Low-Use Features, Sunk Costs, and a Board Deck
You are the Product Owner for a subscription analytics product with 8,000 accounts. Usage telemetry shows that 40 percent of the features built in the last 18 months are used by fewer than 5 percent of accounts, while the three most-used integrations generate most upgrades. The sales team wants a new custom reporting module because two enterprise prospects asked for it, and a large customer is asking for a feature that was started two years ago and never finished. The team spends about 30 percent of capacity maintaining low-use features. Your executive sponsor says the roadmap must show more features for the next board meeting and is reluctant to remove anything already sold. The next quarter has capacity for roughly 120 story points. A competitor has just released a cheaper dashboard with fewer options.
1. How should you evaluate value and decide what to build, maintain, or remove next quarter?
Read the worked answer
Start from outcomes rather than feature count. Segment usage by account value, retention, and expansion. Determine whether the low-use features are harmless, costly to maintain, or actively confusing. Calculate the opportunity cost of the 30 percent maintenance capacity and compare it with the expected value of the custom reporting module and the unfinished feature. For the enterprise prospects, test whether the request is a real buying condition or a preference: ask for a commitment, a pilot, or a paid discovery. For the unfinished feature, assess whether the original problem still exists and whether a smaller slice would deliver value. Prioritize the most-used integrations because they correlate with upgrades, and protect capacity for reliability and onboarding. A sensible next-quarter mix might fund one integration improvement, a time-boxed validation of the enterprise report, and the removal or simplification of several low-use features, rather than adding net new scope.
2. How do you handle sunk cost and executive pressure for a roadmap that shows more features?
Read the worked answer
Name the sunk-cost trap directly. The two years already spent on the unfinished feature are irrelevant to its future value; only remaining cost, risk, and expected outcome matter. The same applies to features already sold: maintenance is a continuing investment, and keeping a low-value feature is a decision to fund it again every quarter. Give the sponsor a board narrative based on customer outcomes, revenue retained, and capacity reallocated, not a longer feature list. Show the cost of the current portfolio and the opportunity cost of not addressing the competitive threat. Offer a compromise that is still honest: a roadmap with fewer, sharper bets and explicit stop or continue decision points. If the sponsor insists on listing more features, label them as options with validation gates, not commitments. Do not let the board deck drive delivery commitments that the evidence does not support.
3. Which metrics would you use to test whether the decision adds value, and how would you avoid vanity measures?
Read the worked answer
Use metrics tied to behavior and money, not output counts. Track adoption by cohort, activation and time to first value, retention and expansion revenue, support tickets per account, task success in the workflows that matter, and cost to serve. For the enterprise reporting request, define a success test before building: for example, a qualified pilot with agreed usage and a commercial commitment. For feature removal, monitor migration, complaints, and whether the accounts that used the feature churn or simply move to a better path. Compare the maintenance capacity freed with the value delivered elsewhere. Avoid measuring story points delivered, number of features shipped, or raw page views without context. Review the evidence at a fixed interval and be willing to stop, continue, or reverse the decision. That is how the team learns whether the portfolio change actually added value.