When the Priority Score Hides the Decision

Two product initiatives sit near the top of a spreadsheet. One improves an existing workflow. The other brings a new customer segment into the product. Both have reach, impact, confidence and effort values. Their scores look comparable, even though one counts active accounts and the other counts sales leads.

The calculation can be correct while the comparison is wrong. Before using a score to assign scarce capacity, establish what its inputs mean, where they came from and which changes would reverse the ranking. The output should help explain the decision in ordinary language.

The examples here concern a fictional team-collaboration product. All numbers are synthetic and illustrate the method. They are not customer results, research findings or suggested industry benchmarks.

Start with a common unit of comparison

Sean McBride's original RICE article combines reach, impact, confidence and effort, using a defined reach period and total team effort in person-months. It also permits dependencies or essential requirements to justify a lower-scoring item being done first. The method is a planning aid, not a mandatory execution order. Intercom's original RICE explanation

For this example, the team is comparing ways to improve completion of a shared work handover. It chooses eligible active accounts reached over the next quarter as its reach unit. A sales forecast must therefore be translated into an evidence-backed estimate for that same population and period before being compared. Counting every lead as a reached account would inflate one option before the multiplication begins.

Agree how repeat use is handled. An account completing ten handovers must not count as ten accounts in one proposal and one account in another. If the team chooses events instead, both options need the same event definition. Put the unit next to the value in the worksheet.

Keep the scoring scale separate from a predicted outcome

Impact points express a judgment in this example. A score of two does not mean the feature doubles retention or produces two units of revenue. Confidence is also a judgment about the supporting evidence, not a measured chance of delivery or success.

The product lead should be able to show the observation behind each input. A support case may establish that a failure exists without establishing how many accounts experience it. A prototype session may suggest a useful interaction without measuring adoption. Keep those evidence limits beside the relevant estimate.

Use the product strategy guide to settle the customer problem before scoring competing solutions. Comparing a retention initiative with a different strategic objective may require an explicit leadership tradeoff that this worksheet cannot resolve.

Work through a ranking that can reverse

The fictional team compares guided handover with a reusable template feature. Both use the same quarter, account definition and impact scale. The numbers below are proposed planning inputs only.

Guided handover scores 600 × 2 × 0.8 ÷ 4 = 240. Templates score 900 × 1 × 0.5 ÷ 2 = 225. These are weighted comparison scores under the selected inputs. They are not expected revenue, a count of successful customers or a return-on-investment measure.

The difference is small enough to inspect the assumptions before treating the order as settled. Holding the handover numerator fixed at 960, its effort threshold for tying the template score is 960 ÷ 225, approximately 4.27 person-months. At 5 person-months, its score becomes 192 and the ranking reverses.

This threshold is arithmetic, not a claim that effort can be forecast to two decimal places. It tells the team where to investigate: an omitted support or migration task could change the choice. The estimate should remain a rough planning input with its evidence and uncertainty visible.

Count the effort needed to make the option usable

Suppose the four-person-month handover estimate includes product, design and engineering but omits customer migration. Adding that work may be warranted. The team should apply the same boundary to templates, including any required rollout or support preparation. Expanding one estimate while leaving the other incomplete creates another misleading comparison.

Person-months describe effort, not elapsed calendar time. A specialist dependency can delay release even when total effort is small. Record timing and capacity constraints beside the score instead of pretending the denominator models every delivery risk.

If both options require the same prerequisite, explain whether it is already funded, included in each independent option or allocated in a combined plan. Do not count it twice in a portfolio total, and do not omit it from both proposals while assuming someone else will pay for it.

The business-case sensitivity guide examines whether an investment itself remains worthwhile. This article asks a narrower question: which of two competing uses of capacity ranks higher under comparable assumptions. The team still needs both kinds of judgment where they affect the decision.

Leave with a decision that survives the spreadsheet

A reasonable outcome might be a short investigation of migration effort before committing the team. Another might be to select templates because the handover estimate is incomplete. If the team chooses guided handover anyway, record the reason and who accepts the uncertainty. A score should not conceal that judgment.

Keep the alternative that loses capacity in the decision record. Otherwise the team can approve both initiatives in separate meetings while still believing that it made a priority choice. Record a revisit trigger tied to new evidence, rather than reopening the debate whenever the same preference returns.

The final explanation should remain intelligible without the formula: the team selected a particular option for a defined group, accepted specific uncertainty and deferred another use of capacity. The arithmetic supports that explanation; the input definitions and decision owner make it usable.