Working with data and metrics
The ability to choose the right metrics, collect and validate data, interpret it without distortion and use it for decisions rather than for decorating a report.
Definition
BT360 v2.0.0The ability to choose the right metrics, collect and validate data, interpret it without distortion and use it for decisions rather than for decorating a report.
What is it made up of?
3 INDICATORS- 01Choosing relevant metrics
- 02Collecting and validating data quality
- 03Interpreting and using data in decisions
Behavior levels 0–5
MANAGER'S GOAL - 4| Lv. | Label | Descriptor |
|---|---|---|
| 0 | Absent | Does not work with data; picks metrics at random or uses ready-made ones without understanding; draws conclusions from a single number; does not notice errors in the data. |
| 1 | Poorly developed | Calculates simple indicators; does not check the source or completeness; confuses absolute and relative values; uses data to confirm a conclusion they already had. |
| 2 | Underdeveloped | Works with a ready-made dashboard; can name a metric but not explain its link to the goal; notices obvious errors; interprets superficially. |
| 3 | Normally developed | Chooses metrics to fit the goal; checks data completeness and quality; distinguishes correlation from causation; qualifies the conclusion; uses data in decisions. |
| 4 | Highly developed leader's goal | Builds a metric system and tests it for distortion; works with incomplete and noisy data; separates signal from noise; explains the limitations; teaches the team data literacy. |
| 5 | Overdeveloped | Data replaces judgment: cannot decide without 'one more number'; metrics are optimized at the expense of meaning; reporting grows, decisions do not change; numbers are used to apply pressure. |
How is it measured?
coating BSJT - BT360 situational test: provides a hypothesis about judgments in work situations and does not replace observation, 360 and interviews.
DMC provides self-assessment of decision competence, but does not test data management. The main method is a practical task with a set of contradictory data.
Development Actions
5 CATEGORIESDMC and a practical task with contradictory data; checks the ability to choose a metric and explain its limitations.
Course 'Metrics and Data for Decisions': metric selection, data quality, causality, visualization, traps.
Rebuild the metrics of one process: goal — metric — source — frequency — owner; remove two useless ones.
Review with an analyst one conclusion that did not hold up; find the error in the interpretation.
Adopt a team rule 'every metric has a decision it changes'.