Measuring Results, Making Decisions, and Using Metrics

A good metric is information that changes the next action, rather than a number for reporting.

Output vs Outcome

A PO should focus on outcomes more than outputs.

Metric Tree

flowchart TD
    B[Business Goal] --> P[Product Outcome]
    P --> U[User Behavior]
    U --> M1[Leading Metric]
    P --> M2[Lagging Metric]

Common Metric Types

Acquisition

Are users arriving?

Activation

Have they experienced the core value for the first time?

Engagement

Are they actually using it repeatedly?

Retention

Do they return?

Efficiency

Have time, clicks, and failures decreased?

Quality

Have crashes, errors, and support issues decreased?

Leading vs Lagging

Leading
= A behavioral indicator that anticipates future results

Lagging
= A final result that has already occurred

Example:

• Leading: Number of Camera Presets saved
• Lagging: Reduction in Camera Setup time

Metric Pitfalls

• Assuming that an increasing number is always good
• Measuring only what is easy to measure
• Looking only at averages and missing differences between segments
• Ignoring the possibility that increased usage reflects increased friction
• Turning a metric into a target and encouraging gaming

Decision-Making Framework

Evidence
↓
Interpretation
↓
Options
↓
Trade-off
↓
Decision
↓
Expected Result
↓
Measure

A decision log should record at least:

• What was decided
• What evidence supported it
• What was given up
• What results would cause the decision to change

For a PO, data is a tool for reducing uncertainty, not "the answer."