Chapter 5: Score Opportunities Without Pretending the Score Is Truth

Once the map has surfaced several opportunities, the scorecard helps you put them beside each other and explain your judgment. It makes assumptions easier to discuss. A total cannot establish that those assumptions are true or choose on the owner's behalf.

Describe each candidate and the first step you're comparing, then give it a whole-number score from zero to five on each of these dimensions:

1. Business impact — if it works, how much does the outcome matter?
2. Likelihood of success — given the process, data, technology, and constraints, how plausible is improvement?
3. Adoption friction — how likely are the people involved to use the changed capability? Five means low friction and easy adoption.
4. Time to evidence — how quickly can the company learn something decision-worthy? Five means evidence arrives quickly.
5. Cost and risk — how favorable is the experiment's cost and risk? Five means low cost and contained risk.

Beside every rating, write the reason, the evidence you have, and what remains unknown. An estimate can be useful if everyone understands what it is. Missing information should stay visible; it isn't a reason to invent a score or silently enter zero. You can prepare or measure before completing the comparison.

Keep the possible outcome separate from the scope of the first learning step. A large potential improvement may be worth investigating through a small record review, but the low cost of that review tells you little about the cost of a full rollout. If you later consider implementation, scope and compare that commitment afresh.

After the ratings are explained, apply the gates and bands below. Use them to focus the conversation on what you should learn or do next, with the owner making the decision.