Problem Framing and Asking the Right Questions

Problem framing is the first step to take when starting a data analysis project. No matter the reason for starting a project, you can use the following who, what, why, where questions to get started:

  • Who has this problem?
  • What is the nature of the problem?
  • Why is this problem worth solving?
  • Where does this problem happen?

Keep the answer to those questions in mind or jot them down. Next, we will get into the problem framing techniques.

Problem Framing Process

You might be wondering, what are the four steps of problem framing?

  1. Understand the business need – time is valuable and nobody wants to do busy work; therefore, the request needs to move the business forward somehow.
  2. Understand the desired outcome – does the stakeholder want a dashboard? a one-off report?
  3. Decide how success will be measured – what KPIs and benchmarks will be used?
  4. Learn who the audience is – a member of the board has much different expectations than a media manager. Knowing your audience is critical to how you present your findings.

Learn about the business need behind the request

Whether you work as a data analyst or aspire to, oftentimes your data analysis project will kick-start when an internal business partner reaches out to you or your team with a data analysis request. These requests can be very vague, therefore it is your job to ask questions to gain a solid understanding of the objective of this analysis. 

Great questions to ask are: 

  • What exactly are they hoping to find out or learn?
  • What is the business need?
  • Experiment to be performed?
  • What KPIs do they want?
  • What are we measuring?
  • What is the desired finished product (e.g., dashboard, analysis, presentation, etc.)?
  • Who is the audience for the finished product (analyst, manager, executive, board member, etc.) ?

Plan the Project

The next steps are to scope the project as best you can. Map out data sources, software to use, the different phases, time for each phase, etc. WWD will cover this part of the process in depth in a future post. Stay tuned!