Two decisions, not twenty
Shrink the decision until it is easy to make. That is work for the team, not a favour the sponsor owes them.
A mid-sized leadership development company is deciding whether to build its own platform. The interesting part of the planning session was not the platform. It was a sentence about the decision itself: across the whole programme, the two owners have to decide exactly twice, once to fund the preparation and once to fund the build. Everything between those two points is the team's job to break down until each step is small enough that nobody has to be brave.
The decision is the bottleneck, not the build
In most digitalisation projects I have worked on, the thing that stalls is not engineering capacity. It is a decision sitting with someone who has eleven other things on their desk and no confident way to compare this one against them. The project then waits, and the waiting is invisible, because nobody logs a decision that has not been asked for yet.
The counter-move in this programme was to fix the number of decisions in advance and design everything else around it. Two sponsor decisions for the whole thing. To make each of them payable, three artefacts have to exist first: a cost case with real numbers in it, a plan that names scope and dependencies, and a clickable prototype that shows what is actually being bought. That is not documentation for its own sake. It is the work required to turn a judgement call into a comparison.
The sponsor added the other half of it, which matters more than it sounds. In case of doubt, quality over timing, and the team is expected to push back when he is the one holding it up. A sponsor who says that out loud converts his own delay from an act of authority into a defect the team is allowed to report.
Requirements have a shelf life
The argument for the pace came from someone who had seen the alternative. If the tempo is not held, the project stretches over years, and the requirements written in this year will be obsolete by release. That is the real cost of a slow decision, and it does not show up in any budget line.
It is worth saying plainly because the usual framing is wrong. Deciding late is normally treated as cautious, and deciding quickly as risky. For anything built on top of a moving technology stack, the arithmetic runs the other way. A specification ages from the day it is written. Every month of deliberation spends some of its remaining value, and the spend is silent.
A plan is not a fixed asset. It loses value the whole time you are thinking about it.
Every input gets an owner and two numbers
The same session handled a second problem with the same logic. Partners and external contributors kept sending good ideas into the programme, which is a nice problem until you notice that none of them had a defined route. Under the new process there are two: either a team asks for input on something specific, and one named owner in that team takes the answer, or an unsolicited idea goes to a single named contact who checks it against the roadmap and the available capacity. It then either gets done or goes on a parked list.
Two details make that more than a flowchart. Every idea is judged against the same two numbers the whole programme is judged against, in this case revenue growth and the digital share of revenue, which stops the debate becoming a contest of enthusiasms. And the answer goes back to the sender quickly and personally, for a no as much as for a yes, rather than in a forum weeks later. A fast no protects the relationship better than a slow maybe, and it keeps the list honest.
The process itself is being introduced live in a meeting rather than announced by email. That is not ceremony. A routing rule that arrives as an attachment gets read once and ignored; one that arrives as a conversation gets questions, and the questions are how you find out whether it survives contact.
How to apply it
- Count the decisions your sponsor has to make in the next quarter. If the number is above three, the project is designed to stall.
- For each remaining decision, write down what has to exist for it to be a comparison rather than a leap of faith, and treat producing those artefacts as the actual work.
- Say the shelf life out loud. Name the date on which this year's requirements stop being worth building, and use it as the cost of delay.
- Give every unsolicited input one named recipient and one visible parked list. An idea with no route becomes a grievance.
- Measure everything, including partner ideas, against the two numbers the programme is already accountable for.
- Ask the sponsor to name himself as a possible bottleneck in front of the team. Without that sentence, nobody will ever report it.
You cannot make a sponsor decide faster. You can make the decision smaller, and that is the team's job, not a favour.

