How Google’s opaque packaging resulted in 3.1 million fewer calories consumed over 7 weeks
In 2012, Google noticed that its employees ate too much of the free candy offered in its offices, and the company feared that this would hinder efforts to keep employees healthy and happy.1 It developed Project M&M to address these concerns, hoping to improve employee well-being by moving unhealthy snacks out of plain sight and instead displaying healthy snacks.
By implementing physical design changes, Google was able to nudge its employees to consume fewer calories.1 Specifically, 3.1 million fewer calories were consumed over the course of seven weeks in its New York office. This calorie decrease was hoped to be associated with happier and healthier employees.
WANT TO WORK TOGETHER ON A RELATED PROBLEM?
Effective interventions start with a nuanced understanding of how decisions are made. Our mission is to help large organizations be better and do better, using behavioral science.
Rating = 5/5 (Direct benefit to participants, easy to implement, effective)
How opaque packages decreased calorie consumption among Google employees
|All snacks visible in clear containers||Pre-intervention measurements were not provided.|
|Sweet snacks placed in opaque containers||3.1 million fewer calories consumed over a 7 week period (the equivalent of 9 M&M packages for each of the 2000 employees)|
Nudges: Techniques to change people’s behaviors in predictable ways without forbidding any options or significantly changing economic incentives.
Physical design change: A type of nudge that focuses on changing the physical environment to facilitate performance of a desired behavior or to create barriers for an undesired behavior.
Too much of a good thing?
Free candy might sound like a dream come true. But what happens when this results in overconsumption? While many technology companies offer perks like free snacks, Google wanted to ensure that said “benefits” didn’t hinder the morale and productivity of its 40,000 employees.1 The organization set an overall goal to improve employee well-being, with a particular focus on healthy eating.
Good food = good mood
Researchers are no strangers to studying the relationship between food and happiness. One psychologist found that increased consumption of calories, carbohydrates, sodium, and saturated fat - all of which can be found in higher amounts in unhealthy snacks - was associated with a negative mood two days later.2 Another study found that eating fruits and vegetables - typically agreed to be healthy foods - was associated with positive mood gains.3
Google’s goal was clear: they wanted to reduce the consumption of unhealthy foods, in order to improve their employees’ mood, productivity, and well-being.
It’s not unusual for Google to gather data from their workplace programs.1 They also had plenty of experience with figuring out the relationship between a given program, e.g., maternity leave, cafeteria size, etc., and employee happiness.
So, when developing Project M&M, the behavioral scientists at Google conducted surveys on employees’ snacking patterns, collected data on the proximity of M&M bins to each employee, and consulted academic papers on food psychology to launch their experiment.
For the study, researchers wanted to know what would happen if they moved chocolates to opaque containers while prominently displaying healthy snacks like pistachios and dried figs in glass jars.1 Google’s Project M&M is an example of using physical design changes in choice architecture to nudge employees towards healthier consumption patterns. While the less healthy snacks were still available, they were less visually prominent than the healthier snacks, prompting employees to reach for items like dried figs instead.
The COM-B framework
Google’s Project M&M exemplifies the COM-B framework, a behavioral change approach that focuses on capability, opportunity, and motivation to influence behavior.5 COM-B is a popular approach in the field of public health.
- C(apability): Google employees had the physical capability to pick both healthy snacks and unhealthy snacks, both before and during the intervention.
- O(pportunity):, Google employees had the opportunity to pick the healthy snacks. Since all snacks are free and such benefits are built into the company’s culture, there are no social restrictions on their eating habits. The intervention only increased the opportunity for healthier choices by increasing the visual salience of the healthier options.
- M(otivaton): The intervention does not affect employees’ reflective motivation for unhealthy snacks. If they think about and decide to choose M&Ms over dried figs, they are still allowed to do so. Instead, the intervention affects the automatic motivation for unhealthy snacks. Since these snacks became less prominent, employees would not automatically think to eat them, increasing the chance they chose a healthier option instead.
- B(ehavior): Since unhealthy snacks were made less visually salient, the behavioral architects hoped that Google employees’ behavior would lean towards making healthier choices. And if the positive effects of healthier foods are to be believed, we might also expect employees to be more motivated to keep making healthy choices--since such choices would lead them to be happier.
Results and Application
2000 employees, 7 weeks, 3.1 million fewer calories consumed
In Google’s New York office alone, the project resulted in 3.1 million fewer calories of M&Ms being consumed by 2,000 employees over the course of seven weeks.1 That’s the equivalent of nine fewer vending-machine-sized M&M packages per employee.
Of course, this calorie reduction alone does not equate to happier employees, which Google recognizes.1 However, its office in Mountain View, California is often ranked in business publications like Fortune magazine as one of the best places to work, as a result of employee surveys. Google believes that this is partly due to the benefits it offers its employees and office design, contributing to the overall work environment.
|Retail & Consumer||A similar design change could be replicated in supermarket packaging to nudge consumers towards healthier snack options.
|Climate & Energy||Environmentally-unfriendly options could be made less visually salient to nudge consumers towards more environmentally-friendly products.|
|Public Policy||Some studies suggest that salience nudges can help citizens use easier, cost-effective online license plate renewal services.6|
- There was a lack of published information on the ethical dimensions of the intervention
- There were no restrictions on employee decisions
|Does the intervention demonstrably improve the lives of those affected by it?||
Room for Improvement
|While the nudge leads employees to healthier choices, it is unclear whether this correlates with employee wellbeing.|
|Does the intervention respect the privacy (including the privacy of identity) of those it affects?||
|It is unclear whether employee data was anonymized (when necessary) before being reported to the decision-makers at Google.|
|Does the intervention have a plan to monitor the safety, effectiveness, and validity of the intervention?||
Room for Improvement
|While effectiveness is quantitatively measured, it is unclear how safety and validity were monitored.|
|Does the intervention abide by a reasonable degree of consent?||
|It is unclear whether employees had to consent before answering the surveys they were provided.|
|Does the intervention respect the ability of those it affects to make their own decisions?||
|Employees were allowed to freely choose what they wanted to eat.|
|Does the intervention increase the number of choices available to those it affects?||
|The number of choices stayed the same.|
|Does the intervention acknowledge the perspectives, interests, and preferences of everyone it affects, including traditionally marginalized groups?||
|It is unclear who was being taken into account when project M&M was first being designed. Only Google employees are explicitly mentioned.|
|Are the participants diverse?||
|It is unclear what the demographics of the participants are.|
|Does the intervention help ensure a just, equitable distribution of welfare?||
|While reducing unhealthy eating is partially an equity issue, it is unclear whether the intervention had any additional, direct effect on equitable causes (e.g., keeping employees from underserved communities).|
Related TDL Content
Would your consumption choices change if a glass jar was constantly restocked with free snacks, compared to if it was always one-third full? Salience bias can explain why food presentation can influence our consumption preferences.
Project M&M set up the Google offices in a way that employees were being exposed to healthier food options. This is an example of the framing effect, which explains how our food preferences can be manipulated. If you want to learn about another intervention using a similar framing effect, read this piece, where we briefly document how framing effects led to a 61.3% increase in vegetable sales!
Google nudges its employees toward healthier food choices in hopes of improving their well-being, but would you expect nudges to also influence sustainable food purchases? Take a look at our interview with marketing researcher Jolien Vandebroele to find out how behavioral science can help us be more environmentally conscious.
- Kang, C. (2013, September 1). Google crunches data on munching in the office. The Washington Post. https://www.washingtonpost.com/business/technology/google-crunches-data-on-munching-in-office/2013/09/01/3902b444-0e83-11e3-85b6-d27422650fd5_story.html
- Hendy, H. M. (2012). Which comes first in food-mood relationships, foods or moods? Appetite, 58(2), 771-775.
- White, B. A., Horwatch, C. C., & Conner, T. S. (2013). Many apples a day keep the blues away – Daily experiences of negative and positive affect and food consumption in young adults. British Journal of Health Psychology, 18(4), 782-798.
- Atkins, L., & Michie, S. (2013). Changing eating behaviour: What can we learn from behavioural science? Nutrition Bulletin, 38(1), 30-35.
- Flanagan, A. E., & Tanner, J. C. (2016). A framework for evaluating behavior change in international development operations (IEG Working Paper 2016/No. 2). Independent Evaluation Group. https://openknowledge.worldbank.org/bitstream/handle/10986/25872/110890-WP-PUBLIC.pdf?sequence=1
- Castelo, N., Hardy, E., House, J., Mazar, N., Tsai, C., & Zhao, M. (2015). Moving citizens online: Using salience & message framing to motivate behavior change. Behavioral Science & Policy, 1(2), 57–68. https://doi.org/10.1353/bsp.2015.0018