Ask how to turn an ending into a beginning and you will mostly be handed a ceremony. Clear the desk. Write the letter you will not send. Decide that the next chapter starts now. The advice is warm and it is untestable, which is part of why it never goes out of print.
There is, though, a measured effect sitting underneath all of it. It is larger than you would guess, and considerably narrower.
People really do start things at boundaries
Hengchen Dai, Katherine Milkman and Jason Riis went looking for this in three sets of records that nobody had created for the purpose, which is the useful thing about them. Their 2014 paper counted Google searches, gym visits and signed commitments, and found the same shape in all three.
The Artful Age
A weekly letter on aging well, family across generations, and the creative life after the kids leave home.
Across 3,104 days of American search data, searches for the word diet rose 14.4% after the start of a new week, 10.2% after a federal holiday, and 82.1% after the new year, relative to baseline.
The gym records are the part that counts, because a search costs nothing and a trip to the gym costs something. At a university where every undergraduate holds an automatic membership, across 442 days, the probability of a student visiting the gym rose 33.4% at the start of a new week, 14.4% at the start of a month, 24.3% after a school break and 47.1% at the start of a semester. For the students whose birthdates the researchers had, it rose 7.5% after a birthday.
And on stickK, a site where people sign binding commitments to their own goals, new contracts rose 62.9% at the start of a week, 55.1% after a federal holiday and 145.3% at the start of a year.
What the boundary seems to be doing
The explanation the authors propose is bookkeeping rather than inspiration. A boundary closes an accounting period. The version of you that failed at this belongs to the period that just ended, filed and shut, and the person starting Monday is technically a different person with a clean ledger.
Two details make this more persuasive than a tidy theory usually is. Searches for gardening, which is something people also put off but do not aspire to in the same way, showed no such rise after the same landmarks. And the pattern broke in exactly one place: students turning 21 went to the gym less afterwards, not more, which the authors report rather than bury, while calling their own explanations for it speculative.
It is worth being clear about what kind of evidence this is. All three studies are archival and correlational, and the authors state that the paper does not directly test the mechanism it proposes.
It is a starting effect, not a finishing effect
Here is the limit that the inspirational version of this research always drops. Not one of these studies measured whether anybody achieved anything. They measured searches, gym entries and signatures, all of which are first days. The authors themselves report that the surge after a holiday spikes and then declines rapidly, while the lift from a new week, month or semester wears off more gradually.
So the finding is not that an ending remakes you. It is that an ending is a cheap and reliable way to get a first day out of yourself.
That is less than the genre promises and more useful, for two reasons the authors give. The calendar hands out dozens of these boundaries a year, so a start that fails does not have to wait for next January. And a brief lift in motivation is sometimes exactly enough, because some things only need one action: the appointment booked, the form signed, the first call made.
There is a catch on the other side of it, which they also flag. If you are counting on wiping the slate clean at an upcoming boundary, you may feel freer to make a mess before it arrives. The clean ledger cuts both ways.
Some of this is a calendar, not a law
A 2026 analysis by Bora Moon, Hyunwoong Pyun and Wonseok Jang at Sungkyunkwan University compared search patterns in the United States and South Korea from 2011 to 2019. The American pattern replicated cleanly. The Korean one did not. Monday still worked for both diet and exercise searches, but the first of the month was not significant for either, and the start of the year did not significantly predict exercise searches at all, at p = 0.27. Korean health searches instead followed an arc that peaked in the middle of the year, around 3.79 months for diet and 6.02 for exercise. No such curve appeared in the American data.
Their own cautions are worth carrying: this is search interest rather than behavior, Google holds roughly 40% of the Korean market against about 80% in the United States, only two search terms were used, and lunar calendar dates are an unruled-out confound. The authors suggest social pressure about summer clothing as the reason for the mid-year peak, and say plainly that they did not test it. This paper also shares a method with the first one, so the agreement between them on the American data is not fully independent confirmation.
Still, the direction is instructive. The feeling that January is the natural place to begin again is partly a habit of the calendar you were raised on.
The ending you probably mean
Now the part that matters most, and the part no article built on this research ever includes. In their own discussion, Dai and colleagues note that landmarks stained by negative emotion, and they name divorce and the death of a family member, may not increase motivation at all, because people have to cope with the experience first. They call it an open question for future research.
Which means the endings that are hardest, and the ones most people are actually thinking of when they ask how to begin again, sit outside what this evidence can speak to. New Year’s Day and a bereavement are not the same kind of boundary, and it would be a small cruelty to hand someone in the second situation a finding drawn from the first and let them conclude they are doing it wrong.
What survives all of this is modest and real. A marked boundary raises the odds that you begin something. On the authors’ own reading, you can also put a boundary wherever you like, since they suggest people may be able to construct turning points for themselves rather than waiting for the calendar to supply one. And the whole thing resets constantly, which means the useful question after a failed start is not what went wrong with you, but which boundary is next, and it is almost always Monday.