Potentially harmful online choice architecture
Do you want to pay a little more for priority boarding? What about express delivery? When shopping online, default options on e-commerce sites can lead people to make choices that they otherwise would not.
Alma Economics was commissioned by the Department for Business and Trade to conduct research into potentially harmful online choice architecture on e-commerce sites. More specifically, we were tasked with researching the prevalence and types of defaults on e-commerce websites and apps used by UK consumers. We also examined the impact of defaults on consumer welfare.
This research was carried out in two phases: (i) a prevalence study and (ii) an online choice experiment and follow-up survey. The prevalence study involved a comprehensive review of 558 of the most popular e-commerce websites and mobile applications in the UK. It found that 49% of sampled websites and apps had defaults - this was highest in the retail sector, where 69% of websites and apps had defaults. Pre-selected options were the most common type of default, usually found on the checkout page, and were most typically found to present options for delivery. Based on the fact that these defaults did not direct the consumer to select more expensive options, the prevalence study found that a majority of defaults used are not harmful.
Informed by these findings, we conducted an online choice experiment with 6,000 participants. The experiment simulated a real online shopping experience and tested the impacts of preselected options and mimicking defaults in two settings: delivery and protection plans for online products.
In our experiment, consumers were 60% more likely to choose the more expensive protection plan and 70% more likely to select the more expensive shipping option when using preselected defaults. However, mimicking defaults – where the options were made to visually look like defaults – did not impact the probability of selecting more expensive options.
We also found that older people had a greater probability of selecting the more expensive default option and, in a follow-up survey, that most consumers did not feel negatively about their choices after being exposed to a default.
Drawing on these findings, we developed policy recommendations that could improve the welfare of online consumers. These include establishing standards for default settings and working to raise awareness about the use of default options among more susceptible groups, such as older adults.
More broadly, we recommended that the government facilitate a consumer-centric culture of innovation where online retailers use defaults when it benefits consumers and protects vulnerable groups from harm.