The same headline number can mean very different things. Before you cite a StatFacts insight, know which unit you are looking at.

Relative percent change

Example: +12–18% signup completion

This is a relative lift on the baseline rate.

• Baseline 20% → upper bound ≈ 23.6% (20 × 1.18)
• Baseline 5% → upper bound ≈ 5.9%

The absolute gain in points shrinks when your baseline is low. A “+15%” story can be a fraction of a point in practice.

When to use: Comparing proportional improvement across teams, prioritizing levers when baselines are similar.

Percentage points

Example: +8% signup conversion (percent_point)

This is an absolute shift on the rate itself.

• 30% → 38% conversion (+8 points)
• Not the same as “+8% relative” (30% → 32.4%)

When to use: Finance models, funnel math, forecasting revenue from conversion changes.

Mixed or directional labels

Some insights (especially sports or health) report mixed effects—one metric up, another down. Read the outcome field and body text together; do not cherry-pick the favorable line.

Quick checklist

1. Find effect_unit on the insight page.
2. Write down your current baseline before applying the range.
3. State both relative and absolute impact if your audience includes finance and product.

Rule of thumb

If someone says “conversion went up 10%,” ask: 10% of what, or 10 points? StatFacts tries to label this explicitly—your job is to carry that precision forward.