The Resolution Effect: What Our Data Reveals About Activity Upticks

Did you make a New Year’s resolution? If so, have you maintained it?

I, Alistair Brownlee, made two health-based resolutions. Firstly, to sleep for 24 hours for each three-night period. I’ve already failed at that one. And secondly, to do some form of exercise every day, which thankfully, is still intact.

The classic "New Year's resolution effect" – that surge of motivation on January 1st followed by a quick fade – is something we’ve all experienced or observed. But does the data support it? Using aggregated insights from a large cohort of users tracked through wearable devices and platforms like Terra API, we wanted to assess whether we could detect a noticeable change in the data.

Our analysis focuses on 2025 data, comparing activity metrics on January 1st, and the days following, to other days in our January - March data set, while controlling for day-of-week effects. We examined steps, active minutes, and distance, applying rigorous statistical methods to separate real behavioral changes from noise. We found a modest, short-lived uptick, strongest among certain users.

We should say that, as we studied anonymised users, we don’t know if any of them actually made a New Year’s Resolution to increase their activity. It would, of course, be more interesting to study how those who made a resolution changed their behavior. Instead, we have a large data set that allows us to determine whether the population as a whole changed their routines and to see the percentage of individuals who demonstrated a different pattern. Then determine what was unique about these pattern changes to see if there is anything interesting we can learn.

The January 1st Boost: Steps Show the Strongest Spike

When we compare January 1st directly to other days across the year, steps stand out:

• January 1st mean: 6,452 steps
• Other days mean: 5,694 steps
• Difference: +13.3% (statistically significant, p=0.0008)

This 13.3% increase in steps is the most pronounced signal of New Year's motivation. People clearly took extra walks, hit the treadmill, or got moving in ways that added up to more total steps.

Active minutes and distance showed smaller lifts:

• Active Minutes: +1.3% (not significant, p=0.74)
• Distance: +2.6% (not significant, p=0.65)

These differences are positive but fall short of statistical significance, suggesting that while more people were moving, the extra effort didn’t always translate into longer sessions or greater overall distance on that single day. This is obviously not proof of a New Year’s resolution effect, as January 1st is a holiday in most Western countries. So we would expect to find a change in behaviour. This data probably just confirms that New Year’s Day walks are a thing!

Day-of-Week Aware Analysis

To avoid bias from weekday/weekend patterns, we benchmarked for day of the week and weekends. For each user, we calculated their personal median activity (January through March 31) and flagged days that exceeded this baseline by 5% or more.

On January 1st:

• Steps: 50.9% of users (653 out of 1,282) were above threshold
• Active Minutes: 47.6% of users (1,550 out of 3,259)
• Distance: 47.2% of users (1,217 out of 2,579)

Roughly half of users showed elevated activity on New Year's Day itself. But how long did it last? The duration of these elevated periods was short for most:

• Mean duration: ~2.7–2.8 days across metrics
• Median duration: 2.0 days
• Maximum duration: 21–31 days

Most users who showed an increase in activity on January 1st (1 day: 40–47%, 2 days: 19–20%) returned to their normal levels after 1 or 2 days. About one in four maintained higher activity for at least 4 days, indicating that, for some, there is a real, perceptible trend difference.

How did people increase their activity?

The standout pattern is that Quantity Only was the most common way people boosted activity (25–35% of users), while adding more sessions without increasing intensity (Frequency Only) was the least common (5–9% of users). A solid portion (16–22%) did both, but nearly half of users (41–47%, rightmost tall bars) showed no meaningful change in either dimension.

This tells a practical story about changes in New Year's activity: most people preferred the easier path of extending or intensifying their existing workouts (e.g., a longer walk or a harder gym session) rather than adding brand-new sessions to their schedules. It suggests that time constraints or habit inertia play a big role; going longer and harder in what you're already doing feels more doable than carving out extra time slots. For those building sustainable habits, this data hints that focusing on quality/intensity upgrades might be a smarter starting point than forcing more frequent workouts right away. We are back at a principle we see time and again in this research. Small and consistent wins when it comes to health.

The population vs the individual

When analysing data like this, we try to distinguish genuine behavioural shifts - like a true New Year's motivation boost - from statistical flukes or random daily variations. We did this in two main ways. Firstly, by population-level significance (did the overall group trends hold up across the board?). Then we looked at individual-level significance (did each person's January 1st spike stand out against their own typical patterns?)

The population-level test has much greater statistical power because it averages thousands of users together, making even small, widespread increases (like an extra 700–800 steps across many people) highly detectable. In contrast, individual tests compare just one day against each person's own baseline, where day-to-day noise is much larger and makes it harder for most personal increases to reach statistical significance. This tells us the New Year's effect at a population level was real but subtle

At the person-by-person level, we ran statistical tests to determine whether a user's January 1st activity was significantly higher than their personal baseline (p < 0.05). This filters out one-off anomalies.

• Steps: 44.1% of users showed a statistically significant increase (566 out of 1,282 total users with data).
• Active Minutes: 45.4% (1,479 out of 3,259 users).
• Distance: 42.2% (1,089 out of 2,579 users).

In short, about 42–45% of users had a "real" elevation on New Year's Day, not just luck or a random good day. This is a strong hint that resolutions did drive action for nearly half the group.

Statistical significance tells us if something happened, but Cohen's d measures how much it mattered, like the practical size of the change. Our effect sizes were consistently small, which fits the idea of a subtle "nudge" rather than a seismic shift in habits.

• Steps: -0.009 (negligible — basically no meaningful group difference after accounting for variance).
• Active Minutes: 0.038 (small).
• Distance: 0.123 (small).

These tiny numbers suggest the resolution effect was widespread but gentle: think a collective extra stroll or slightly longer jog, not everyone suddenly training for a marathon.

Who Increases? Less Active Users Lead the Charge

We found that the increase in activity is not uniform. Less-active users, those in the lowest baseline quartiles, were more likely to increase their steps. The lowest quartile had a 20% increase rate; the highest quartile only 2%. Comparing the least active quartile to the most active quartile:

Active Minutes: 63.6% vs. 29.9%

Distance: 46.4% vs. 39.5%

Less-active individuals showed the largest behavioural shift. Already-active users tended to maintain rather than amplify their routines. There could be a regression to the mean effect here - it’s more likely that previously low activity users can, and do, increase their activity levels.

Interestingly, this trend almost reversed for distance. Where the most and least active quartiles have a similar (4-5%) increase.

What’s happening here? Our theory is that people walk more on January 1st, but don’t necessarily engage in more intense activities. But we are guessing!

Is It Just Regression to the Mean? The Evidence Says No

Regression to the mean would show up among the lowest-activity users, increasing activity purely by chance, and a high-activity day might follow a low one purely by chance. The results showed that those in the lowest quartile had a high coefficient of variation in the number of steps (much higher than the Q4 activity users), suggesting a regression-to-the-mean effect on this analysis. But our tests show the January 1st effect is real. Increase rates on January 1st were 15–25 percentage points higher than on random Wednesdays (p<0.001). The effect was specific to January 1st, not a random fluctuation. While variance plays a role, the data confirms a genuine behavioural response tied to the New Year's "fresh start" mindset.

What This Means

Our analysis shows a New Year's resolution effect on activity: about half of users increased activity on January 1st, with the strongest signal in steps and among less-active individuals. Yet for most, the boost lasted only days, not weeks. This aligns with broader patterns where enthusiasm peaks early but fades quickly.

The big unknown here is how many people in our sample actually set a Resolution to increase their activity. It could have been every user, none, or more likely somewhere in the middle! So we don’t know how many of those who set an exercise resolution stuck to them.

The good news is that the 24–27% who sustained elevated activity for 4+ days prove persistence happens for many. In the future, it would be great to test if interventions (apps, gym membership, etc.) can help with these behaviour trends, by both behavioural nudges and by providing deeper insights – tracking strain scores, spotting early burnout. These are the kinds of insights that could really help users and Terra customers create meaningful health impact.

Whether we were part of the January spike or watching from the sidelines, the numbers remind us: every day is an opportunity to move a little more. Remember, this analysis was set up to determine whether we could see different patterns in activity on January 1st and the days that followed. But we can all start a new activity, “trend”, or resolution in any of the other 365 days of the year!