How to read a Sanrui case
What makes a recommendation worth following?
A recommendation should do more than sound convincing. You should be able to see what was considered, where the conclusion could fail and what would justify changing course.
What other choices were considered?
Begin with the actual question, not a preferred answer. List the relevant alternatives, including keeping the current arrangement, seeking other expertise and deciding not to buy a consultation. “All relevant alternatives” is a bounded working list, not a claim that every possible future is known.
- Example
- An office considering a move can keep its layout, change one operating arrangement, commission a relevant assessment or compare actual alternative premises.
- Exclude for a reason
- An option may be unavailable because of safety, permission, timing or cost. Record that reason; an unfavourable traditional reading alone is not a safety finding.
What would show the interpretation is wrong?
For each important hypothesis, state the strongest alternative explanation and the observation that would weaken it. An outcome that did not occur remains in the record. Do not reinterpret the original question after learning the result.
| Check | Question |
|---|---|
| Within the method | Are the inputs, rule version and chart relationships recorded consistently? A consistent calculation is not proof of predictive accuracy. |
| In everyday life | Did the recorded problem change under comparable conditions? What else changed at the same time? A before/after change alone does not establish causation. |
Which conditions must be true?
Name the assumptions that carry the recommendation. “Given the available material” is not enough: say which missing fact affects which conclusion. An unverified input stays unverified, even if an exact-looking chart can be displayed.
- Space: the orientation, reference north, building period, centre and room boundaries used for this reading.
- Life stages: the birth input, its uncertainty and the actual options available to the person.
- One decision: the question, chart time, practical deadline, decision authority and constraints.
- Execution: the people involved agree, costs are acceptable and the action does not breach safety or contractual requirements.
What happens beyond the first result?
A low purchase price can conceal time, maintenance, disruption and the cost of reversing a change. Moving a desk affects other people; moving premises creates further commitments. Repeatedly seeking a new reading can also make an unresolved choice feel more urgent without adding evidence.
Judge the original problem, adverse effects and total cost together. Do not change a working arrangement merely because a calendar period or a software score has changed. A fresh reading needs a fresh question or a material change in the facts.
What is recommended now, and when should it stop?
The recommendation can be “not enough information” or “this needs a different professional”. A review date is a management checkpoint, not a forecasted date of success. More methods, a longer report or frequent follow-ups do not automatically create more value.