Checklist · 17 August 2026
Eleven questions to answer before you sign a large build.
Every one of these is cheap to establish before signature and expensive to discover afterwards. Use them with any supplier, including us.
Most programmes are not lost during delivery. They are lost in the four weeks before anyone starts, when a set of reasonable-sounding assumptions get written into a contract without being tested.
None of what follows requires a consultant. It requires someone willing to ask uncomfortable questions while it is still cheap to hear the answers.
Who is actually accountable
Who owns the outcome, not the deliverable? Name one person. If two names come up, or the answer is a committee, you have found your first structural problem before spending anything.
Can that person kill it? Someone who can approve a programme but not stop it is a sponsor in title only, and the programme will outlive its own usefulness.
Will they still be in post in twelve months? Inherited programmes are rarely killed and rarely championed. They are continued at reduced conviction, which is the most expensive of the options available.
Whether the organisation can absorb it
Who has to change how they work? Name the teams. Then ask what those teams are being asked to stop doing to make room. If the answer is nothing, the change has nowhere to go.
What happened last time? If a comparable change was attempted here before and did not survive, the reason is almost certainly still present and has not been named.
Who runs it after handover? If that person is not identified before the build starts, you are buying something that will be operated by whoever happens to be available, which is how a new system quietly becomes a parallel process.
Whether you are buying the right thing
What is the evidence that this is the constraint? Complete the sentence: performance is limited by ___, and the evidence is ___. If the second half will not complete, you have a hypothesis and are about to fund it as a certainty.
Who decided the brief, and what would change their mind? A brief nobody can argue against is usually a brief nobody has examined.
Does the supplier have an interest in this answer? Not an accusation but a structural fact. A firm that builds platforms will diagnose a platform problem. Ask what they would do if the constraint turned out to be elsewhere.
Whether you can stop
How is success measured, and was it agreed before the work started? Measures written after delivery are written to be met. Agree them in advance or expect the argument later to be about opinion.
What does it cost to stop at each stage? Ask for the number at three points, not one. A programme you cannot exit is not a programme, it is an annuity, and you should know that before signing, not during the first difficult quarter.
What to do with the answers
If most of them come back clean, sign with more confidence than you had an hour ago. If three or more will not answer, the constraint is not in the thing you were about to buy. It is in the conditions around it, and buying will not touch it.
Questions four, five and six are the receptivity gate in compressed form. If they fail, the useful next step is preparation rather than procurement. The full gate and constraint map are published if you want to run the longer version yourself.
If you would rather have it examined properly
That is what the Assessment is for: three weeks, one fixed fee, and a recommendation you own, including the recommendation not to proceed.