The Fresh Start Effect: Why Mondays and January Actually Work

Mondays and January do work, in a narrow and measurable sense: dates that feel like the start of something reliably increase the rate at which people begin pursuing a goal. What they do not do is help you continue. That asymmetry is the entire story, and it explains both the January gym queue and the empty gym in February using the same mechanism.

The study that named it

The term comes from Hengchen Dai, Katherine Milkman and Jason Riis, then all at Wharton, in a 2014 paper in Management Science (volume 60, issue 10, pages 2563 to 2582) titled "The Fresh Start Effect: Temporal Landmarks Motivate Aspirational Behavior." They looked at three unrelated archives of real behavior rather than running a laboratory study.

The first was Google search volume for the word "diet." The second was gym attendance records from a university. The third was commitments made on a goal-commitment website, where people formally pledge to pursue a specific outcome.

All three moved in the same direction after the same kinds of dates: the start of a new week, a new month, a new year, a new semester, and after a birthday or a holiday. The landmarks were not equally powerful, and the calendar-scale ones did more than the weekly ones, but the pattern held across three completely different measures of aspiration.

That convergence matters more than any single number. Search volume, gym turnstiles and commitment contracts have almost nothing in common as data sources, and they are not vulnerable to the same confounds. When they all bend at the same points on the calendar, something is happening on the calendar.

The mechanism: a mental accounting break

Dai, Milkman and Riis argue that a temporal landmark opens a new mental accounting period. The past is filed under a previous period, which puts some psychological distance between the person you were and the person deciding right now. Their phrasing is that landmarks "relegate past imperfections to a previous period" and prompt a big-picture view.

The consequence is that a landmark makes the current self feel discontinuous with the record. Three failed attempts belong to last year. This year has no attempts in it yet. That is not a rational update, since nothing about your circumstances changed at midnight, but it is a useful one, because a track record of failure is one of the more effective deterrents to trying again.

This connects to a broader finding about how self-concept drives action, covered in the research on identity-based motivation: people do what feels congruent with who they currently take themselves to be. A landmark gives you temporary permission to take yourself to be someone else.

Habit discontinuity: the version with a stronger causal story

There is a related literature that gets less attention and arguably has more practical weight. Bas Verplanken and colleagues at the University of Bath have argued that habits are most changeable during genuine context disruption, not symbolic disruption.

Their most cited demonstration, published in the Journal of Environmental Psychology in 2008 (volume 28, issue 2, pages 121 to 127), looked at university employees and their commuting choices. Employees who had recently moved house and held strong environmental values drove to work less than equally environmentally concerned colleagues who had not moved. Neither factor alone did it. The move without the values changed nothing, and the values without the move changed nothing.

That interaction is the interesting part. A context change does not push behavior in a good direction by itself. It suspends the automatic response, which opens a window in which whatever you actually value gets a vote. Verplanken and Deborah Roy later tested this directly in a field experiment, also in the Journal of Environmental Psychology, delivering an intervention to recent movers and non-movers and finding the discontinuity effect at the aggregate level.

January is a symbolic landmark. Moving house is a structural one. If you have both available, the structural one is doing more work, because it actually removes the cues that were running the old behavior. That is the same argument developed at length in the case for reducing friction rather than raising motivation.

5 temporal landmarks, ranked by how much they give you

1. An actual move or life transition. New house, new job, new city, new schedule. This is the only item on the list that changes the physical cues, which is why it is the only one with a mechanism beyond framing.
2. The start of a year. The largest of the symbolic landmarks in the Dai data, and the one with the most cultural scaffolding behind it, which is not nothing: everyone around you is also treating it as a boundary.
3. A birthday or the start of a semester. Personally salient, which is what appears to matter. A landmark works to the extent that it feels like a boundary to you specifically.
4. The start of a month. Real in the data, smaller than the year.
5. Monday. The weakest of the effects that showed up, and also the most frequently available, which is the tradeoff. Fifty-two chances at a small effect is a reasonable deal.

The honest accounting

Three limits belong on this finding, and the third is the one that matters.

First, the original studies are archival and correlational. They document that behavior spikes after landmarks. They do not randomly assign people to landmarks, because nobody can. The mental accounting explanation is a reasonable interpretation of the pattern rather than something the field data demonstrate directly.

Second, the effects are modest. These are shifts in rates, not transformations. A measurable increase in the number of people who start something is worth having and is not the same as a measurable increase in the number who finish.

Third, and most importantly: everything the fresh start effect measures is initiation. Nothing in it addresses persistence, and persistence is where behavior change actually fails. For a sense of the gap, consider the largest field study of gym attendance ever run, a 2021 megastudy of 61,293 members of a national gym chain published in Nature by Milkman, Angela Duckworth and dozens of collaborators. Roughly 45 percent of the 53 tested interventions raised weekly gym visits during the four-week program. Only about 8 percent produced effects that were still statistically detectable afterward.

That is the February problem stated precisely. Starting is cheap and can be nudged. Continuing is expensive and mostly cannot.

There is also a darker reading of the mental accounting mechanism, and it follows from the same logic. If a landmark works by separating the current period from the last one, then a lapse inside the current period is not a neutral event. It contaminates the new account. The clean slate that made starting feel possible is the same structure that makes a single missed day feel like the slate is dirty, and the natural response is to wait for the next landmark. That is why the resolution architecture produces both the January spike and the February collapse.

The practical implication is not to ignore landmarks. It is to use them for what they demonstrably do, which is lower the cost of starting, and then to hand the behavior over to something that does not depend on the calendar at all. The research on building a morning ritual that survives low-motivation days is about exactly that handoff: anchoring to an existing behavior so that continuation does not require a reason.