Skip to main content

​
Overview

The Sleep Score is an Open Wearables native score - it is computed directly from the raw sensor data synced through the platform, independently of any manufacturer score. The same algorithm runs regardless of the device or provider the user wears, giving you a consistent, comparable score across your entire user base.

​
Requirements

For a Sleep Score to be generated, the following conditions must be met: Data syncing
  • The user must have at least one connected provider that syncs sleep data.
  • The provider must report total sleep duration (net sleep time, awake periods excluded).
Device and provider requirements by component

​
Score Components

The overall Sleep Score is a weighted average of four independently-scored pillars.

​
Duration

Scores how long the user actually slept (net sleep time, after subtracting awake periods).

​
Stages

Scores the quality of sleep architecture by comparing Deep and REM stage durations against targets. Each stage is scored independently and combined with equal weighting. The two stage scores are combined at 50/50 to produce the Stages component score.

​
Consistency

Scores how closely the user’s bedtime matched their own historical routine - not against a fixed clock time. This encourages habit stability rather than conforming to an external schedule.
  • A 14-night rolling median bedtime is computed from prior nights.
  • A 15-minute grace period is applied symmetrically - minor variation is not penalised.
  • Going to bed late incurs a linear penalty up to 100 points across a 105-minute window past the grace period.
  • Going to bed early incurs a softer penalty, capped at 20 points - early nights are penalised less than late ones.
No historical data means no consistency score can be computed (returns 0). The score becomes meaningful after a few nights and stabilises after ~14.

​
Interruptions

Scores sleep continuity based on wake-after-sleep-onset (WASO) - time spent awake between falling asleep and the final wake-up, excluding sleep latency. The component has two sub-scores that are summed: WASO duration (80 points available) Awakening frequency (20 points available)

​
Weights

The four pillars are combined into the overall score using a weighted average. The default weights reflect clinical research on the relative importance of each factor to sleep quality. Weights are adjustable. If your application has different priorities - the pillar weights can be customised. The only constraint is that they must sum to 100%. Contact us or refer to the configuration reference to adjust weights for your deployment.

​
Score Breakdown

Every Sleep Score response includes a full breakdown payload so you can surface the component scores in your UI: