What are Archetypes and how do they work

Archetypes turn weeks of health and lifestyle signals into stable, human-readable labels for segmentation and personalization. Learn how Sahha archetypes work, the science behind the behaviors they represent, and how to use them in your product.

Archetypes convert weeks of health, lifestyle, and behavioral data into human-readable labels (e.g., night_owl, short_sleeper, highly_active). Unlike daily scores, archetypes are designed to be stable traits you can use for segmentation, targeting, personalization, and analytics — without overreacting to one-off days like travel, illness, or deadlines.

Key Takeaways

What they are: long-term behavioral labels derived from weeks of data.
Why they matter: archetypes are stable, segmentable, and easy to use in product logic.
What makes them credible: archetypes summarize validated behavioral dimensions (sleep, activity, sedentary behavior, recovery patterns) that have strong scientific links to health and wellbeing.
Best practice: treat archetypes as “profile traits,” and use daily scores/biomarkers for real-time nudges.

Metric Spec

Item	Value
Output type	Archetype assignment (label + metadata)
Typical cadence	Weekly and monthly refresh (some schemas may support longer cycles)
Best used for	Segmentation, targeting, personalization, analytics/BI
Data requirements	Varies by archetype; most are smartphone compatible
Delivery methods	API (pull) or Webhooks (push)
Retroactive support	New integrations can receive recent historical archetypes depending on availability

How Archetypes Work

Sahha analyzes weeks of data and assigns users to behavioral categories. Compared to day-level metrics, this “smoothing” reveals stable patterns and reduces noise.

There are two archetype types:

Ordinal archetypes (ranked progression)

Ordinal archetypes represent a ranked scale where categories move from lower → higher states in a meaningful order.

Example concept:

sleep_duration: very_short_sleeper → short_sleeper → average_sleeper → long_sleeper

These are useful for:

cohort comparisons
“improving vs declining” narratives
simple threshold-based journeys

Categorical archetypes (distinct groups)

Categorical archetypes group users into distinct categories without implying “better” or “worse”.

Examples in Sahha include:

primary_exercise_type (e.g., strength_oriented, cardio_oriented)
sleep_pattern (distinct timing/consistency patterns)

These are useful for:

content routing
preferences and persona-style segmentation
personalization that shouldn’t be framed as a performance ranking

The Science Behind Archetypes

Archetypes are credible when they represent constructs that are (1) measurable from passive data, (2) behaviorally meaningful over time, and (3) supported by evidence linking them to outcomes.

Sahha archetypes draw their meaning from the same evidence-backed dimensions used in Sahha’s score models:

1) Sleep archetypes map to validated sleep health dimensions

Modern sleep science treats sleep health as multi-dimensional. Beyond duration, dimensions like regularity, continuity, circadian alignment, sleep debt, and restorative stages each independently relate to health and functioning. Sahha’s Sleep Score science outlines these dimensions and why they matter, and archetypes summarise these same dimensions as stable patterns over weeks.

How this builds credibility: archetypes like sleep_duration, sleep_regularity, sleep_quality, bed_schedule, and wake_schedule are long-term labels for sleep timing and consistency — the same dimensions used in the Sleep Score scientific model.

2) Activity archetypes map to validated dimensions of physical activity behavior

Physical activity cannot be captured by a single metric like step count. Evidence supports multiple independent dimensions: volume, intensity, frequency, energy expenditure, and sedentary behavior. Sahha’s Activity Score science lays out this multi-dimensional model and its links to outcomes.

How this builds credibility: archetypes like activity_level and exercise_frequency summarize movement behavior over time using these validated dimensions.

3) Mental and overall wellness archetypes reflect behavioral signals linked to wellbeing trends

Behavioral patterns in sleep regularity, circadian alignment, activity, and sedentary behavior are each associated with emotional balance and resilience. Sahha’s Mental Wellbeing Score science describes a model built from these behavioral dimensions (steps, active hours, extended inactivity, activity regularity, sleep regularity, circadian alignment). Archetypes provide stable “trait-like” versions of these behavioral patterns for segmentation.

Similarly, Sahha’s Wellbeing Score science describes a holistic approach combining sleep and activity factors. Archetypes like overall_wellness provide a stable label aligned to holistic, multi-factor wellbeing.

4) Exercise preference archetypes have direct internal validation for recommendation use

Sahha archetypes include primary_exercise, secondary_exercise, and primary_exercise_type. Sahha’s research on archetype-based class matching shows statistically significant evidence that a user’s preferred sports (captured in archetypes) can help predict other sports they may engage with — supporting recommendation experiences built on these archetypes.

5) Scientific references for compliance and transparency

Sahha provides dedicated scientific reference pages for scores (often used for app store submission transparency). These are useful to link in your product for credibility when you explain “why” health signals matter.

Science Links by Archetype Family

Archetype family	What it summarizes	Closest Sahha science references
Activity	Volume, frequency, intensity, energy expenditure, sedentary behavior	Activity Score science; Activity Score scientific reference
Sleep	Duration, timing, regularity, quality/continuity patterns	Sleep Score science; Sleep Score scientific reference
Mental wellness	Behavioral signals linked to emotional balance/resilience (sleep + activity rhythms)	Mental Wellbeing Score science
Overall wellness	Combined sleep + activity lifestyle patterns	Wellbeing Score science
Exercise preference	Stable preference signals from exercise logs	Archetype-based class recommendation research; Exercise types taxonomy

List of Archetypes

Below is the current set of archetypes and their possible values.

Archetype	Type	Possible Values	Periodicity	Description	Requires Wearable
`activity_level`	Ordinal	`sedentary`, `lightly_active`, `moderately_active`, `highly_active`	Weekly, Monthly	Overall level of physical activity including movement and exercise	No
`exercise_frequency`	Ordinal	`rare_exerciser`, `occasional_exerciser`, `regular_exerciser`, `frequent_exerciser`	Weekly, Monthly	How often the individual exercises	No
`mental_wellness`	Ordinal	`poor_mental_wellness`, `fair_mental_wellness`, `good_mental_wellness`, `optimal_mental_wellness`	Weekly, Monthly	Mental wellness and resiliency based on physical activity, sleep, and stress indicators	No
`overall_wellness`	Ordinal	`poor_wellness`, `fair_wellness`, `good_wellness`, `optimal_wellness`	Weekly, Monthly	Overall wellbeing across all aspects of health	No
`primary_exercise`	Categorical	Most frequent exercise (e.g., running, weightlifting, yoga)	Weekly, Monthly	Most commonly performed exercise	No
`primary_exercise_type`	Categorical	`strength_oriented`, `cardio_oriented`, `mind_body_oriented`, `hybrid_oriented`, `sport_oriented`, `outdoor_oriented`	Weekly, Monthly	Categorizes the primary exercise into strength, cardio, sports, etc.	No
`secondary_exercise`	Categorical	Second most frequent exercise (e.g., swimming, cycling, hiking)	Weekly, Monthly	Second most commonly performed exercise	No
`sleep_duration`	Ordinal	`very_short_sleeper`, `short_sleeper`, `average_sleeper`, `long_sleeper`	Weekly, Monthly	Typical sleep duration relative to recommended norms	No
`sleep_efficiency`	Ordinal	`highly_inefficient_sleeper`, `inefficient_sleeper`, `efficient_sleeper`, `highly_efficient_sleeper`	Weekly, Monthly	How effectively the individual maintains uninterrupted sleep	Yes
`sleep_pattern`	Categorical	`consistent_early_riser`, `inconsistent_early_riser`, `consistent_late_sleeper`, `inconsistent_late_sleeper`, `early_morning_sleeper`, `chronic_short_sleeper`, `inconsistent_short_sleeper`	Weekly, Monthly	Overall sleep behavior based on timing and consistency	No
`sleep_quality`	Ordinal	`poor_sleep_quality`, `fair_sleep_quality`, `good_sleep_quality`, `optimal_sleep_quality`	Weekly, Monthly	Long-term quality of sleep based on duration, regularity, recovery, and debt	No
`sleep_regularity`	Ordinal	`highly_irregular_sleeper`, `irregular_sleeper`, `regular_sleeper`, `highly_regular_sleeper`	Weekly, Monthly	Consistency in sleep timings	No
`bed_schedule`	Ordinal	`very_early_sleeper`, `early_sleeper`, `late_sleeper`, `very_late_sleeper`	Weekly, Monthly	Typical bedtime	No
`wake_schedule`	Ordinal	`very_early_riser`, `early_riser`, `late_riser`, `very_late_riser`	Weekly, Monthly	Typical wake-up time	No

Output Schema

Archetypes are delivered as objects with stable metadata. Typical fields include:

id: unique identifier for the archetype assignment
profileId, accountId, externalId: identifiers for joining to your user model
name: archetype name (e.g., sleep_duration)
value: assigned category (e.g., short_sleeper)
dataType: ordinal or categorical
ordinality: numeric rank for ordinal archetypes
periodicity: e.g., weekly, monthly (and where supported, longer cycles)
startDateTime, endDateTime: time window the archetype represents
createdAtUtc: when the archetype was generated

Example response:

{
  "id": "91ced284-5355-57f0-b162-1ac920a42371",
  "name": "sleep_duration",
  "value": "short_sleeper",
  "dataType": "ordinal",
  "ordinality": 1,
  "periodicity": "monthly",
  "startDateTime": "2025-01-01T00:00:00+13:00",
  "endDateTime": "2025-01-31T00:00:00+13:00",
  "createdAtUtc": "2025-02-01T13:08:53.322886Z"
}

How to Use Archetypes in Your Product

1) Segmentation and targeting

Use archetypes as audience traits:

activity_level = sedentary → onboarding focused on small movement wins
sleep_regularity = highly_irregular_sleeper → routine builder journey
primary_exercise_type = strength_oriented → strength-oriented plans and recovery content

2) Personalization rules (deterministic and explainable)

Examples:

If wake_schedule = very_late_riser, avoid early-morning notifications by default
If sleep_pattern = inconsistent_late_sleeper, prioritize schedule-stabilization content
If exercise_frequency = rare_exerciser, start with 2–3 days/week goals (not daily streaks)

3) Recommendations and content matching

primary_exercise and secondary_exercise support “more like this” content feeds.
primary_exercise_type supports stable preference routing when exact exercise types are noisy.

4) Analytics and BI

Archetypes provide clean cohort filters:

retention by overall_wellness
conversion by exercise_frequency
engagement by sleep_quality

Implementation Suggestions for your Products

Pick the right role for archetypes
- Use archetypes for journeys and segmentation.
- Use scores/biomarkers for day-level decisions and nudges.
Store archetypes as user attributes
- Persist the latest value per name plus its periodicity and time window.
- Keep createdAtUtc and start/endDateTime for auditability.
Design for change over time
- Archetypes can shift as behavior changes.
- Avoid “locking” users permanently; re-route journeys when archetypes change.
Handle missing data and onboarding gaps
- New users may not have enough history.
- If missing, fall back to simpler signals (steps, sleep duration, onboarding questions).
Be transparent about uncertainty
- Use language like “based on recent patterns” and “compared to your typical behavior.”
- Avoid diagnostic framing.

FAQ

Are archetypes medical diagnoses?

No. Archetypes describe behavioral patterns inferred from passive data over time. They are designed for personalization and analytics, not diagnosis.

Why do archetypes feel “more accurate” than daily metrics?

Because archetypes use weeks of data and smooth out day-to-day noise. This reduces overreaction to single events (travel, illness, late nights).

Do archetypes require wearables?

Most do not. Some sleep-related archetypes (e.g., sleep_efficiency) require wearable-grade signals.

How should I justify archetypes scientifically in my product?

Link to the Sahha score science pages that describe the validated behavioral dimensions underlying the archetypes (sleep health dimensions, activity behavior dimensions, and behavioral signals linked to mental wellbeing). For compliance transparency, include the score scientific reference links where relevant.

Scores (daily, explainable): Activity, Sleep, Mental Wellbeing, Readiness, Wellbeing
Biomarkers (daily aggregates): steps, active duration, sleep duration, etc.
Insights (trend and comparison analytics)

Notes

This content is educational and intended for product personalization and engagement. It is not medical advice and should not be used to diagnose health conditions.

References

Sahha (Research)

Class recommendation with archetypes (archetype-based sport similarity / recommendations)
https://resources.sahha.ai/resources/research/sport-class-recommendation/