Measuring success

Agree on the evidence before claiming the result.

Measures should be selected during discovery or early delivery so the team knows what improvement means and how it will be evaluated.

Data charts on a screen
Measuring success

Process

A clear, staged approach.

  1. 01

    Define the outcome

    State what behavior, workflow, system quality or business result should improve.

  2. 02

    Establish a baseline

    Capture a reliable current-state measurement where a before/after comparison is intended.

  3. 03

    Instrument the system

    Add analytics, telemetry or operational reporting needed to observe the target measure.

  4. 04

    Validate the definition

    Make sure client and delivery teams use the same metric definition.

  5. 05

    Measure over an appropriate period

    Avoid drawing conclusions from an unrepresentative launch spike or a period that is too short.

  6. 06

    Publish with context

    Include source, period, baseline and limitations when they materially affect interpretation.

Metric Source Types

Application/product analytics

Infrastructure/observability telemetry

Business system reporting

Client-provided operational reporting

User research or surveys

Support/incident data

Engineering delivery metrics

Anti Patterns

Selecting a metric only after seeing a favorable result.

Publishing percentage improvements without the original baseline.

Using vanity metrics as proof of business value.

Attributing an outcome entirely to software when other changes contributed materially.

Converting qualitative client feedback into a numerical score without a valid method.