Unit 13.04: The story, the limits and what you would build next

The handover names the analysis that could overturn the finding.

Limits, and what would change the conclusion

Decision, summary, the reader test and its fix, limits, and next steps.

The code prints the handover note.

import json

handover = {
    "decision_supported": "which two suppliers to renegotiate with",
    "summary_sentence": "Two suppliers account for 60% of spend at above-median "
                        "unit cost.",
    "charts": 3,
    "reader_test": {"readers": 3, "reached_the_finding": 2,
                    "question_all_failed": "what period is this",
                    "fixed": "period added to every subtitle"},
    "known_limits": ["one quarter", "list price, not landed cost",
                     "excludes freight, which may reverse the ranking"],
    "not_shown_deliberately": ["individual buyer performance"],
    "what_would_change_the_conclusion": "landed cost including freight",
    "next_build": "add freight and re-run; extend to four quarters",
}
print(json.dumps(handover, indent=2))

print("\n`what_would_change_the_conclusion` is the field that makes this")
print("honest -- it names the analysis that could overturn the finding.")

what_would_change_the_conclusion is the field that makes this honest. Freight is excluded, freight could reverse the supplier ranking, and saying so is stronger than letting someone discover it after the renegotiation.

The reader test is recorded with its outcome and the fix, so whoever inherits this knows the charts were tested and what was found.

The mistake this prevents

The mistake is listing limits that do not threaten the conclusion - sample period, data freshness, the usual list - while omitting the one that does. The useful limit is the one that could change the answer.

Takeaway

State what would change the conclusion, not just what the analysis excluded. A limits list with nothing threatening in it has not been thought about.

Mark complete