Cobalt Robotics: One Quarter, Five Demands
You are the PM for Cobalt Robotics's industrial robotics platform. You have 5 engineers for a 12-week quarter — roughly 60 engineer-weeks of capacity, before the usual 20-30% goes to support and unplanned work.
Five things are on the table:
A. Enterprise SSO. Blocking $660k of annual contract value across three deals in late-stage procurement. Estimated 11 engineer-weeks.
B. Performance work. The app's slowest screen takes 6.5s to load. Support cites it as the top complaint. Estimated 11 engineer-weeks.
C. Onboarding redesign. Activation sits at 44%; research suggests a redesign could add 6 points. Estimated 15 engineer-weeks.
D. Platform migration. Technical debt is slowing every team by an estimated 40%. Estimated 26 engineer-weeks.
E. The CEO's feature. The CEO promised a specific customer a custom reporting feature at a conference. 5 engineer-weeks. No other customer has asked for it.
Sales says A. Support says B. Growth says C. Engineering says D. The CEO says E.
risk
- enterprise churn risk accounts
- 4
items
- C onboarding
- [object Object]
- B performance
- [object Object]
- E ceo feature
- [object Object]
- A enterprise sso
- [object Object]
- D platform migration
- [object Object]
capacity
- weeks
- 12
- engineers
- 5
- gross engineer weeks
- 60
- realistic capacity pct
- 70
Set the quarterly roadmap. Provide:
- Analysis — how you evaluate and rank these, with your reasoning shown.
- Risks — of your choices, including what breaks if you're wrong.
- Recommendation — the committed plan, and how you communicate the no's.
You cannot do everything. Be explicit about what you are cutting.
100 points, 60% to pass.
- recommendation30
- framework application25
- quantitative reasoning25
- stakeholder management20
Reveal suggested structure
- Real capacity = 60 × ~70% = ~42 engineer-weeks. Planning against gross capacity is the most common roadmap error.
- Score consistently — RICE (reach × impact × confidence ÷ effort) or cost of delay ÷ duration.
- Convert to money or users where possible; A is already in revenue terms, C can be, B and D are indirect.
- Treat D as an investment — it compounds, so delaying it is a growing cost, not a fixed one.
- Handle E on its merits, not its source.
- Sequence, leave slack, and communicate the trade-offs.