Behavioral Archetypes

Archetypes categorize users based on their behavioral patterns, lifestyle, and wellness states. These classifications help understand long-term trends rather than short-term fluctuations. Unlike numerical health scores, archetypes provide an intuitive, high-level interpretation of an individual's health. Each archetype is derived from multiple digital biomarkers and health scores, aggregated over weeks or months. Archetypes enable personalized UX, tailored recommendations, and effective cohort-based analysis.

Key Features

How It Works

Health data is gathered from wearables, smartphone sensors, and health scores over time. The data is analyzed over weeks or months, smoothing out short-term fluctuations. Based on defined thresholds, users are assigned a qualitative label for each archetype. Archetypes are refreshed weekly and monthly, allowing tracking of behavioral shifts over time. There are two types of archetypes:

Ordinal Archetypes (Ranked Progression)

Ordinal archetypes represent a ranked scale, where values move from lower to higher states in a meaningful order. Each level in the archetype reflects an increase in the measured behavior.

• Example: sleep_duration
  • very_short_sleeper → short_sleeper → average_sleeper → long_sleeper
  • What it means: As users get more sleep, they progress toward higher sleep duration categories.
• Why This Matters:
  • Enables comparisons between different user groups and population benchmarks.
  • Helps identify trends and improvements over time.

Categorical Archetypes (Distinct Groups)

Categorical archetypes group users into distinct categories without any hierarchical ranking. Unlike ordinal archetypes, these categories are independent of each other and do not imply a progression.

• Example: sleep_chronotype
  • Categories: early_bird, night_owl, intermediate
  • What it means: These categories describe natural sleep-wake tendencies but are not ordered from "worse" to "better" or "less" to "more".
• Why This Matters:
  • Useful for segmentation and personalization (e.g., optimizing app notifications based on chronotype).
  • Provides insights into natural inclinations rather than performance-based measures.

By having both ordinal and categorical archetypes, Sahha provides a comprehensive view of measurable behaviors (e.g., activity level, sleep efficiency) to inherent traits (e.g., chronotype). This approach enables more actionable information for a variety of use cases.

List of Archetypes

Output Schema

Archetypes are delivered as a JSON object with the following fields:

{ 
    "id": "91ced284-5355-57f0-b162-1ac920a42371",
    "profileId": "6be989eb-813c-4380-be85-a6a7d787da70",
    "accountId": "a17fd912-46a2-48aa-be3e-1146ee2cd258",
    "externalId": "edd9afa-7012-4c30-8121-53fc3a9be461",
    "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" 
}

FAQs

Q: How are Archetypes different from Health Scores?
A: Health Scores provide numerical values (e.g., 0.0-1.0), while Archetypes classify users into predefined categories. Archetypes offer a more intuitive interpretation of long-term behavior.

Q: How does ordinality work?
A: Ordinality represents the ranking of each category, with lower values indicating less value alongt that dimension and higher values representing more value.

• Example for sleep_duration:
  • 0 = very_short_sleeper
  • 1 = short_sleeper
  • 2 = average_sleeper
  • 3 = long_sleeper

Q: How frequently are Archetypes updated?
A: Archetypes are generated at the end of each week, month, or quarter, depending on the appropriate periodicity.

Getting Started

To integrate Archetypes into your product, subscribe to the archetypes Webhook, or use the Sahha API to fetch the archetype assignments for profiles.

Support

For additional assistance or more detailed information on Archetypes, please contact support@sahha.ai or reach out in the slack community .