All open work
BUILD 02 / Decision Brief Studio

A vendor comparison shows where the recommendation can break.

A synthetic three-option reporting-platform choice is scored on workflow fit, three-year cost, data control, and time to usable. The example intentionally includes one low-confidence migration estimate so the recommendation cannot pretend every input is equally strong.

Demonstration with fictional vendors and synthetic evidence. The method is real; the product claims and recommendation are not market facts.
01 / INPUT CONTRACT

Reporting-platform decision

3 options · 4 weighted criteria · 3 evidence records

  • the team has three operators
  • data export is mandatory
  • the current process consumes two days monthly
  • the decision is due within four weeks
02 / METHOD
01

Define a finish line

The decision statement names the date, constraints, and the work the selected option must support.

02

Write criterion definitions

Each score is anchored to a stated dimension instead of being a free-floating preference.

03

Record evidence confidence

Official documentation, a direct test, and an internal estimate do not receive the same confidence treatment.

04

Stress the ranking

The recommendation is reviewed for weight sensitivity and low-confidence dependencies before approval.

03 / OBSERVED OUTPUT
3.61

top adjusted score

The leading option combines a high raw score with stronger evidence confidence.

100%

weight visibility

Every point in the scoring rule is named and can be challenged.

1

weak assumption

The migration-time estimate needs a direct test before it supports a final choice.

4

exports

Brief, score matrix, evidence register, and reopenable project record.

DECISIONS FROM THE BUILD
  1. Keep the current leader as the provisional recommendation.
  2. Run a two-hour migration spike before approval.
  3. Reverse the choice if data export fails the representative-file test.
  4. Record the approval owner and review the decision after the first monthly cycle.
05 / RELEASE CHECK

What had to be true.

  • Weights total 100%
  • Every criterion has a definition
  • Low-confidence claims are visible
  • Recommendation is tied to the declared rule
TRY THE PUBLIC EDITION

Run the method on your own input.

Open the free tool See the complete paid edition