September 29, 2022

Lessons in leveraging behavioral science to encourage food waste prevention

By: Tabitha Sugumar, Kadeem Khan, and Eoghan Griffin

Rotting food emits vast amounts of greenhouse gases, so much so that if lost and wasted food were a country, it would be the third-largest emitter. Households generate 61 percent of food wasted at the consumer level, and encouraging the prevention of food waste among this population supports a significant step forward in reducing carbon emissions and the warming that results. The question, then, is: How do we convince people to take action to prevent food waste at a large enough scale to make a difference?

Researchers from the World Resource Institute (WRI) collaborated with Meta’s Data for Good team to address this problem using Facebook ads as a platform to reach the public at scale. The researchers tested different messaging approaches to drive awareness of the impact of food waste and intent to act in the United Kingdom (U.K.) and Germany, reaching over forty million people. In recognition of International Food Loss and Waste Day, we’re sharing what we learned about effective messaging for food waste reduction in the hope that it will be useful for future work on communication campaigns focused on climate action and sustainable living.

Campaign messages and content

The campaigns tested an informational approach as well as three different social norming techniques, which ran as short video ads. The informational approach strove to raise awareness about the extent of the food waste problem by providing information on how much food is currently wasted. For the research, the goal was to focus on two Global North countries in the same region with a large enough Facebook audience to test four different messages. We focused on the Global North because, in these countries, food is wasted largely in the later stages of the supply chain due to consumer behaviors. We selected Germany and the U.K. because we were interested in observing whether the messages tested would perform similarly in both markets.

The social norming campaigns attempted to encourage a change in user behavior by setting a social expectation of lessening food waste. This approach came from the behavioral science concept that people align their behavior to the actions and expectations of those around them. The hope was that the campaigns would make an impact by tapping into that pattern. Social norm messages have already proved effective in promoting other sustainability-related behaviors, such as water and energy conservation.

Each social norming campaign took a different approach to setting a social expectation of less food waste. The first campaign used dynamic norms, which emphasize that the norm is changing, indicating that more people are taking action to reduce waste. This approach communicates that change is possible, the specific change is trendy, and behavior is improving, even when the status quo is far from ideal, as is the case with food waste.

The second campaign highlighted the current state of behavior, a static descriptive norm, which is an approach that attempts to decrease the amount of food waste produced by those who waste more than most.

The third emphasized what people should do—in this case, waste less food—known as an injunctive norm.

In addition to the message itself, all campaigns contained a link to a WRI’s resource page, “5 Tips to Make Your Food Last,” so that people could click through for more information on how to prevent food waste.

After launching the campaigns, we evaluated the effectiveness of each message by comparing the post-campaign survey responses of those who saw the ads and those who did not. The survey asked respondents whether they remembered seeing the ad, as well as several questions regarding awareness of the food waste problem, attitude toward the issue, and inclination to act. A post-campaign lift, or a statistically significant percent difference in the responses between the group that saw the campaign and the group that didn’t, suggested that the campaign succeeded.

What we learned

Although all the campaigns had a distinct effect on at least one metric related to reported awareness of, attitude toward, or intent to take action to mitigate food waste, the dynamic and static norming messages proved most successful. These two approaches had an impact on responses to multiple survey questions, suggesting they have the potential to make a broad impact across user-reported awareness, attitude, and intent. These messages also proved to be the best strategy in terms of cost and magnitude of lift for one question per country. Overall, the success of these campaigns indicates that messaging informed by behavioral science is a valuable tool in driving awareness and information seeking through digital ad campaigns.

The dynamic norm message, which emphasized that households are committing to greater efforts in reducing food waste, was markedly successful in Germany and had a positive effect in the U.K. In both countries, dynamic norms proved to be the most effective strategy in increasing likelihood of sharing knowledge about food waste with family and friends, resulting in both the lowest cost per brand lift and the highest percentage lift, with a +1.7-percentage point lift in Germany and +2.3-percentage point lift in the U.K. Additionally, in the U.K., the dynamic norm message increased awareness that food waste is a big problem in the country. In Germany, the message was even more successful, achieving lifts of +1.2, +1.3, and +1.4 percentage points on questions about the importance of minimizing food waste in one’s household, self-efficacy in reducing household food waste, and agreement that people should try their best to waste as little as possible, respectively. It was also the most effective in driving traffic to the linked resource about how to reduce food waste.

Similarly, the static norm campaign, which highlighted the percentage of households trying to reduce food waste, also had an impact in both countries. This message was particularly successful in the U.K., where it achieved the lowest cost per brand lift and the highest lift (+1.8 percentage points) on the question about the importance of minimizing household food waste. The campaign also increased the likelihood of people agreeing that food waste is a big problem and that they have self-efficacy in reducing household food waste. In Germany, the static norm campaign had the lowest cost per brand lift and highest lift (+2.8 percentage points) on the question about self-efficacy in reducing food waste but didn’t attain a lift on any other question.

In both Germany and the U.K., the injunctive norm campaign achieved the highest brand lift on the question about whether people should try their best to waste food as little as possible, with lifts of +1.5 percentage points in Germany and +1.7 percentage points in the U.K. This is unsurprising, as the campaign emphasized that message, and this result reiterates the importance of ensuring that the questions used for measurement reflect the content of the ads they evaluate. However, in both countries, this campaign failed to achieve lift on any other question.

Finally, while the informational campaign had little effect in the U.K., in Germany it was the best strategy for driving impact on the question about the importance of reducing household food waste in terms of cost and size of lift.

Looking forward

These campaigns reached over forty million people and encouraged over 60,000 to click on a link for more information about preventing food waste at home. This impact demonstrates that social media can be leveraged as a communication channel to help inform people about building a more sustainable world. They show that behavioral science concepts are effective tools in developing online campaign messages, as the success of dynamic and static social norms in decreasing food waste demonstrated and the measurable impact of both an informational approach and injunctive norms support. Although it’s reasonable to assume that these findings are applicable in European and Western contexts more generally, the differences in results across countries suggest that it’s worth carefully considering the local context when deciding which messages to try. More research is needed to more fully understand the best way to apply strategies informed by behavior science to digital campaign messages.

While these campaigns represent a step forward in understanding how to reduce food waste through online campaigns, Meta continues to explore opportunities to deepen our understanding of the nuances of ad campaign messaging in promoting sustainable behavior change. We look forward to continuing to share our learnings on how digital campaigns, grounded in behavioral science, can have a positive impact on driving climate action.

Special thanks to Sophie Attwood, Senior Behavioral Scientist, and Stacy Blondin, Behavioral Research Associate, at the World Resources Institute (WRI) for conducting the research behind these findings.