Read your delivery plan again and ask which of its commitments were made by people who report to the project. It will be a short list.
A commercial shipyard approves a vessel programme. The plan carries eight signatures — the yard director, the design office, procurement, the pipe shop, electrical, paint, quality and finance.
Eleven months later the pipe shop is running at sixty per cent of assumed availability and the programme slips a quarter. In the review nobody is found to have broken a promise, and the reason is uncomfortable: nobody in the pipe shop believes they made one. A manager signed a document. What they understood themselves to be signing was that they had seen it.
The programme was never the delivery team's plan. It was a bundle of undertakings by eight parties, six of whom the team could not direct, collected in a way that guaranteed most were not undertakings at all.
The Strategic Context
The teaching material behind this article is unusually precise about what a signature on a plan means. Each stakeholder signs "attesting that as far as his or her area of concern goes, the plan seems to be feasible and appears to address those concerns properly." And then it gives the example that matters: a functional manager signs "signifying that he or she feels confident that the number of support people required to get the job done will be available when required, and that they will be able to handle the technical requirements of the job."
Read literally, that is a capacity pledge: not an opinion, not a courtesy, not an acknowledgement of receipt, but a forward commitment of people by someone who owns them and does not report to the initiative that needs them.
That is what a plan mostly consists of. The delivery team's own contribution is real but small — sequencing, coordination, the management of the work. Almost everything else belongs to someone else and has been promised.
Enterprises govern this asset badly, because it does not look like an asset. It looks like a document.
What Leaders Commonly Misread
The first misreading is that a plan is the delivery team's commitment. It is a register of other people's commitments with a delivery team's coordination wrapped around it. When it fails, the enterprise investigates the wrapper.
The second misreading is that a signature is a commitment. The source is blunt about this, and its warning is procedural rather than moral. Sign-off "should be done in a meeting. It should not be done by circulating copies of the notebook through company mail and asking people to read and sign it. Under those conditions, they will often fail to read carefully, and when they are later required to 'deliver' on their part of the project, they may be unable to do so."
That last clause is the whole mechanism. The failure is not bad faith. It is that a person who signs without reading has not formed the belief that the signature asserts, and cannot be surprised later to discover they cannot honour it — because they never checked.
The third misreading is that approval and commitment are the same event. They occupy the same meeting and they are different objects. Approval is a decision to proceed, and whether that decision is ever genuinely capable of going the other way is a separate governance question examined in [Related article: What a Stage Gate Is Actually For]. Commitment is a set of individual undertakings that either exist or do not. An initiative can be approved with none of them in place, and frequently is.
The fourth misreading is that the signatory can pledge what they are signing for. A functional head attesting that people will be available may hold that capacity only nominally — already promised elsewhere, or contested by a structure designed to contest it. The arithmetic of that contest, and why fractional allocation does not add up, is argued in [Related article: There Is No Such Thing as Half a Project Manager]. This article takes it as given and asks a narrower question: when the pledge is made, does anybody test whether the person making it can keep it?
Reframing the Issue
The reframing is to treat plan sign-off as the collection of counterparty commitments, and to design it as an enterprise would design any other process for collecting commitments from parties it cannot compel.
Three consequences follow immediately.
The method of collection determines whether the commitment exists. Every organisation understands this commercially — nobody closes a supply agreement by emailing a document and asking for initials. Internally the same enterprise does exactly that, and treats the result as binding.
A refusal is information, not obstruction. A functional head who declines to attest that people will be available has told you something the plan needs and would otherwise have concealed until the eleventh month. Organisations that punish that answer stop receiving it, and are left with a plan whose signatures are uniformly present and uniformly uninformative. The source is explicit that the purpose "is to ensure that it is realistic and feasible, not to use it as a club to 'beat up' people if there are problems!"
The uncollectable pledge has to be priced. Where a commitment genuinely cannot be given — because the capacity is contested, or the decision belongs elsewhere, or the answer is not yet knowable — that is a live exposure with an owner and a date, and it belongs in the plan as one.
A boundary is worth naming here, because it is easy to blur. A separate failure occurs when a plan never states what it will not produce, and every reader fills the silence with an expectation — the argument in [Related article: Silence Reads as a Promise]. That concerns an unstated negative read as a commitment. This concerns a stated positive that was never a commitment at all. Both are failures of what a document conveys; they are opposite in shape and they need different remedies.
Strategic Analysis
What the room is for
The source's insistence on a meeting is not ceremony, and it is worth understanding why the format changes the output.
A circulated document is read privately, by someone with other work, where the cost of signing is zero and the cost of not signing is a conversation. A meeting inverts both: the commitment is made in front of the people whose plans depend on it, raising the cost of an unconsidered yes, and the objection is voiced where it can be resolved, lowering the cost of a considered no.
The material goes further, and asks the delivery leader to do something counterintuitive: to "highlight key areas of the project that could be problems and encourage all members of the group to ask 'embarrassing' questions", on the reasoning that "better that those questions be asked during the meeting than later."
That is an instruction to make one's own plan look worse in the room where it is approved. Very few delivery leaders are incentivised to do it, which is precisely why it has to be the sponsoring organisation's expectation rather than the delivery leader's choice. Note what that expectation is not: it is not a duty of the sponsor's office, whose obligations are a separate matter set out in [Related article: Sponsorship Is an Office, Not an Endorsement]. It is a property of how this particular meeting is run.
The scope-of-concern problem
"As far as his or her area of concern goes" is a sensible limit and it creates a specific gap.
Eight signatories attesting within their own scope produce eight local feasibility judgements and no global one. The interfaces — the moment the pipe shop's output becomes electrical's input — sit inside nobody's area of concern. Every party can be individually confident and the plan still infeasible, with no signature on the page that would have caught it.
The remedy is not a ninth signature. It is to name, before the meeting, the three or four handovers on which the programme actually turns, and to require that both parties to each handover attest to it jointly. Two people agreeing a date in front of each other is a different artefact from two people separately agreeing their own halves of it.
The same failure in an operations setting
An insurer replaces its underwriting platform. The plan assumes that experienced underwriters will spend one day a week for eight months on rules validation — a substantial commitment, and one the underwriting director signs for.
What that director controls is a team with an annual renewal cycle and a hard external deadline. For seven of the eight months the pledge is honourable; in the eighth it is not, and the eighth is when validation matters. Nothing in the sign-off asked when the capacity would be available against the initiative's need — only that it would be — and the plan's schedule was never put in front of the person whose calendar would break it.
A pledge without a shape is not a pledge. Availability has a profile, and the profile is where the pledge either meets the plan or misses it entirely.
Decision Framework
Five steps, applicable to any initiative before its plan is baselined.
1. List the commitments the plan requires, by party. Not the tasks — the undertakings. Underwriting provides two FTE from March to October. Most plans have six to fifteen. Writing them out separately from the schedule is the whole diagnosis, and takes about an hour.
2. Establish who can actually give each one. The person who owns the capacity, not the person who owns the relationship. Where those are different people, the pledge is being collected from someone who cannot make it.
3. Collect them in a room, with a profile attached. Each party states what they are committing, over what period, and at what intensity. The delivery leader's job in that meeting is to surface what could go wrong, not to secure agreement — and the sponsoring organisation should say so explicitly beforehand, because otherwise nobody will do it.
4. Record refusals and qualifications as findings. A pledge that cannot be given goes on the plan as an exposure with an owner and a review date. This is the step organisations skip, and skipping it converts good information into a surprise eleven months later.
5. Re-collect at each material replan. Commitments are made against a shape. When the shape changes the commitment has lapsed, whether or not anybody says so.
A supporting convention: name the three handovers the programme turns on, and have both sides attest jointly. It costs ten minutes in the meeting and it is the only mechanism above that addresses the space between scopes of concern.
From Strategy to Execution
Immediate. Take the initiative closest to baseline and complete step one — the list of undertakings, by party. Then ask each party's owner a single question: what did you understand yourself to be committing to? The gap between the two lists is available in an afternoon and is usually larger than anyone expects.
Medium term. Change how sign-off is run. One meeting, commitments stated aloud with their profiles, refusals recorded rather than resolved on the spot. This is a change to a calendar invitation and a template, and it converts a document-collection exercise into a commitment-collection one.
Long term. Build the enterprise's memory of which commitments hold. Most organisations know informally which functions honour a pledge and use it nowhere. A record of pledged-versus-delivered capacity, kept across initiatives, turns reputation into a planning input — and shows where the problem is a function that over-promises rather than a team that under-performs.
Signals to Monitor
- Sign-off obtained by circulation. The single strongest predictor of a pledge that will not hold. Count how many of your last five plans were approved without a meeting.
- Uniform agreement. Where every signatory signs without qualification on every plan, the process is collecting signatures. Some qualification is evidence that people are reading.
- Commitments with no profile. A pledge expressed as a headcount but not as a period and an intensity has not been tested against the schedule it is meant to serve.
- Failures attributed to the delivery team where the shortfall was somebody else's capacity. Look back at the last two post-mortems and ask whose commitment actually failed.
- Signatures obtained against superseded plans. Where a baseline has moved and nobody was re-asked, the plan carries consent to something that no longer exists.
- Functions that never say no. Not good citizenship. Usually a sign that saying no has a cost — a different mechanism from the reporting losses examined in [Related article: Whose Knowledge Does Your Governance System Actually Hear?], and one that operates before anything is reported at all.
Questions for the Leadership Team
- On our largest initiative, how many of the commitments in the plan were made by people who report to it?
- When we last approved a plan, was it approved in a room — and did anyone qualify what they were agreeing to?
- Which functions have we asked for capacity three times in the last two years, and how often did the capacity arrive in the shape promised?
- What does the person who owns our scarcest specialist capability believe they have committed to this year, in total?
- When an initiative slipped last year because people were not available, who did we hold accountable — and was it the party who made the pledge?
- Where a function has declined to commit, do we record that as a finding, or do we treat it as something to be managed around?
Closing Perspective
A delivery plan is a strange document. It is written by the party with the least authority over the resources it describes, approved by parties who read it least carefully, and enforced against the only party who cannot supply most of what it needs.
The correction is not a better template, a stricter governance forum or a more detailed schedule. It is to recognise what the signature page actually is — a set of forward commitments from parties the enterprise cannot compel — and to collect it accordingly: out loud, with a profile, in a room where the objection is cheaper than the failure.
Organisations that do this look, from the outside, slower to approve. What they are doing is finding out in the second month what everyone else finds out in the eleventh, and one of the conditions of finding out that early is that the schedule they are committing to has itself been examined rather than inherited, which is the argument in [Related article: Which of Your Dependencies Are Real?]. A plan built on untested orderings and uncollected promises is not a plan. It is a hope with dates on it, and the dates are the most convincing part.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.