Business Context

You’re the analytics partner for Uber Eats Merchant Growth. The team is building a new feature called “Smart Promotions” that helps restaurants automatically create discounts during slow hours. The feature is expected to drive +0.6% marketplace gross bookings in the next quarter and is tied to a public commitment in the quarterly business review (QBR). The org runs 2-week sprints with a 6-person engineering squad, 1 PM, 1 designer, and shared QA.

Over the last 4 sprints, leadership has noticed that the team is missing milestones: the beta launch slipped by 3 weeks, and cross-functional partners (Sales, Support, Legal) are escalating because merchant comms and training depend on the delivery date. The EM claims the team is “working hard” and points to high commit counts and long hours. The PM believes scope creep is the issue. Finance is worried because the delay pushes expected revenue impact out of quarter.

You have access to Jira, incident logs, and production deploy metadata. You’re asked to use velocity and related delivery metrics to correct the team’s trajectory—not just report what happened, but identify the most actionable drivers and propose a plan that can be executed within the next 2 sprints.

Metric Scenario

• Sprint cadence: 2 weeks; typical sprint capacity is ~60 story points historically.
• Recent trend (last 4 sprints):
  • Story points committed: 62, 65, 70, 68
  • Story points completed: 55, 44, 41, 39
  • Unplanned work tickets (production issues, urgent requests): rising from ~5% to ~25% of completed work.
• There was a platform migration (payments service) and a spike in on-call pages in the same period.

Stakeholders are asking:

• “Is this a real productivity problem or a measurement artifact?”
• “What changed, and what should we do next sprint?”
• “Can we still credibly hit the beta date? If not, what scope should we cut?”

Data Available

Requirements (What you must produce)

1. Define the primary velocity metric you will use (and explicitly state what it is not measuring). Include how you handle spillover work, bugs, and partial completion.
2. Compute and interpret the last 4 sprints’ trajectory: is the team trending off-plan relative to baseline and commitments? Quantify the gap and forecast the next 2 sprints under current conditions.
3. Decompose the velocity change into actionable drivers (e.g., unplanned work, incident load, workflow bottlenecks, scope volatility, estimation drift). Prioritize the top 2–3 drivers with evidence you’d look for in the available data.
4. Propose specific interventions for the next 2 sprints (process changes, scope cuts, staffing/on-call changes, quality investments). For each, state the expected metric movement and what would convince you it worked.
5. List guardrail metrics to ensure “velocity recovery” doesn’t come from harmful behavior (e.g., point inflation, quality regressions, burnout).

Constraints:

• You cannot add headcount this quarter.
• Any plan must maintain reliability: Sev-0/Sev-1 incident response SLAs cannot degrade.
• The PM needs a recommendation in 72 hours for a roadmap re-plan meeting.