When Apple released iOS 18 in late 2024, it quietly removed a feature that millions of users depended on: the ability to track sleep using just an iPhone. The “Time in Bed” feature — which used the phone’s sensors to estimate when a user fell asleep and woke up — was discontinued. Sleep tracking on Apple devices now requires an Apple Watch.
For developers building health-aware apps, this created an immediate problem. A significant portion of their user base — those with iPhones but no Apple Watch — suddenly had no passive sleep data. Sleep scores couldn’t be computed. Sleep-dependent features went dark for these users. And there was no alternative from Apple.
Sahha has developed a proprietary algorithm that fills this gap, estimating sleep timing from passive smartphone data alone — restoring the sleep signal for smartphone-only users without requiring any additional hardware.
What Apple removed and why it matters
Prior to iOS 18, iPhones tracked a “Time in Bed” metric by monitoring device usage patterns — when the phone was put down, left untouched, and picked up again. It wasn’t true physiological sleep tracking (no heart rate, no motion-based sleep staging), but it provided a useful approximation of sleep timing and duration that the Health app recorded automatically.
Apple’s decision to remove this feature and require Apple Watch for all sleep tracking left a meaningful gap in the health data ecosystem. The majority of iPhone users don’t own an Apple Watch. For them, passive sleep data simply stopped being collected.
For product teams, the impact was concrete:
- Sleep scores that relied on Apple Health sleep data stopped generating for iPhone-only users
- Readiness scores that used sleep duration as a contributing factor lost a key input
- Sleep-related features like bedtime reminders, sleep trend analysis, and recovery recommendations lost their data source
- User engagement around sleep-related content dropped for the affected segment
How Sahha’s smartphone sleep estimation works
Sahha’s algorithm estimates sleep timing using signals that smartphones passively collect — device interaction patterns, ambient sensor data, and behavioral indicators. The system infers when a user likely fell asleep and woke up without requiring any manual input or additional hardware.
The approach differs from what Apple’s Time in Bed feature did. Rather than relying on a single signal (device untouched), Sahha’s algorithm combines multiple passive data streams to produce a more robust estimate. The algorithm was developed and validated against wearable-derived sleep data to ensure the estimates are meaningful.
What it provides
Sleep timing — estimated sleep onset and wake time for each night. This is the core signal that enables downstream sleep analytics.
Sleep duration — total estimated sleep time derived from the timing data. This feeds directly into sleep scores, readiness scores, and sleep debt calculations.
Sleep consistency — with timing data across multiple nights, the system can assess regularity of sleep patterns, which research shows is independently predictive of health outcomes.
What it doesn’t replace
Smartphone-based estimation does not detect sleep stages — deep sleep, REM sleep, and light sleep require physiological sensors (typically optical heart rate) that only wearable devices provide. For product teams that need sleep stage data, a wearable integration remains necessary.
The practical implication: smartphone sleep estimation covers the foundational metrics (duration, timing, debt, consistency) that the broadest user base needs, while wearable data adds the deeper recovery signals (deep sleep, REM, HRV during sleep) for users who have the hardware.
Why smartphone sleep data matters for product teams
Sleep is the single most impactful health behavior for daily performance, mood, and long-term health outcomes. It’s also the health metric users engage with most consistently. Losing sleep data for a significant user segment doesn’t just affect one feature — it degrades the entire health experience.
Coverage over precision
Most health apps serve a mixed user base: some users have wearables, most don’t. A product that only delivers sleep features to wearable owners is leaving the majority of its users without a core feature.
Smartphone sleep estimation extends coverage to 100% of the user base. The estimates are less granular than wearable data, but they’re sufficient for the features that drive daily engagement: sleep scores, trends, bedtime recommendations, and readiness assessment.
Continuity across platform changes
Apple’s iOS 18 change illustrates a broader risk: platform dependencies create fragility. When a health data source depends on a single platform vendor’s implementation decisions, a single OS update can break features for millions of users.
Building on a data infrastructure that can fill these gaps — using proprietary algorithms when platform-level data disappears — provides resilience against future platform changes.
Availability
Smartphone sleep estimation is available now through the Sahha SDK on both iOS and Android. The feature works automatically — no additional permissions or configuration are required beyond the standard SDK integration.
Sleep data from the estimation algorithm flows through the same pipeline as wearable-derived sleep data: it feeds into sleep scores, readiness scores, sleep debt calculations, and Insights (trends and comparisons). For developers already integrated with Sahha, the data appears alongside existing sleep metrics with no code changes needed.