User Engagement Score is a composite metric—a weighted index that summarizes how deeply a user interacts with your product or content across multiple signals. Unlike a single metric such as Session duration or Pageview count, it combines behaviors like Time on Page, Scroll Depth, meaningful Click Events, and downstream outcomes like Conversion Rate. There’s no universal standard; you define the inputs and weights to match your business model. Content site? Weight reading depth and time. E-commerce? Give more weight to add-to-cart and conversion.
How to calculate it (framework)
- Choose signals (events/metrics) that reflect real engagement: time, depth, actions, return frequency, conversions.
- Normalize each signal to 0–1 (min–max, percentile, or z-score bucketed to 0–1) so they’re comparable.
- Assign weights that sum to 1 based on impact.
- Compute an index and scale it to 0–100 for readability.
Formula:
User Engagement Score = 100 × Σ (wᵢ × xᵢ), where xᵢ are normalized signals and Σwᵢ = 1.
Mini example (session-level)
| Signal | Normalized (xᵢ) | Weight (wᵢ) | Contribution (wᵢ×xᵢ) |
|---|---|---|---|
| Time on Page | 0.60 | 0.30 | 0.18 |
| Scroll Depth | 0.80 | 0.20 | 0.16 |
| Meaningful Click Events | 0.40 | 0.20 | 0.08 |
| Goal hit → Conversion Rate proxy | 1.00 | 0.30 | 0.30 |
| Total | 0.72 → Score = 72/100 |
Why it matters
- Reduces noisy, siloed metrics into one interpretable signal for dashboards and alerts.
- Aligns teams around what “good engagement” means for your product, not a vendor’s default.
- Correlates engagement with outcomes (revenue, retention, churn) better than raw Bounce Rate or average Session length.
Implementation tips
- Start simple (3–5 inputs). Avoid double-counting highly correlated signals (e.g., Average Session Duration and Time on Page).
- Validate weights against business outcomes and iterate quarterly.
- Track distribution (p50/p90) and segment by channel, device, and cohort (Cohort Analysis).
- Document normalization and weights to keep the score comparable over time.