Just how far are people willing to go to achieve “completeness”? Recent research I conducted with Leslie John, Elizabeth Keenan and Michael Norton of Harvard Business School investigated whether it’s possible to harness this desire to motivate people in specific ways.
In a series of studies, we used visual cues and verbal descriptions to artificially reframe individual items, from donations to tasks, as cohesive but otherwise arbitrary groups. We then measured the effect of this pseudo-set framing on people’s effort levels and completion rates, and found that behavior changed in significant and meaningful ways.
Our first test was in the field; we teamed up with the Canadian Red Cross (CRC) to conduct an experiment during their 2016 Holiday Campaign. The CRC randomly directed more than 7,000 donors to one of three nearly identical web sites. One version displayed the usual web site, a platform offering donors the option to give money and/or up to six aid items (like hot meals and blankets). A second version encouraged donors to give the six aid items—the more, the better—and “rewarded” them with a badge for each one added to the cart. A third version also encouraged donors to give the six aid items, but, this time described them as component parts of a “global survival kit,” presented a graphic that filled in as items were added and showed text marking progress toward “100 percent” completion, indicating all six goods were in the cart.
Once the final donations were tallied, we found that the global survival kit framing led four to seven times as many people to donate all six items (as compared to the business-as-usual site and the badge prompts). By merely tweaking the framing, and without changing anything about the choices themselves, we were able to systematically shift what donors chose to give. Next, we wondered: Just how arbitrary could these sets be and still elicit the same behavior? We tested this in several follow-up laboratory studies, which depicted pseudo-sets in different ways. In all cases, the framing made people significantly more likely to reach completion—spending more time, or in the case of the gambles, incurring more risk—relative to control conditions, even though there were no rewards for doing so, and even when the total arbitrariness of the grouping was made explicit.
So what exactly makes pseudo-set framing work? In an online experiment, we showed one group of study subjects images of one, two or three loose beers with no product packaging, and then asked how many additional bottles they’d want to buy. Most said they’d purchase either nothing more or the number needed to add up to six. However, when we presented a second group of subjects with a four-pack container—pre-filled with one, two or three bottles—they overwhelmingly said they would purchase only the extras needed to fill all four slots, no more and no less.
Kate Barasz is an assistant professor of marketing at IESE Business School in Barcelona, Spain.