I had been doing one-month tracking experiments for years. They produced specific insights about specific interventions. What I had not done was sustained tracking long enough to see the patterns of what produces my baseline state.
For 6 months I tracked daily: energy 1-5, mood 1-5, stress 1-5, plus what I had done that day in several categories (sleep, exercise, meditation, social interaction, alcohol, work intensity, screen time).
Here is what the data showed about what actually drives my baseline wellness.
The setup
Tracked daily, end of day, in a notebook:
- Energy 1-5
- Mood 1-5
- Stress 1-5
- Hours of sleep previous night
- Exercise yes/no and intensity
- Meditation yes/no
- Social interaction (none/light/moderate/heavy)
- Alcohol yes/no
- Work intensity 1-5
- Phone screen time hours
Took about 2 minutes per day. Sustained for 184 days (missed maybe 6 days during travel).
The strongest correlations with mood
After 6 months, the data showed clear patterns. Strongest correlations with high mood days:
1. Sleep duration the previous night. Single strongest predictor. Days following 7+ hours averaged mood 4.0; days following less than 6.5 hours averaged 3.0. A full point difference.
2. Exercise that day. Days with exercise averaged 0.8 points higher than days without.
3. Meaningful social interaction. Not just any social interaction — meaningful interaction (close friends, deep conversations) averaged 0.6 points higher than days with only superficial interaction.
4. Time outside. Days with substantial outdoor time averaged 0.5 points higher.
5. Meditation. Smaller effect than expected — about 0.2 points higher on meditation days. The cumulative effect across many days was larger than the daily effect.
The strongest correlations with stress
For stress, the patterns were partly the inverse but with some surprises:
1. Sleep again. Sleep deprivation predicted stress as strongly as it predicted low mood.
2. Work intensity. Strong correlation. High-intensity work days had average stress 1.0+ points higher.
3. Phone screen time. Surprising correlation. Days with 4+ hours of phone time had higher stress, even controlling for work.
4. Alcohol the day before. Real effect. Day after drinking averaged stress 0.5 points higher.
5. Lack of solitude. Days with no solo quiet time had higher stress regardless of other factors.
The surprising weak correlations
Several factors I expected to matter showed weaker effects than predicted:
Diet quality. I had been tracking what I ate qualitatively. The correlation with mood and stress was weak. Either diet matters less than I thought, or the effects are too lagged for daily tracking to capture.
Caffeine intake. Almost no correlation with energy. Either it works through other mechanisms or my baseline tolerance has equalized the effect.
Specific exercise type. Exercise mattered, but specific type (cardio vs strength vs yoga) did not significantly differ in mood/stress effects. Movement itself seemed to matter more than the type.
Time of day for activities. I had assumed morning workouts would be better than evening. The data did not confirm this.
The seasonal patterns
Six months covered late summer through early spring, which let me see seasonal effects:
- Mood ticked down in November-January (about 0.5 points lower than peak)
- Energy similarly affected
- Stress slightly higher in winter
- The seasonal effect was real but smaller than expected
- Morning sunlight in winter substantially reduced the effect on days when I got it
The day-of-week patterns
- Mondays averaged about 0.4 points lower in mood than other weekdays
- Sundays were higher than other days but with elevated stress in the evening
- Friday-Saturday were the highest-mood days
- Tuesday-Thursday were essentially indistinguishable
The "Sunday Scaries" effect was real and small. The "weekend high" was larger than I had appreciated.
What the data revealed about my own habits
The most consequential insight was about how often I did the things that the data showed mattered most:
- Sleep 7+ hours: only 58% of nights
- Exercise: 4 days per week average
- Meaningful social interaction: 2-3 days per week average
- Time outside (>30 min): 3-4 days per week average
- Meditation: about 5 days per week
The things that mattered most were the things I was least consistent about. The corrective implication was clear.
What I changed because of the data
Three changes based on what the data revealed:
1. Sleep became non-negotiable. Treating 7+ hours as a hard target rather than aspirational. Implemented harder phone-off-by-9 PM rule.
2. Daily outdoor time. The data made the cost of skipping it visible. Implemented daily walk regardless of weather.
3. Phone use restriction. The phone-stress correlation was strong enough to justify substantial reduction. Removed social media apps.
What I learned about tracking itself
1. 6 months is the right minimum. Patterns at one month were noise. Patterns at three months were directional. Patterns at six months were reliable.
2. Multiple variables tracked together reveal interactions. Single-variable tracking misses patterns that emerge across variables.
3. The act of tracking changes behavior. Knowing I would be writing things down sometimes nudged me toward the behaviors I knew I would record positively.
4. The data will tell you something uncomfortable. For me it was the alcohol-day-after correlation. Yours will be different but there will be one.
What I learned about myself
The macro insight: my wellness is more dependent on a small number of fundamentals (sleep, movement, time outside, meaningful social interaction) than on the elaborate optimization I had been pursuing. Most of the wellness optimization industry sells variations on these fundamentals as if they were novel.
The fundamentals are dull. They are also disproportionately effective.
This is not advice
Your data will be different. Your patterns will be different. The point of writing this is the methodology — track over time, look at the patterns, change based on what your specific data shows.
Most wellness advice is generic because it is written for anyone. Your wellness data is specific to you. The patterns you find will inform decisions better than any general advice.
The takeaway
Sustained self-tracking reveals the actual drivers of your baseline wellness. The drivers are usually fewer and more fundamental than wellness culture suggests.
If you want to actually improve your wellness rather than just consume content about wellness, track yourself for at least 3 months. Look at what the data says actually matters. Make changes based on that. The compounding effect over a year is substantial.
