This piece was co-authored by Cait Lamberton, Associate Professor of Business Administration at the University of Pittsburgh's Joseph M. Katz Graduate School of Business, and Benjamin Castleman, Assistant Professor of Education and Public Policy at the University of Virginia's Curry School of Education.
On one hand, nudging is getting creative - and good. Take a recent project in which researchers either encouraged people to think of using a coupon when they saw a stuffed alien on a cash register--a distinctive memory cue--or to think of using their coupon when they saw a cash register--an everyday memory cue. They found that the distinctive cue reminded consumers to use a coupon much more effectively than did the cue we see every day, like the cash register itself. Sending patients low-cost text message reminders that state the precise economic costs of medical appointments lowers no-show rates from 11.1% to 8.4%. Portioning a shopping cart into produce and non-produce "sections" increased the purchase of fresh fruits and vegetables from $3.95 to $8.95 per customer.
On the other hand though, there are cases where even beloved nudges, like social norms and informational prompts fail. Employer wellness programs that often focus on creating a social norm related to weight-loss appear to have vanishingly small effects, and may leave more damaged morale in their wake than they increase productivity. Providing people with more salient information also doesn't always work: a recent analysis indicates that on the whole, providing nutrition data on menus hasn't reduced restaurant calorie consumption significantly, either for adults or children.
Even in "successful" nudges, a substantial portion of the population seems immune. For example, while using stuffed aliens as reminders was more effective than was using a non-distinctive cue, it still prompted less than a quarter of participants to use the researchers' coupon, despite the fact that using the coupon would save the shopper money. So even in the best case, 75% of participants were not doing something in their own self-interest.
What's getting in the way of these nudges?
The problem may be something we're calling "sludge" - (1) characteristics of the decision environment or (2) characteristics of the decision-maker that limit the effectiveness of a nudge. To design more effective interventions, we need to address both of these sludge sources. We propose two next-generation approaches to take on this task:
- Bundling Nudges to Raise Overall Effectiveness
Sometimes one nudge just isn't enough.
Say that we want to nudge people toward taking the stairs. We decide to use a salience manipulation - we put a neon arrow over the staircase and leave the elevator area unmarked. This prompts the staircase to visually jump out, making it more likely that we factor the stair climbing option into our decision making. As we'd expect, based on existing nudge practice, salience helps: we see 10% more stair climbing.
But we think we can do better than just increase stair climbing by 10%, and we think we can get there by designing a nudge-bundle.
What should be in the nudge-bundle? To answer this, we'd recommend a "Sludge Audit," where we ask, "What attentional, motivational, or identity-based barriers may keep a nudge from working?" For example, some people may erroneously believe the elevator is faster than climbing the stairs and don't want to waste time. Perhaps we also find an emotional barrier: people also feel that the stairwells are dingy or spooky and therefore, associate negative emotions with stair-climbing.
This doesn't mean we ditch our neon arrow. Instead, we bundle our salience cue with representations about the time people can save using the stairs, so that people who want to be maximally efficient see the virtue in making the climb. We may also install better lighting or play cheerful music in the stairwells, replacing prior negative affective associations with better ones. These "bundled nudges" may cost only a little bit more than the simple neon arrow, but they are likely to be much more effective, because they account for the multiple factors that get in the way of the desired behavior.
To date, nudging has largely been a one-size-fits-all game. But technology and data science increasingly enable us to personalize nudges to people based on their circumstance and even based on their response to or engagement with earlier nudges or information.
- We know defaults in charitable request letters can prompt giving, in one study increasing the proportion of individuals who give to a charity from 19% to 81%. But some research suggests that defaults may not always increase aggregate donation amounts received by charities, because they prompt some prospective givers who would have otherwise given more generously to peg their giving to a smaller default amount. So might it make sense to customize defaults? For example, perhaps a charitable donor is given defaults calibrated to their personal past contribution plus 5%, rather than the same default categories with which all donors are presented. Might this nudge to give slightly above a familiar reference point be more successful than a nudge that fails to recognize the level of giving the individual was comfortable with in the past?
Beyond allowing us to develop more effective interventions, these approaches may illuminate decisions that are truly un-nudge-able, even given our most comprehensive efforts. In these cases, it makes sense to push for more costly interventions, deep attitude change and policy work - a nudge to remind ourselves that not all important policy problems have simple, low-cost solutions.