The Feature Factory
Sales wants the feature. The VP wants the feature. Nobody wants to ask if the last five features anyone wanted actually worked.
You're advising Alicia, a product manager at Ledgerly, invoicing software for freelancers. Her roadmap for the next quarter is already full: five items, every one of them requested by a stakeholder. She pulls up usage data on the last five "must-have" features her team shipped, all sales- or exec-requested, none scoped with a target outcome beyond "ship it":
| Feature (last 2 quarters) | Requested by | Adoption (30 days post-launch) | Still used weekly (90 days) |
|---|---|---|---|
| Custom PDF branding | Sales (deal-specific ask) | 11% of eligible accounts | 4% |
| Multi-currency invoicing | Exec (competitor parity) | 18% | 9% |
| Recurring invoice templates | Sales | 24% | 15% |
| Slack notifications | Support (ticket volume) | 6% | 2% |
| Client portal v1 | Exec (board deck promise) | 14% | 7% |
None of these moved the metric that actually matters for Ledgerly: percentage of active accounts that send a second invoice within 30 days of their first (their best proxy for habit formation and retention), which has been flat at 61% for three quarters regardless of what's shipped.
Now the newest ask lands: Bram, the VP of Sales, is in Alicia's inbox and on her calendar. Ledgerly's largest customer, an agency worth 8% of total revenue, wants a custom approval-workflow feature — multi-step invoice sign-off before a client sees it — or they're evaluating a competitor at renewal in six weeks. Bram wants it on the roadmap now, ahead of everything else, and has already told the customer's champion it's "in progress." Building it properly, generalized enough to not be a one-off, is a 6-8 week effort for two of Alicia's four engineers — more than half the team, most of the quarter.
Alicia has a hunch, but no proof yet, that the approval-workflow request is really a proxy for a narrower job: this one client's finance team doesn't trust what goes out the door without a second set of eyes. She hasn't talked to the client directly — only heard the ask secondhand through Bram.
Data snapshot
Bram wants an answer today: is the approval-workflow feature going on the roadmap, full scope, starting now? Write your answer as Alicia. Take a clear position on what you'll say yes and no to this quarter, explain what's broken about how the last five features were chosen (using the adoption data), and lay out specifically what you'll do about the at-risk customer in the next two weeks that isn't simply building what Bram already promised.
Rubric
- Clear verdict. States plainly whether the full approval-workflow feature goes on the roadmap as scoped, and what happens instead — no answer that avoids committing to a position on this quarter's plan.
- Diagnoses the feature-factory pattern. Names the actual failure mode using the data: the team is a feature team executing a stakeholder wish list (output), not a product team accountable for an outcome, and the 61%-flat retention metric proves the last five 'must-haves' didn't move anything that mattered.
- Separates the customer's problem from the customer's solution. Recognizes that 'approval workflow' is Bram's/the client's proposed solution, not a validated problem, and proposes finding the underlying job (e.g. trust in what goes out the door) before committing engineering time to the full generalized build.
- Concrete near-term plan for the at-risk account. Gives a specific action within the 6-week renewal window that protects the relationship without abandoning outcome discipline — e.g. a fast narrow/manual fix, a direct discovery conversation with the client, or a scoped prototype — not silence and not a blank check.
- Handles the stakeholder relationship. Addresses how Alicia manages Bram directly: correcting the 'already promised' commitment, engaging him as a partner in discovery rather than a request-taker, without being either a pushover or dismissive of a real revenue risk.