Portfolio commitment is the organization agreeing to protect the capacity an approval assigned, for the length of the stability window. It is the half of the decision most intake processes omit, and without it an approval reserves nothing. Commitment is established in the Align practice at the same moment as the approval it protects, and unlike the approval, it is tested continuously rather than at a point in time.
How does portfolio commitment work?
Commitment turns an assignment into a protection by adding three rules that apply after the decision is made, for as long as the stability window runs.
The first is the protection rule: the people named in an approval are not available for other work during the window. That is the entire substance of the commitment, and it is a constraint on the people who normally reallocate capacity informally, which is to say most of the leadership team.
The second is the displacement rule: anything that does enter mid window removes an equivalent amount of committed capacity, by name, and the removal is decided by whoever holds the relevant decision right. Work is exchanged rather than stacked. This is what stops a commitment from being honored on paper while the same teams quietly absorb three additional requests.
The third is the visibility rule: every breach is recorded. Not punished, recorded. The count of mid-window interruptions and what they displaced is the only honest measure of whether the organization commits or merely approves.
Commitment is therefore not a mood or a level of enthusiasm. It is a small set of rules with a person enforcing each one, and an organization either operates them or it does not.
How do you tell whether a commitment is real?
Watch what happens to the next urgent request. If it displaces committed work without going through the reopening conditions, there was an approval but no commitment. That is the whole test, and it can be run in any organization without instrumenting anything.
Four secondary signals are worth watching between tests:
- Named people appearing on unapproved work. Check the calendars of the two or three scarce specialists in your portfolio. If they are spending meaningful time on initiatives that hold no approval, the protection rule is not operating.
- Committed initiatives slipping without a renegotiation. A committed piece of work that is late and has never been formally rescoped usually lost capacity nobody recorded losing.
- Teams keeping a private buffer. When delivery leads routinely under-commit because they expect interruptions, they are pricing in the breach rate. That is rational behavior and a reliable diagnostic.
- A growing gap between the portfolio view and what people are doing. If the board and the week do not match, the board is describing approvals and the week is describing commitments.
What does good portfolio commitment look like?
Good commitment is specific enough to check. Every approval has a stated window, because an approval with no duration cannot be protected. Scarce roles are protected by name: protection of "two analysts" is unenforceable, while protection of two named analysts can be checked. Interruptions have one legitimate route, and delivery leads know explicitly what protection they now have. A commitment nobody communicated changes no behavior, because teams keep accommodating requests they were never told they could refuse.
The clearest sign it is working is the first refusal. The organization learns what the commitment means from the first request that is queued instead of absorbed, and it learns the opposite from the first one that is absorbed quietly. Over time the recorded breach count falls, and the trend is what persuades people.
What does a commitment look like under pressure?
Say a retail technology group approves four initiatives for a 10 week window and commits three teams to them, including the only two people who know the order management integration. In week three, a senior vice president asks one of those two for "a couple of days" on a supplier onboarding issue that was never ranked.
Without commitment, that request is granted in a hallway. Two days becomes three weeks of intermittent work, the committed initiative slips, and the slip is reported as a delivery problem at the next steering meeting, where nobody connects it to the hallway.
With commitment, the request goes to the person holding the reopening right. It does not meet any listed condition, so it is offered two routes: wait for the decision point in seven weeks, or displace something now. The senior vice president looks at what would be displaced, decides the onboarding issue is not worth pausing the order management work, and the request is queued. The whole exchange takes ten minutes, and it is recorded.
The interesting part is what happens next. Two weeks later the same request comes back, this time with a customer contract attached, which does meet a listed condition. It is taken, the lowest-ranked committed initiative is paused, its team is released, and the pause is recorded with its reason. The commitment held in both cases, because holding a commitment is not the same as refusing everything.
What are the common mistakes?
Approving and stopping. A frequent failure. Portfolio approval is treated as the end of the decision, and the protection is left to good intentions.
Committing capacity the organization does not have. Commitment made against optimistic capacity collapses in the first week, and the collapse discredits the mechanism rather than the estimate. Commit against validated capacity, within your WIP limits.
Letting small interruptions bypass the rule. The protection rule is broken by half days, not by projects. Set a threshold low enough to catch them, and route anything above it.
No record of breaches. Without a count, every conversation about whether commitments hold becomes anecdotal, and the anecdotes favor whoever has the better memory for their own interruptions.
Commitment that only binds the teams. If delivery leads are held to their dates while leadership reallocates their people, the arrangement is not a commitment. It is an unfunded expectation, and it is the fastest way to lose the credibility of the whole operating system.
Treating a broken commitment as a personal failure. The useful response to a breach is to ask which rule did not operate and who was supposed to operate it. The answer is usually a missing route rather than a missing will.
Portfolio commitment vs portfolio approval
Approval and commitment are two halves of one decision, made at the same moment, and they fail differently. Approval is a permission granted at a point in time. Commitment is a protection maintained over a period. An organization can be excellent at the first and absent at the second, which is common, and it produces portfolios where everything is approved and nothing is protected.
| Portfolio approval | Portfolio commitment | |
|---|---|---|
| What it does | Permits an initiative to consume capacity | Protects that capacity for the window |
| When it happens | At the decision point | Continuously, until the next decision point |
| Held by | The approver at the relevant threshold | Everyone who could reallocate the named people |
| Failure looks like | An approved list nobody staffed | Staffed work quietly interrupted |
| Evidence it worked | A written approval record | A low, recorded count of mid-window breaches |
The dependency runs in one direction. Commitment without approval is just people being busy on something nobody chose. Approval without commitment is a document. You need the approval to say what capacity is assigned and to whom, and the commitment to make that assignment survive contact with the next urgent request.
For how these terms sit alongside the rest, see the glossary of agile portfolio management terms. If your approvals are already sound and the capacity keeps leaking anyway, that is a good pattern to bring to a 60-minute Portfolio Clarity Session, and a PPM4Agile implementation puts the protection, displacement, and visibility rules in place.