Dashboard design.

A short list of numbers that each drive a named action and have a proven, current source, then a dashboard built on that list with honest data states instead of blanks and zeros.

You bring:

• the request or spec, treated as a hypothesis
• access to probe each data source
• who looks at it and how often

The steps, in order. The method itself is served whole through the connector, current on every use.

1. 1

   Gate every metric on the action it drives

2. 2

   Probe whether the data is real and current

3. 3

   Scope the operator, isolation and auth

4. 4

   Freeze definitions and data states

5. 5

   Build one number end to end first

6. 6

   Verify like an adversary

• Rebuilding the analytics tool you pull from
• Screens of numbers nobody acts on

Decide-then-build method for any dashboard, monitor, KPI/metrics view, scorecard, control tower, status board, or reporting panel that displays numbers to watch. Triggers across every domain — finance/cash/runway, sales/pipeline, support, eng/ops reliability, product, marketing/ads, personal health. The differentiator: it pulls you from "what chart?" back to "what decision does this drive, and is the data even real?"