Project Background
NovaMart, a mid-sized e-commerce marketplace, produces 14 recurring business reports each week for executives, finance, operations, and category managers. Today, analysts spend about 32 hours per week manually pulling data, cleaning spreadsheets, validating numbers, and formatting slides. You are the program manager assigned to improve this workflow before Q4 planning begins in 10 weeks.
Key Stakeholders
The CFO wants fewer reporting errors and faster month-end close. The VP of Operations wants daily visibility into fulfillment and returns. The Head of Analytics wants to reduce analyst time spent on repetitive work, while the Engineering Manager is concerned about adding fragile automations that create maintenance burden. These priorities conflict: executives want speed and polish, but engineering wants to limit scope and technical debt.
Constraints
- Timeline: 10 weeks, with a hard deadline before the Q4 business review
- Team: 2 data engineers, 1 analytics engineer, 2 business analysts, and you as PM
- Budget: $85,000 for tooling, contractor support, and QA
- Dependencies: finance metric definitions must be finalized by Week 3; data platform team can only support 6 hours per week
- Current state: 14 reports, 6 data sources, and 11 known data quality issues
Complications
- Two executive reports use slightly different definitions of "net revenue," and Finance has not approved a single source of truth.
- One critical fulfillment report depends on a vendor CSV file that arrives late 30% of the time.
- The CEO has asked whether all reporting can be fully automated this quarter, increasing pressure to overcommit.
Your Task
Produce the following:
- A framework for deciding which parts of the reporting workflow to automate first
- A 10-week execution plan with milestones, owners, and dependencies
- A recommendation on what should remain manual for now and why
- Success metrics for the project, including time savings, accuracy, and adoption
- A risk mitigation and stakeholder communication plan if full automation is not feasible by the deadline
