The Science of Relationship Health Scores: How AI Measures Connection Strength
Relationship health isn't just a feeling — it's measurable. Decades of research in social psychology, communication science, and evolutionary biology have identified the behavioral signals that predict whether a relationship will thrive or fade. Yenesow's scoring model is built on this research. Here's how it works.
What Is a Relationship Health Score?
A relationship health score is a number from 0 to 100 that represents the current strength of your connection with someone. Green (70-100) means the relationship is healthy. Yellow (40-69) means it needs attention. Red (0-39) means you're at risk of losing touch.
It is not a measure of how much you care — it is a measurement of behavioral signals that research shows predict relationship quality and longevity.
Example: A Real Relationship Map
The Four Factors Behind the Score
The Research Foundation
Connection Goals: The Strongest Signal
Oswald, Clark, and Kelly (2004) in Communication Research showed that regular, brief interactions maintain friendships more effectively than infrequent, lengthy ones. A 5-minute call every week does more than a 3-hour dinner every 6 months.
In Tune Up, you set how often you'd like to connect with each person. The AI measures whether you're meeting those goals — not by tracking your messages, but by the context you share in-app and when you mark outreach as done.
The Friendship Formation Timeline
Hall (2018) quantified friendship formation: roughly 50 hours for a casual friend, 90 for a friend, 200+ for a close friend. This informs our decay model — newer relationships need more frequent contact, while established ones are more resilient.
That is why Mom stays at 92 even if you skip a week, while a new professional contact drops faster. The relationship has more stored capital.
Dunbar's Layers
Robin Dunbar's research (2010) identified concentric circles: 5 intimate contacts, 15 close, 50 good friends, 150 acquaintances. Each layer requires different maintenance. Our model adapts — we do not expect you to call a professional contact as often as your sister.
Responsiveness: The Quality Multiplier
Reis and Shaver's (1988) intimacy process model shows that perceived responsiveness — the feeling that someone notices and cares — is the foundation of deep relationships. Missing a birthday measurably hurts. Remembering it has an outsized positive effect.
What the Score Does Not Measure
Emotional depth. We do not read messages, so we cannot assess conversation quality.
In-person interactions. If you see someone daily but never text, our score will not capture that.
Relationship satisfaction. A high-frequency relationship could still be stressful.
Reciprocity. Currently we measure your outreach, not whether the other person reciprocates.
A relationship health score is like a step counter for your social life. It does not tell you everything about your fitness, but it tells you when you have been sitting too long.
See your relationship health scores
Yenesow generates AI-powered health scores for every connection. See which relationships are thriving and which need attention.
Download Yenesow FreeSources and References
- Oswald, D. L., Clark, E. M., and Kelly, C. M. (2004). Friendship maintenance behaviors. Journal of Social and Clinical Psychology, 23(3), 413-441.
- Hall, J. A. (2018). How many hours does it take to make a friend? Journal of Social and Personal Relationships, 36(4), 1278-1296.
- Dunbar, R. I. M. (2010). The social brain hypothesis. Annals of Human Biology, 36(5), 562-572.
- Reis, H. T. and Shaver, P. (1988). Intimacy as an interpersonal process. In Handbook of Personal Relationships. Wiley.
- Sprecher, S., et al. (2013). Effects of self-disclosure on liking and closeness. Journal of Social and Personal Relationships, 30(4), 497-514.
- Holt-Lunstad, J., et al. (2010). Social relationships and mortality risk. PLoS Medicine, 7(7), e1000316.