There is growing interest in applying principles from the field of behavioral science to
shift individual decisions in desirable directions through “nudges.” However, an important but
oft-overlooked feature of nudges is that they can leak information about the implicit
recommendations of their designers. In this paper, we report on paired field and laboratory
experiments highlighting that a widely-used nudge—encouraging pre-commitment—can
backfire, and that this appears to be due to unintentional information leakage. We also highlight
the benefits of a previously untested nudge—framing a future opportunity for behavior change as
a “fresh start”—as a means of encouraging future-oriented behaviors. We conducted a field
experiment intended to nudge increased retirement savings with 8,682 employees from four
major U.S. universities. Offering people the opportunity to choose to save now or to save at a
time delay significantly decreased savings over a nine-month follow-up period compared to
offering people the opportunity to choose to save now, but highlighting that delayed savings will
begin after a fresh start (e.g., an employee’s birthday) counteracted this negative effect. A
laboratory experiment suggests an explanation: when a nudge encourages good behavior “later”
(i.e., pre-commitment), it leaks the message that the target behavior is not an urgent priority,
undermining the nudge’s efficacy. Our findings highlight (1) the critical importance of pilot
testing nudges to assess the implicit messages that may be unintentionally leaked by their design
and (2) the potential of nudges that rely on fresh start framing