Nudge Examples That Worked
Do nudges work? Here are 5 nudges with measured results from published studies, from tax letters and organ-donation defaults to energy reports and retirement saving. Each one changed how a choice was presented and moved the outcome it targeted. The outcomes differ (payment rates, consent, energy use, saving rates), so the results are listed one by one rather than averaged, and the strongest study designs come first. New to the idea? Start with what a nudge is.
- #1 · Public Policy · Natural field experiment
Social-Norm Tax Reminder Letters: The HMRC Natural Field Experiments
In two randomized trials run with HMRC and the Behavioural Insights Team, a single short message was added to the standard reminder letter. Messages included social norms ("Nine out of ten people in the UK pay their tax on time"), a minority framing ("You are currently in the very small minority of people who have not paid us yet") and, in Experiment Two, a local-and-debt norm: "The great majority of people in your local area pay their tax on time. Most people with a debt like yours have paid it by now."
Taxpayers paying within 23 days (Experiment Two): Standard reminder letter 33.6% → Letter + local & debt norm 38.6% (+5.0 pts) · Sample: 119,527 taxpayers in Experiment Two; 101,471 in Experiment One
38.6% = the 33.6% control rate plus the reported +5.0-point effect (a 14.9% relative increase).
Source: Hallsworth, M., List, J. A., Metcalfe, R. D., & Vlaev, I. (2017). The behavioralist as tax collector: Using natural field experiments to enhance tax compliance. Journal of Public Economics, 148, 14-31.
- #2 · Energy · Natural field experiment
Opower Home Energy Reports: Neighbor Comparisons Cut Energy Use by 2%
Opower mailed Home Energy Reports, on behalf of utilities, to randomly selected households. Each report compared the household's energy use with that of similar homes nearby and gave tips for saving energy. Households were also rated: the best categories ("Great", "Good") came with smiley-face icons.
Household energy use relative to the randomized control group: Households receiving reports: −2.0% vs. control households without reports · Sample: About 600,000 treatment and control households across the United States
Average across programs; range by pre-treatment usage: −6.3% (highest decile) to −0.3% (lowest decile).
Source: Allcott, H. (2011). Social norms and energy conservation. Journal of Public Economics, 95(9-10), 1082-1095.
- #3 · Healthcare · Online experiment
Johnson & Goldstein Organ Donation Defaults: Opt-In vs. Opt-Out
In an online experiment, respondents imagined they had just moved to a new state and were asked whether they would be donors. In the opt-in version, the default was not being a donor. In the opt-out version, the default was being a donor. In the neutral version, they had to choose with no default. Changing status took one mouse click, which largely rules out effort as the explanation.
Respondents agreeing to be organ donors (online experiment): Opt-in default 42% → Opt-out default 82% (+40.0 pts) · Sample: 161 online respondents; cross-country consent data for European countries
Neutral condition (forced choice, no default): 79%. This is stated willingness in a hypothetical scenario, not actual registrations or donations.
Source: Johnson, E. J., & Goldstein, D. (2003). Do defaults save lives? Science, 302(5649), 1338-1339.
- #4 · Technology · Lab experiment
Overcoming Algorithm Aversion: Letting People Slightly Modify an Algorithm's Forecasts
In incentivized forecasting tasks, participants estimated real high-school seniors' percentile scores on a standardized math test. They were told a statistical model was off by 17.5 percentiles on average, then chose between their own forecasts and the model's. In the treatment conditions, choosing the model still allowed some control: adjusting each forecast by up to 10 percentiles, or overriding up to 10 of the 20 forecasts.
Participants choosing to use the algorithm's forecasts (Study 1): Algorithm could not be changed 32% → Could adjust each forecast by up to 10 percentiles 76% (+44.0 pts) · Sample: Three incentivized experiments: Study 1 in the Wharton Behavioral Lab; Studies 2-3 online (Study 2: 816 participants)
Third condition (could override up to 10 of 20 forecasts): 73%.
Source: Dietvorst, B. J., Simmons, J. P., & Massey, C. (2018). Overcoming algorithm aversion: People will use imperfect algorithms if they can (even slightly) modify them. Management Science, 64(3), 1155-1170.
- #5 · Finance · Before–after (participants)
Save More Tomorrow (SMarT): Committing Future Pay Raises to Retirement Saving
Save More Tomorrow (SMarT) asks employees to commit now to contributing more later. Contributions rise with the first paycheck after each pay raise and keep rising at each scheduled raise until a preset maximum. Employees can opt out at any time. Because each increase comes with a raise, take-home pay does not fall.
Average saving rate of SMarT participants (% of pay): Before joining SMarT 3.5% → After four pay raises (40 months) 13.6% (+10.1 pts) · Sample: First implementation at one midsize manufacturing firm: 315 eligible employees; 162 joined SMarT (78% of those offered the plan)
This compares the same participants before and after; there is no randomized control group.
Source: Thaler, R. H., & Benartzi, S. (2004). Save More Tomorrow: Using behavioral economics to increase employee saving. Journal of Political Economy, 112(S1), S164-S187.
Not listed here: studies in our findings database that are not nudges (such as machine-learning evaluations) or whose intervention did not improve the outcome, and illustrative design examples, which are not evidence.