How would you apply RICE scoring to prioritize these three initiatives?

WHAT THIS TESTS: This question evaluates whether you can bridge engineering execution and product strategy by quantifying technical work in business terms. Interviewers want to see that you do not prioritize by gut feel or loudest stakeholder, but instead apply a repeatable framework with metrics that expose real cost and real value. The scenario deliberately mixes a pure tech-debt play, a speculative integration, and a user-facing feature to see how you compare dissimilar initiatives.

A GOOD ANSWER COVERS: First, define Reach for each initiative in the same unit, such as monthly active users affected or transactions per quarter. For the feature improvements, Reach is the existing user base; for the refactor, Reach is every request hitting that service; for the integration, Reach is projected trial signups. Second, score Impact using concrete metrics. For the refactor, Impact is the 15 percent latency gain translated into conversion or retention lift plus reduced incidents. For the integration, Impact is new revenue. For the feature improvements, Impact is support ticket reduction or satisfaction gain. Third, assign Confidence honestly. The existing feature has usage data, so Confidence is high. The refactor impact is medium if latency gains are proven in staging but production patterns differ. The integration is lowest Confidence because adoption is speculative. Fourth, estimate Effort in person-weeks, not t-shirt sizes. The refactor requires regression testing and rollback planning. The integration demands API review, contract testing, and vendor dependency. The feature improvements have the most predictable scope. Finally, compute RICE scores and rank them, but note that the score is an input, not a dictator, and that strategic context like a looming partnership deadline might override the math.

COMMON WRONG ANSWERS: A major red flag is treating all three initiatives with identical Confidence. Another is using vague effort estimates like small, medium, or large, which collapses under scrutiny when one initiative involves an unknown third-party SLA. Some candidates also conflate Reach with Impact, claiming the refactor reaches the whole system therefore has massive impact, when Reach is about the number of people or events and Impact is the degree of change per person. Finally, failing to mention risk or ongoing operational burden in Effort shows shallow estimation.

LIKELY FOLLOW-UPS: The interviewer may ask what you would do if the lowest RICE score is the CEO's pet project. They might also ask how you would validate the 15 percent latency claim before committing, or how you would sequence the work if two initiatives share the same engineering team. Another common follow-up is how you would adjust the framework for infrastructure work that has no direct user reach.

ONE CONCRETE EXAMPLE: Suppose the feature improvement affects 500,000 monthly users with medium impact and high confidence, requiring four person-weeks. Its RICE score is 500,000 times 1 times 1.0 divided by 4, or 125,000. The refactor affects 10 million requests with low per-user impact and medium confidence, requiring 12 person-weeks, giving roughly 166,667. The integration reaches 2,000 prospects with high impact but low confidence, requiring 8 person-weeks, yielding 250. Here the refactor wins numerically, but you ship the feature first because it is de-risked and user-facing, while scheduling the integration for a later quarter pending a proof of concept.

Source: productplan.com