Online Choice Architecture and Vulnerability: Understanding the Impact of Reference Pricing on Consumer Behaviour
Executive Summary
This study examines how three common online sales practices – reference pricing, sensory manipulation, and time-bound elements – influence people’s choices and well-being when shopping online. Reference pricing refers to showing a higher previous or future price next to the current price (for example, “was £22.99, now £19.99”). Sensory manipulation refers to using eye-catching design on certain deals (for example, bold colours or large fonts). Timebound elements include countdown timers (for example, “20 seconds left on this offer”) and time-limited discounts (for example, “first 3 months free”).
These practices are widely used, particularly in three types of online marketplaces: broadband plans, gym memberships, and hotel bookings. We replicated these marketplaces to test how the practices work under real-world conditions. Specifically, we set out to measure whether they reduce decision quality, whether people facing financial insecurity or poor mental health are affected differently, and whether these practices discourage shopping around, which matters for competition and consumer welfare. Our findings are intended to support better design and oversight of online marketplaces, so the benefits of digital commerce are widely accessible while the risks of confusion or manipulation are reduced.
To study the effects of these practices, we built a realistic simulation experiment that took place across these three marketplaces. Participants were asked to shop and select the best offer in each one. A nationally representative sample of over 8,000 adults took part, including large sub-samples of people facing financial insecurity and poor mental health. Participants were entered into a lottery for each correct choice so they were motivated to behave much more similarly to how they would when shopping online.
People were randomly assigned to one of four groups: (1) reference pricing only; (2) reference pricing plus sensory manipulation; (3) reference pricing plus sensory manipulation and a time-bound element; or (4) a control group that saw none of these practices. We measured decision quality (did they pick the best-value option), time spent shopping, confidence in selected offers, shopping-around (whether they visited a competitor site), and the monetary amount spent shopping.
What we found
● Reference pricing reduced consumer decision quality. People shown “was/now” prices were less likely to choose the best-value option.
● Adding sensory manipulation and time-bound elements typically strengthened the negative effect of reference pricing, but this effect varied by market.
● Vulnerable people started from a lower baseline decision quality. People experiencing poor mental health or financial insecurity were less likely to pick the best-value option in the control group, but their choices were not disproportionately affected by reference pricing, relative to others.
● There is suggestive evidence that a combination of reference pricing, sensory manipulation and time-bound elements made vulnerable people spend more time picking a broadband plan, relative to others.
● Broadband decisions were hard for everyone, even without reference pricing. The complexity of broadband plans seemed to be the reason.
● Reference pricing reduced shopping-around behaviour. Reference pricing on its own reduced the visits of participants to a competitor site.
● Confidence and accuracy moved in opposite directions, with people who picked the best-value option feeling less confident in their selection. Even in the absence of reference pricing, sensory manipulation and time-bound elements, people who were more confident generally made poorer choices.
● Reference pricing countered the advantages of bargain-hunting and financial literacy. Self-described bargain-hunters and those with high financial literacy started out making more correct choices, but reference pricing, especially when combined with sensory manipulation and time-bound elements, eroded that advantage.
● Reference pricing tends to make consumers overspend, not underspend. This effect got stronger when combined with sensory manipulation and time-bound elements.
Based on these results, we recommend that regulators do the following:
● The Competition and Markets Authority (CMA) should conduct further investigations into the use and impact of reference pricing across markets. Previously, the CMA have done market-specific investigations on reference pricing 6 and published relevant principles for firms (CMA, 2024), but these focused primarily on whether or not the price comparison is genuine. Our research shows that reference pricing can harm consumer decision making regardless of whether the information shown is fairly presented. We recommend that the CMA explore the use of reference pricing across markets and that they expand and reframe their guidance for firms around consumer harm rather than whether or not information is genuine.
● Ofcom should investigate broadband pricing practices to address the serious difficulties consumers face in making the right choice in this market. They should consider our experimental findings on the harm caused by OCA practices, but also the findings from our control group which show how difficult it is for consumers to make good decisions in this market in the first place.
● All the regulators we engaged throughout the design of this experiment should continue to conduct research on the impact of OCA practices, especially on vulnerable consumers, to address the gap in existing evidence.