MODULE 3 · LESSON 2
Free — no login requiredSign in to track progress, save quiz attempts and enrol in the full course.
Sign in to track progress / enrolScoring and the Portfolio View
Candidates that survive qualification are all possible. Scoring is about which to do, and in what order.
Four dimensions
Score each surviving candidate 1–5 on four axes. Keep the scale coarse: precision here is false, and a 1–10 scale invites arguments about whether something is a 6 or a 7 that carry no information.
Value. How much does the winning value pattern from module 1 actually deliver? Score against a stated number, not a feeling. A 5 has a quantified benefit somebody outside the project agrees with.
Feasibility. How likely is it to work technically? The hundred-case sample gives you this almost directly.
Readiness. How much has to be true before you can start? Is the data reachable, is there an owner, does the integration point exist? This is the axis that most often makes an attractive candidate a bad first candidate.
Risk. What is the exposure if it goes wrong — to customers, to compliance, to reputation? Score this inverted, so 5 means low risk. Otherwise your total silently rewards danger.
Scoring is a conversation, not a calculation
The single most important thing about this method: the number is not the decision. The purpose is to make disagreement explicit and locate it.
When two people score the same candidate 4 and 2 on feasibility, that gap is the useful output. One of them knows something the other does not — usually about data quality, an integration constraint, or a previous attempt. Scoring separately and then comparing surfaces that in minutes; discussing without scoring buries it under whoever is most senior or most fluent.
So: score independently, compare, and spend the meeting on the disagreements rather than the totals.
Two consequences follow. Never let the highest total automatically win — a candidate scoring 16 with a readiness of 1 is not startable, whatever its total. And never let scoring substitute for the sequencing judgement in the next lesson.
The portfolio view
Do not run one project. Run a small portfolio with different jobs, because a set of projects chosen purely by score will cluster in exactly the wrong place.
The capability project is the one that gets cut, and cutting it is why organisations run five years of pilots without accumulating anything. If invoice extraction, contract review and claims triage are all blocked by the same document store nobody can query, fixing that store is worth more than any single one of them — and it will never win on a score sheet, because its value is entirely indirect.
Name it as a capability project explicitly, so it is not judged against a return it was never meant to produce.
How scoring gets gamed
Three patterns to watch for, all of which mean the numbers have stopped carrying information:
Value inflation. Benefits assume 100% adoption, zero exception handling and immediate headcount reduction. The correction is module 4's discipline: benefit numbers must come with the assumption that produces them.
Feasibility optimism from whoever wants the project. The sponsor scores feasibility, and it is a 5. Feasibility should be scored by whoever will build it, or by the hundred-case sample, which does not have a preference.
Risk minimisation by omission. Risk is scored on what goes wrong when the system is right, ignoring what happens when it is confidently wrong at scale. Score the failure mode, not the happy path.
A logistics business scored six qualified candidates. Value, feasibility, readiness, risk — each 1–5, risk inverted.
| Candidate | Value | Feas. | Ready | Risk | Total | |---|---|---|---|---|---| | Delivery-note extraction | 3 | 5 | 4 | 5 | 17 | | Customer email triage | 2 | 4 | 5 | 4 | 15 | | Dynamic route optimisation | 5 | 2 | 2 | 3 | 12 | | Damage claims assessment | 4 | 3 | 2 | 2 | 11 | | Demand forecasting rebuild | 4 | 3 | 3 | 4 | 14 | | Driver-app query assistant | 2 | 4 | 3 | 4 | 13 |
The naive reading picks delivery-note extraction and moves on.
The useful part happened in the comparison. Two people had scored damage claims assessment at readiness 4 and 1. The gap turned out to be decisive: the operations lead assumed claim photographs were stored against the claim record; the systems manager knew they were attached to emails in a shared mailbox and had never been linked to claims at all. Every historical claim was therefore missing its evidence.
That single disagreement, surfaced in four minutes, saved a project that would have spent three months discovering the same thing. It also generated the capability project: link the photograph store to claim records — which unlocked claims assessment and a fraud-review idea that had been rejected at gate 3 the previous quarter.
The portfolio they ran was: delivery-note extraction as the proving project, demand forecasting as the value project, and the photograph-linking work as the capability project. Route optimisation — the highest value score on the sheet — was deliberately not started, on readiness grounds, with a disqualification note explaining exactly what would change the answer.
Two managers independently score the same candidate 4 and 1 on readiness. What is the most valuable use of that gap?
The four scoring axes
Click to flipValue, feasibility, readiness and risk — each 1–5, with risk inverted so that 5 means low risk. Keep the scale coarse; precision here is false.
Click to flip backScore surviving candidates on value, feasibility, readiness and inverted risk — but treat the total as a conversation opener rather than a verdict, because the disagreements between scorers are where the real information sits. Run a small portfolio with distinct jobs: one project to prove you can ship, one to justify the programme, and one capability project to remove a precondition blocking several others. That last one always looks unjustifiable on a score sheet, and cutting it is how organisations accumulate pilots instead of capability.