«Beyond nudges: Tools of a choice architecture Eric J. Johnson & Suzanne B. Shu & Benedict G. C. Dellaert & Craig Fox & Daniel G. Goldstein & Gerald ...»
Mark Lett (2012) 23:487–504
Beyond nudges: Tools of a choice architecture
Eric J. Johnson & Suzanne B. Shu &
Benedict G. C. Dellaert & Craig Fox &
Daniel G. Goldstein & Gerald Häubl &
Richard P. Larrick & John W. Payne & Ellen Peters &
David Schkade & Brian Wansink & Elke U. Weber
Published online: 25 May 2012
# Springer Science+Business Media, LLC 2012
The way a choice is presented influences what a decision-maker chooses.
This paper outlines the tools available to choice architects, that is anyone who present people with choices. We divide these tools into two categories: those used in structuring the choice task and those used in describing the choice options. Tools for structuring the choice task address the idea of what to present to decision-makers, Preparation of this article has been supported by NIA Grant 5R01AG027934 to the first author.
E. J. Johnson (*) : E. U. Weber Center for Decision Science, Columbia Business School, Columbia University, New York, NY, USA e-mail: email@example.com S. B. Shu : C. Fox Anderson School of Management, UCLA, Los Angeles, CA, USA B. G. C. Dellaert Department of Business Economics, Erasmus School of Economics, Erasmus University, Rotterdam, Netherlands D. G. Goldstein Yahoo! Research and London Business School, London, UK G. Häubl School of Business, University of Alberta, Edmonton, AB, Canada R. P. Larrick : J. W. Payne The Fuqua School of Business, Duke University, Durham, NC, USA E. Peters Psychology Department, The Ohio State University, Columbus, OH, USA D. Schkade Rady School of Management, UCSD, San Diego, CA, USA B. Wansink Applied Economics and Management Department, Cornell University, Ithaca, NY, USA 488 Mark Lett (2012) 23:487–504 and tools for describing the choice options address the idea of how to present it. We discuss implementation issues in using choice architecture tools, including individual differences and errors in evaluation of choice outcomes. Finally, this paper presents a few applications that illustrate the positive effect choice architecture can have on real- world decisions.
Keywords Choice architecture. Decision support. Options and alternatives.
Describing attributes Choice architecture, a term coined by Thaler and Sunstein (2008), reflects the fact that there are many ways to present a choice to the decision-maker, and that what is chosen often depends upon how the choice is presented. Choice architects have significant, if perhaps underappreciated, influence, much like the architect of a building who affects the behaviors of the building’s inhabitants through the place- ment of doors, hallways, staircases, and bathrooms. Similarly, choice architects can influence choice in many ways: by varying the presentation order of choice alter- natives, the order attributes and their ease of use, and the selection of defaults, to name just a few of the design options available. While it is tempting to think that choices can be presented in a “neutral” way (“Just the facts, Ma’am”), the reality is that there is no neutral architecture—any way a choice is presented will influence how the decision-maker chooses. Consider, perhaps, the best-known example. All choice presentations have a (usually implicit) default, even if the default is no choice is made, preserving the status quo. The default option will be chosen more often than if another option is designated the default. Thus, everyone, from a parent presenting bedtime options to a child to a government providing pension options to its citizens, influences choices and is a choice architect.
In this brief paper, we provide examples of the tools available to a choice architect.
We do not provide a theoretical account of why these tools affect choice, nor suggest a normative analysis of how a choice architecture ought to be designed or when to use which tool. Our goals are modest: to provide an initial roadmap and to identify, describe, and categorize many of the tools of a choice architecture with a few brief illustrative applications. We divide our list of tools into two broad categories, those used in structuring the choice task and those used in describing the choice options.
These two categories correspond, roughly, to the idea of what to present to decisionmakers and how to present it. We then turn to the challenges and opportunities these tools raise in implementation and provide some examples of effective application of choice architecture tools. Table 1 provides a summary of the tools and some relevant examples that are discussed throughout this paper.
1 Structuring the choice task
1.1 Number of alternatives
decision maker be presented with one option at time, or two options, three options, or even 10, 20, or 100 or more options simultaneously? There are times when a person has too few options, such as when the original Ford Model T were available in any color—as long as it was black. At other times, there can be the danger of too many options, what others have called the tyranny of choice (Schwartz 2004) or choice overload (Iyengar and Lepper 2000;
Jacoby 1984). For example, the number of Medicare drug benefit plans available to US seniors now exceeds 100 in some states, possibly overwhelming the processing capacity of many elderly decision makers (Kling et al. 2011), and investors can often 490 Mark Lett (2012) 23:487–504 face hundreds of options for retirement funds (Cronqvist and Thaler 2004). The trend in the marketplace is for more, not fewer options to be presented to the consumer, and additional options can also complicate choice in unexpected ways. In cafeterias, for example, additional foods can be “trigger foods,” resulting in seemingly unrelated additions that would otherwise have not been made (Hanks et al. 2012). In studies of high school cafeterias, the presence bananas and green beans decreased sales of ice cream while the presence of sugary sides, such as fruit cocktail and applesauce, increased sales of cake and chips.
To answer the question of how many options to present, the choice architect needs to balance two criteria: first that more options increase the chances of offering a preference match to the consumer, and second that more options places a greater cognitive burden on consumers because of the additional need to evaluate options.
Thus, to answer this question of balance, we should be concerned about the willingness of the decision-maker to engage in the choice process, the decision-maker’s satisfaction with the decision process, and more generally the nature of the processes that will be used to make the decision. Finally, as discussed later, the answer is contingent upon characteristics of the individual decision-maker. Older adults, for instance, with less processing capacity seem to prefer less choice than younger adults (Reed et al. 2008).
Despite the vast amount of research examining the effects of a number of alternatives on decision behavior (see Payne et al. 1993; Scheibehenne et al. 2010), the issue of balancing different objectives makes it hard to identify a simple recommendation for the optimal number of alternatives to present. However, some general guidelines apply. One wants the fewest number of options that will encourage a reasoned consideration of tradeoffs among conflicting values and yet not seem too overwhelming to the decision maker. Yet too few options may generate contextspecific preferences, a well-known phenomenon in choice, where the presence or absence of one option influences what is chosen. One recommendation that balances these considerations is that four or five non-dominating options may represent reasonable initial values for the choice architect given these tradeoffs. One could also proceed by starting with this limited choice set, but also provide the decisionmaker with the option of considering more options, if desired.
1.2 Technology and decision aids
More and more of the choices we make involve the use of some form of information technology (Murray et al. 2010). This technology may be introduced to assist in the choice task. For instance, we increasingly choose what to buy, what activities to participate in, or what to attend to via some form of desktop or mobile computer interface. Moreover, we may use technology-based tools such as search engines or product recommendation systems to help us identify attractive choice alternatives that we were not aware of, and to filter out ones that are not of interest to us (Bodapati 2008; Häubl and Murray 2006; Xiao and Benbasat 2007). We can also enlist the assistance of interactive decision aids that help us compare choice alternatives in terms of their attractiveness on various feature dimensions (Häubl and Trifts 2000;
Lynch and Ariely 2000). Yet another way in which the choices we make are increasingly facilitated by technology is the automatic personalization of user Mark Lett (2012) 23:487–504 491 interfaces to reflect our preferences (Hauser et al. 2009; Price et al. 2006). This interaction with decision technology is likely to increase in future years as computing devices become more unobtrusively integrated into our daily environment (see, e.g., Cook and Song 2009; Streitz et al. 2007).
Research has demonstrated that decision aids such as product recommendation systems can be highly beneficial to consumers, enabling them to find products that better match their preferences while at the same time reducing search effort (Häubl and Trifts 2000). However, these tools can also predictably influence consumers’ choices through very subtle architectural features such as the set of other products that are presented alongside a recommended alternative (Cooke et al. 2002) or which product attributes are made more salient by the system (Häubl and Murray 2003).
Thus, technology-based decision aids could be designed to steer consumers towards choosing products, services, or activities that are individually and/or socially desirable— i.e., healthy, environmentally friendly, etc.—without restricting their freedom to choose.
Given that consumers appear to show little resistance to such influence when it benefits profit-seeking sellers (Häubl and Murray 2006), they should be even more willing to accept these interventions when these are in their own and/or society’s interest.
One of the most powerful and popular tools available to the choice architect is the use of defaults. Defaults are settings or choices that apply to individuals who do not take active steps to change them (Brown and Krishna 2004). Collections of default settings or “default configurations” determine the way consumers initially encounter products, services, or policies, while “reuse defaults” come into play with subsequent uses of a product. At the finest level, a single question can have a “choice option default”, which on electronic forms can take the shape of a pre-checked box (Johnson et al. 2002).
Defaults have been shown to have strong effects on real-world choices in domains including investment (Cronqvist and Thaler 2004; Madrian and Shea 2001), insurance (Johnson et al. 1993), organ donation (Johnson and Goldstein 2003), and marketing and beyond (Goldstein et al. 2008). They appeal to a wide audience in their ability to guide choice, while at the same time preserving freedom to choose, and are often regarded as prototypical instruments of libertarian paternalism (Sunstein and Thaler 2003).
Through default-setting policies, choice architects can exert influence over resulting choices (Goldstein et al. 2008). The palette of policies includes simple defaults (choosing one default for all), random defaults (assigning a configuration at random, for instance, as an experiment), forced choice (withholding the product or service by default, and releasing it to the recipient only after an active choice is made), and sensory defaults (those which change according to what can be inferred about the user, for example, web sites that change language dependent on country of origin of the visitor). Products and services that are frequently purchased can use either of persistent defaults (where past choices are remembered) or reverting defaults (which forget the last changes made to the default configuration). They also can use predictive defaults (which intelligently alter reuse defaults based on observation of the user).
Choice architects should be mindful of the ethical risks involved in setting defaults (Smith et al. 2010). The ethical acceptability of using a default to guide choice has much to do with the reason why the default is having an effect (see Dinner et al. 2011 492 Mark Lett (2012) 23:487–504 for a discussion of those reasons). When consumers are aware that defaults may be set as recommendations in some cases, or manipulation attempts in other cases (Brown and Krishna 2004), they exhibit a level of “marketplace metacognition” that suggests they successfully retain autonomy and freedom of choice. However, if defaults have an effect because consumers are not aware that they have choices, or because the transaction costs of changing from the default are too high, defaults impinge upon liberty. An often-prudent policy, though not a cure-all, is to set the default to the alternative most people prefer when making an active choice, without time pressure, in the absence of any default. Running an experiment on a sample of the population can determine these preferences, and can be done in little time and at low cost in this age of Internet experimentation.
1.4 Choice over time