STRATEGY

Prioritisation scoring of a backlog

Use when you have more ideas than capacity and the arguments about order have gone in circles.

prioritisation-scoring.md

Download .md

You are a prioritisation facilitator. You score consistently and you show your reasoning per score.

Backlog items: {{BACKLOG_ITEMS}}
What we are optimizing for this quarter: {{PRIMARY_OBJECTIVE}}
Capacity available: {{CAPACITY}}
Known dependencies or blockers: {{DEPENDENCIES}}

Method:
1. Score each item 1 to 5 on: Impact on the primary objective, Confidence in that impact, Speed to first result, and Strategic fit. Score Effort separately in person days, not on a 1 to 5 scale.
2. Compute a score of (Impact x Confidence x Fit) divided by Effort in days. Show the numbers.
3. Output a ranked table with columns: Rank | Item | Impact | Confidence | Speed | Fit | Effort days | Score | One line reason for the Impact score.
4. Draw a cut line where cumulative effort reaches the stated capacity. Everything below the line is explicitly not being done this quarter.
5. List any item whose Confidence score is 2 or below and propose the cheapest test that would raise it.
6. Flag any item that is blocked by a dependency and cannot start regardless of rank.

Constraints: do not adjust scores to make a favored item rank higher. If two items score within 10 percent, say the ranking between them is not meaningful. No em dashes.

Fill in before running

Replace each placeholder with your own detail. The more specific you are, the less the model invents.

{{BACKLOG_ITEMS}}
{{PRIMARY_OBJECTIVE}}
{{CAPACITY}}
{{DEPENDENCIES}}

Getting a better result

Score effort in days rather than t-shirt sizes; the division only works with a real number.
Run it twice, once with a different primary objective, to see which items are robust to the goal changing.
Share the cut line rather than the top of the list - the argument is always about what got dropped.

Questions about this prompt

When should I use this rather than settling it in a planning meeting?

When the same argument has gone round twice. The value is not the score itself, it is that every item is judged on the same four dimensions against one stated objective, with a one line reason attached to each impact score. Disagreement then happens about the reason rather than about the order.

What do I need to score the backlog properly?

Backlog items described well enough to score, effort in person days rather than t-shirt sizes, capacity in that same unit, and known dependencies. The score divides by effort, so a size label breaks the arithmetic. Decide what you are optimising for this quarter before you paste anything, because every impact score hangs off it.

What comes back, and which part should I share?

A ranked table with the four scores, effort in days, the computed score and the reason behind each impact score, a cut line at your capacity, low confidence items paired with the cheapest test, and blocked items flagged. Share the cut line rather than the top of the list, since the argument is always about what got dropped.

What is the mistake that wastes the scoring exercise?

Treating a narrow gap as a real ranking. Items within 10 percent of each other are called out as not meaningfully ordered, and reshuffling those is wasted meeting time. Also resist nudging a favoured item upwards. If you are going to override the model, say so openly rather than editing the inputs.