An independent study reference written by Dr Phuc V. Nguyen. It is not official subject material — for assessment requirements always follow your subject outline and vUWS.
Structuring a messy problem
Business situations rarely arrive as problems. They arrive as problem situations, with several parties, conflicting goals and no agreed boundary. Soft systems work separates hard problems, where the goal is agreed and only the means are in question, from soft problems, where the goal itself is contested. A rich picture is the standard device for the soft case: a deliberately unstructured drawing of who is involved, what flows between them, what each party wants, and where the tensions sit. It is drawn before any model and it is not a process diagram. Its whole purpose is to surface what would otherwise be assumed.
Why it matters
Imagine sketching an office on one page, with people, arrows, and speech bubbles saying what each person is annoyed about. That sketch is the point. A flowchart has to pretend everyone agrees on the sequence, so it quietly deletes the disagreement. A rich picture has room for two people wanting opposite things, and the disagreement it exposes is usually the reason the original request made no sense.
A rich picture drawn by one analyst alone comes out neat, symmetric and free of any visible conflict. What is the most likely reading?
Worked examples
A hospital outpatient unit asks for "a dashboard to reduce waiting times". A rich picture is drawn on a whiteboard with the clerks, two consultants, a nurse manager and a patient representative present.
The drawing puts clerks, consultants, nurses, patients and the booking system on one page, with arrows for referrals and rescheduling, and annotations for what each party wants. Two conflicts appear immediately. Consultants want low idle time, which favours over-booking. Patients want short waits, which favours under-booking. Clerks absorb the difference by phoning people. Waiting time was never one number. It was the outcome of a trade-off nobody had written down. The useful analytics question becomes measuring the exchange rate between idle clinic minutes and patient waiting minutes, so the trade-off can be set deliberately.
A retail operations manager asks, "why are our customers unhappy?" The analyst has twenty-one days and access to transaction, roster and survey data.
That request is unanswerable as written, because it names no decision and no unit of analysis. Structuring converts it into a question with four parts: the decision (whether to reverse a recent rostering change), the decision-maker (the operations manager), the unit of analysis (a metropolitan store over a seven-day trading period), and a measurable outcome (repeat purchase within 90 days). The result is a question that can be answered wrongly, which is what makes it worth answering. The original phrasing could only ever produce a report.
Common mistakes
- ✗A rich picture is a rough first draft of a process diagram. It deliberately avoids boxes and sequence so that it can hold opinion, emotion and conflict. Redrawing it as a flowchart removes precisely the content it was made to capture.
- ✗A good rich picture is tidy. A neat, symmetric picture with no visible tension usually means one person drew it alone. Its value comes from disagreement becoming visible, so it is normally drawn with the parties in the room and it normally looks untidy.
- ✗Once the picture exists the problem is defined. The picture is an input. A defined analytics question still needs four things named: the decision, the person who makes it, the unit of analysis and the measurable outcome.
- ✗Problem structuring is only worth doing on large projects. The question is the cheapest place in the whole process to be wrong and the most expensive place to stay wrong. Even a two-day task benefits from ten minutes spent naming the decision and the unit of analysis.
Revision bullets
- •Situations arrive as problem situations, not problems
- •Hard problem = agreed goal, contested means; soft problem = contested goal
- •Rich picture: actors, flows, wants and tensions on one page, no sequence
- •Drawn with stakeholders, not for them; untidiness is expected
- •A structured question names decision, decision-maker, unit of analysis, measurable outcome
- •CATWOE is a companion checklist for stating whose view a proposed change serves
Quick check
A rich picture drawn by one analyst alone comes out neat, symmetric and free of any visible conflict. What is the most likely reading?
Which of these is a properly structured analytics question?
Connected topics
More in How Analytics Gets Built
Sources
- Checkland (1981)Checkland, P. Systems Thinking, Systems Practice. John Wiley & Sons, 1981.Origin of soft systems methodology and of the hard-versus-soft problem distinction used here.
- Checkland & Scholes (1990)Checkland, P., & Scholes, J. Soft Systems Methodology in Action. John Wiley & Sons, 1990.Working accounts of rich pictures and CATWOE applied to real organisational situations.
- Monk & Howard (1998)Monk, A., & Howard, S. "The Rich Picture: A Tool for Reasoning About Work Context." Interactions, 5(2), 21-30, 1998.A practical treatment of what belongs in a rich picture and why it should not be tidied into a process diagram.