1.5 From concern to analytical question

Suppose a service organization says, “Demand seems to be increasing, and we do not know whether current capacity will be enough.” This statement identifies a concern, but it does not yet specify what should be measured.

The concern can be separated into several analytical questions:

• How has service volume changed over time?
• How has utilization changed by location and age group?
• Is the change associated with population growth, changing utilization rates, or both?
• What demand would be expected if current group-specific utilization rates continued?
• Which groups or locations account for the greatest projected increase?
• Which operational constraints would become important under each scenario?

Each question implies different data, definitions, methods, and limitations. The first is descriptive. The third requires compatible population denominators. The fourth is a conditional projection rather than a promise about the future. The final question may require information that is not present in the service dataset at all.

Worked Example: Refining a Broad Request

A retailer asks, “Which customers should we target?”

A useful refinement process might produce the following questions:

1. Which customer groups differ meaningfully in purchase frequency, contribution margin, product mix, or retention?
2. Are the differences stable across time periods and sales channels?
3. Which groups have sufficient size and potential value to justify a distinct strategy?
4. What action could the retailer take for each selected group?
5. Which metric would indicate that the action was working?

The refined questions prevent the team from creating attractive customer labels that have no analytical or operational purpose.