Soft vs. Hard Problems

Problems often range on a continuum between:

"Hard" Problems

Problems that are well defined or well structured. These are routinely solved by application of a well-understood formula, process or design.
Well-understood process, stable
Examples: Scheduling, lease versus buy decisions

"Soft" Problems

Problems that are highly dependent upon how they are perceived by the participants.
Ambiguous, volatile, poorly-understood and dynamics.
Examples: Motivation, change management,

Many problems fall between the extremes: Organization reengineering, policy, compensation, and quality.+

Some problems that appear to be "hard" have many "soft" elements. Consider a competitive bid to supply a computer component. Costing the design and production is an engineering function. However, in pricing the component, one would consider:

Competitiveness of competitors
Value as perceived by the purchaser (and possibly also the end user)
Relationship with the purchaser, e.g., potential to do other future business

The challenge for both the analyst and client is to incorporate everything that is important to the problem into the analysis.

Highly subjective beliefs can be modeled or quantified as part of a numerical assault on the problem. Quantification allows elements of the problem to be kept in perspective. This is better than relying upon our intuitions alone.

Inspired by remarks in "Creativity in OR/MS: From Technique to Epistemology" by Haridimos Tsoukas and Demetrios B. Papoulias, Interfaces, 26: 2 Mar-Apr 96, pp. 73-79.

—John Schuyler, March 1998.