room, it is how a strategy office answers three questions in front of a board. What did we commit to. What is actually happening. Did it work. You have portfolio governance when those three answers come out of the same record, and you do not have it when they come out of three documents assembled by three different people in the week before the meeting.

Most entities that believe they lack governance have plenty of governance activity. What is missing is a record that all of it points at.

How is portfolio governance different from project and programme governance?

Three altitudes, three time horizons, three kinds of authority.

Project governance governs one defined piece of delivery: scope, schedule, cost, quality, risk, for a single output. PRINCE2 describes this layer well. Its question is whether the thing will arrive as agreed, and its authority ends at closure.

Programme governance governs a group of related projects pursuing one shared outcome. It coordinates dependencies, resolves contention for the same scarce team, and asks whether the outcome is assembling across the parts. MSP is the reference most UAE and Saudi entities already have on the shelf.

Portfolio governance governs the whole set of funded investments against strategy, and it does not end. Given finite money and finite delivery capacity, is this still the right mix. What should be started, slowed, merged, stopped. ISO 21504 and MoP address this layer, though most offices assemble their own version out of whatever their mandate requires.

The practical difference is authority over money. Project and programme governance decide how to deliver what is already funded. Portfolio governance decides what gets funded and can take the funding back.

Want a quick read on a forum that calls itself a governance board? Skip the terms of reference and look at what it has stopped in the last two years. A board that cannot cancel an initiative or move its budget is a reporting body. Fine thing to be. Not what it is called.

What does a portfolio governance record contain?

A connected chain, not a set of registers.

An objective carries a statement, an accountable owner, the mandate it derives from (Vision 2030, We the UAE 2031, a sector strategy), its place in a hierarchy, and its weight against its siblings. Skip the weight and everything is equally important, which is operationally identical to nothing being important.

A KPI carries a baseline value with the date it was taken, a target with its date, a unit, a measurement method, a frequency, a named owner, and a series of actuals. A measure with no baseline is an opinion. It shows position but not movement, and governance is almost entirely about movement.

An initiative carries approved, committed and spent cost, a sponsor, a delivery owner, the objectives it serves, the KPIs it claims to move, and its current gate.

A milestone carries a baseline date, a forecast date and an actual date. Three dates, not one. With three dates, slip is a subtraction. With one, slip is a negotiation between the person reporting and the person receiving.

A benefit carries what will change, in which units, by when, measured on which KPI, and who stays accountable once the delivery team has dispersed.

None of that is exotic. What is rare is holding the links between them as data rather than as a slide. Anyone can produce five registers. The value is being able to ask which objectives are exposed if this initiative slips two quarters, and getting the answer from the record instead of a workshop.

One element is routinely left out: the audit line. Every status that moved, every gate passed should carry who changed it, when, and what it was before. That is the difference between a number you can defend and a number you can only repeat.

Why do spreadsheets fail at portfolio governance?

Not because they are unsophisticated. Most of the best portfolio analysis you will ever see was done in a spreadsheet. It fails at governance for structural reasons no skill in the author can fix.

It stores a snapshot, not a history. When a status moves from amber to green, the amber is gone. No record that it was ever amber, who changed it, or on what evidence.

It has no referential integrity. Rename an initiative on one tab and its links elsewhere break quietly, still returning a number, just the wrong one.

Ownership is a text field, not a person. Nothing stops “Finance” being an owner, and nothing notices when the individual who actually understood the number transfers out.

Consolidation is manual, which makes it periodic, which makes the portfolio view at least as old as the retyping cycle. It is also tedious and error-prone.

And it never refuses. It will accept a KPI with no baseline, a benefit with no measure, a milestone with no date. A large part of governance is refusing a partial record at the moment someone tries to enter it, and a blank cell cannot refuse anything.

Where does this definition get uncomfortable?

Everything above argues that governance improves when more of the portfolio is measured and linked. That is mostly true, and not entirely true.

Goodhart’s law is the obvious problem. Once a measure becomes the basis on which funding is granted, people optimise the measure rather than the thing it stands for. Attach a KPI to every objective, publish it, tie budget to it, and in eighteen months you will have a portfolio that scores well. Whether it scores well because outcomes arrived or because definitions were adjusted is a question the record cannot answer.

There is a quieter distortion too. Measurable work crowds out important work. An initiative with a clean baseline and a monthly actual is easy to defend in a committee. An initiative whose value is a non-event, the breach that did not happen, the capability held in reserve, is nearly impossible to defend against it, because its success looks identical to it never having been needed. Strong governance systematically favours the first kind. Nobody has a satisfying general answer. Some offices carve out a protected allocation and decide not to govern it the same way, which is an admission rather than a solution.

What does good portfolio governance look like?

Concrete markers, none from a maturity model. Every KPI has a baseline, because the system will not store one without it. Every number carries a person’s name rather than a department’s. Actuals arrive on their own cadence instead of being requested. Slip is arithmetic. Decisions carry the name and date of whoever made them and stay attached after that person moves on. Benefits outlive the projects that promised them. And the forum spends its time on exceptions: if a governance meeting opens by reading aloud the rows that are green, it has already failed.

Where should an office start?

Not with a tool. Take one objective and trace it downward.

Pick something a minister or a board member would recognise by name. Find every initiative claiming to serve it. Find the KPI supposed to show it moving, and ask when the baseline was taken and by whom. Find the benefits promised by initiatives under it that have already closed, and ask who measured them.

You will almost certainly find the same four things: initiatives attached to no objective, objectives with no measure, KPIs with no baseline, benefits nobody owns. It takes about a week and it beats a maturity assessment, mostly because it is specific enough to be uncomfortable.

Common questions

Is portfolio governance the same as PPM?

No. Project portfolio management describes the operating practice and its tooling: intake, prioritisation, resourcing, scheduling, reporting. Governance is the decision layer above it, covering who may decide what, on what evidence, and where the decision is written down. A capable PPM tool with no governance produces well-formatted reports nobody acts on.

Who should own portfolio governance in a government entity?

Usually the strategy office or the EPMO, and where it sits matters more than what it is called. The function needs enough authority to stop things and enough independence to publish a number the delivery organisation would rather not see.

What is the minimum a governance record must hold?

For every funded initiative: the objective it serves, the KPI it claims to move with a baseline value and date, a named accountable person, baseline and actual dates on its milestones, and the benefit it promised with a measure and a date after closure. With less than that you can report activity, but you cannot judge value.

Where this comes from

We build TruMandate, a portfolio governance platform used by government entities and large enterprises in the UAE and Saudi Arabia. The above is what we learned holding that chain as one connected record instead of five.