Skip to content
Structured Decision Making | Part 6 of 6

Decide, Record and Review: The Decision Note and What It Teaches

What belongs in a decision note, why assumptions must be numbers with a range, how a decision register works and what a review after twelve months shows.

After eight weeks the plant had everything on the table: three scored options, a cost comparison, three cases and a list of what would change the ranking. The plant manager decided for option C, a 3 kW laser cell for the thin sheet family, with thick plate purchased. The decision took ten minutes. What made it last was the page written afterward.

A decision that is not written down was not made

Six months after a decision, three things are usually forgotten: which options were rejected and why, what was assumed, and under which conditions the yes was given. The first bad month then reopens everything. A decision note prevents that. It has seven lines.

  • Question. The decision in one sentence, with a date.
  • Options. All of them, including the excluded ones with the reason.
  • Reason. Two lines: the criteria and the number that decided.
  • Assumptions. Each as a number with a range and a source. “Volume stays stable” cannot be reviewed, “420,000 blanks, range 336,000 to 483,000” can.
  • Conditions. What must be true or signed before the decision is carried out.
  • Reconsider if. The threshold at which the decision is opened again.
  • Owner and review date. One name, one date.

The note of the plant

  • Decision: option C.
  • Reason: highest weighted result at 7.70 and the lowest five-year cost in all three cases. The extra EUR 430,000 for option B would pay back in 7.7 years, beyond the horizon of five.
  • Assumptions: 420,000 blanks per year, range 336,000 to 483,000. Thick plate at EUR 4.70 per blank, range 4.50 to 5.20. Variable cost in house EUR 2.52 per blank.
  • Conditions: a framework agreement for thick plate with a price clause and a 24-month term, signed before the cell is ordered.
  • Reconsider if: volume passes 500,000 blanks per year, or the thick plate price rises by more than 11 percent.
  • Excluded: option A failed the capacity must, option E failed on hall height and shift model.
  • Review: twelve months after the first part.

The line “reconsider if” is the least common and the most useful. It turns reopening a decision from a mood into a rule.

The review

Twelve months after the first part, the plant put the numbers side by side.

Assumed Happened
420,000 blanks per year 408,000
Variable cost EUR 2.52 per blank 2.61
Annual cost EUR 1,556,700 1,546,700
Payback in years 1.5 1.9
Thick plate lead time in working days 12 14

The volume was 3 percent below plan and inside the range. The annual cost was lower, but only because the volume fell. The variable cost was 4 percent above the estimate, because cutting gas had risen by 12 percent. The payback moved from 1.5 to 1.9 years and stayed well inside the horizon.

A review does not ask whether the outcome was good. A good outcome can follow a poor decision, and the reverse. It asks whether the assumptions held, and what the method should do differently next time.

What the plant learned

The decision held. The method changed in one place. The variable cost had been entered as a single number where a range belonged. From the next decision of this weight on, every cost estimate enters the sheet as a low, an expected and a high value, and the sensitivity check is filled before the decision, not after it. The safety stock for thick plate was raised by two days.

The register

One decision teaches little. Twenty teach a lot. The plant keeps one register with a line per decision: number, question, class, decider, date of request and of decision, review date, status and one sentence on what was learned. Two things become visible that no single note shows: how long decisions take and where they queue, and which reviews are overdue. After a year, the column with the lessons is the best training material the plant owns.

Newsletter

One Lean idea.
Every two weeks.