The Davenport-Kim Three-Stage Decision Model
The Davenport-Kim Three-Stage Decision Model
Plain-Language Definition
The Davenport-Kim model, developed by Thomas Davenport and Jinho Kim, breaks any data analysis project into three stages: framing the problem, solving the problem, and communicating the results. It’s less a statistical technique and more a project structure — a reminder that the math is only the middle third of the work.
Why It Matters
Students often jump straight into calculations without properly framing the problem first, then struggle to communicate results clearly at the end. Naming these as three distinct stages — each with its own required skill set — helps you allocate effort correctly instead of treating the whole task as “just do the statistics.”
The Three Stages
- Framing the problem — recognizing what the actual problem is, identifying stakeholders, reviewing what’s already known, and getting specific about what you’re trying to find out. This stage is conceptual, not mathematical, and skipping it is why some analyses answer the wrong question well.
- Solving the problem — the modeling, data collection, and data analysis work. This is the stage requiring the most intensive statistics and technical effort.
- Communicating results — translating findings into a business recommendation a non-technical stakeholder can act on. Getting the math right but failing here is one of the most common ways students lose points on written reports.
Worked Example (Fictitious Company)
Applying the model to Alderbrook Coffee Roasters’ decision tree task:
- Frame: The COO needs to choose between three alternatives under uncertain demand — the problem is a resource-allocation decision under risk, not simply “which option is biggest.”
- Solve: Calculate payoffs and expected values for each alternative using the provided probability and demand data.
- Communicate: Present a primary recommendation and a backup option, explained in terms the CEO — who isn’t going to read the raw calculations — can act on immediately.
Key Takeaways
- Framing, solving, and communicating are three distinct skills, not one undifferentiated task
- Most of the technical statistics work belongs in the “solving” stage — the other two stages are about clarity and judgment, not calculation
- A technically correct “solve” stage still fails the assignment if the “communicate” stage doesn’t translate the results into an actionable recommendation
- Rushing through “framing” to get to the numbers faster is one of the most common, avoidable causes of a wrong or irrelevant business question
Related Content
- C207 Course Roadmap
- C207 QUM3 Task 2 Guide: Decision Tree Analysis
- Decision Tree Analysis, Explained
References & Further Reading
- Davenport, T. H., & Kim, J. (2013). Keeping Up with the Quants: Your Guide to Understanding and Using Analytics. Harvard Business Review Press. — The original source of the three-stage framework this page explains. Note: this is the model’s originating practitioner text, not a peer-reviewed journal article; no recent journal-published evaluation of this specific named framework was found, so the primary source is cited directly per standard academic practice for attributing a named model to its creators.