In a trial with 226 people, researchers kept tempting audiobooks on iPods that one group could access only at the gym — that group initially visited 51% more often than the controls, though the advantage faded as the study crossed the Thanksgiving break.

The next chapter was waiting at the gym. In an experiment with 226 people, access to an enticing audiobook became part of the incentive to turn up for a workout.

Katherine Milkman, Julia Minson and Kevin Volpp tested the arrangement in “Holding the Hunger Games Hostage at the Gym,” published online in 2013 and in Management Science in 2014. One group could access audiobook-loaded iPods only at the gym. Another received audiobook access and encouragement to restrict listening to workouts; a third served as the control.

The gym-only group initially visited 51% more often than controls. That advantage declined, particularly after Thanksgiving. This was an early attendance effect, not a sustained 51% increase across the entire trial or a measurement of improved fitness.

A reward available during the task

The arrangement illustrates temptation bundling: pairing something immediately appealing with an activity whose benefits may feel more distant. The pleasure becomes part of doing the activity, rather than a prize delivered only after it is finished.

That timing creates a practical distinction. A reward promised for completing a month of exercise requires waiting. An absorbing story can make the current visit more appealing. Whether the pairing works will depend on both sides: the entertainment must be something the participant wants, and the activity must allow them to enjoy it.

In Wharton’s account of the research, Milkman highlighted the challenge of sustaining the initial benefit and re-engaging participants after time away. The university participants had already expressed a desire to exercise more, an important feature of the population being studied.

The story supplied a reason to return.

The break reveals why follow-up matters

An intervention can improve the first few decisions without becoming a lasting routine. A holiday interrupts schedules, locations and access to the things that supported those decisions. If the effect changes afterward, that is useful evidence about durability.

It is not, by itself, a clean experiment on holidays. Separating the effect of a break from the passage of time would require an appropriate comparison. The cautious interpretation is that the original benefit weakened and the interruption was a notable point in that trajectory.

For a team designing a behavior-change product, this suggests measuring more than the launch period. The first week, the following month and the return after an interruption may reveal different problems. Initial appeal and successful restarting deserve their own measures, even when they belong to the same service.

A larger trial tested a lighter version

A 2020 field experiment extended the idea to 6,792 participants in a four-week exercise program. Participants received an audiobook with encouragement to combine listening and exercise, an audiobook alone, or neither.

The audiobook-plus-encouragement condition increased the likelihood of a weekly workout by roughly 10% to 14% and average weekly workouts by about 10% to 12% in the reported analyses during the program and follow-up, which extended to seventeen weeks afterward.

The additional benefit of explicitly explaining temptation bundling, compared with supplying an audiobook alone, was modest. The researchers noted a possible reason: giving someone an audiobook within an exercise program may already suggest listening while working out.

This was a different intervention from keeping iPods at a university gym. It matters because a self-imposed listening rule is easier to distribute at scale, but it also gives participants more freedom to listen elsewhere. The later study tested that less restrictive approach rather than simply repeating the original arrangement.

Returning after a missed visit is another target

A 2021 Nature megastudy compared 54 four-week digital programs among 61,293 members of an American fitness chain. Forty-five percent of the interventions significantly increased weekly gym visits, with reported gains ranging from 9% to 27%.

The best-performing intervention offered small rewards for returning after a missed workout. That result points to recovery from a lapse as a useful target, alongside making the activity attractive in the first place.

It does not establish that the same incentive would have prevented the earlier audiobook trial’s decline. The studies used different participants and programs. Together, however, they show why getting started, continuing and restarting should not automatically be treated as one behavioral problem.

Measure the behavior the product actually changes

A gym entry is a concrete, observable outcome. It does not reveal the length or intensity of the workout, and it cannot on its own establish a health improvement. Clear reporting keeps the measured action separate from hoped-for consequences.

The same principle applies outside exercise. An app open, a completed lesson and a retained skill are different outcomes. A product can influence the first without proving the last. Testing a pleasant pairing means deciding in advance which action matters and checking whether the effect lasts.

The audiobook experiment offers a specific possibility: an activity people already intend to do can become more attractive when something enjoyable accompanies it. Its declining effect also leaves a practical question for any implementation: what brings the person back when the routine is interrupted?