Price personalisation in digital markets

Briefing Note

DeepDive Podcast, accessible on Chrome

https://notebooklm.google.com/notebook/16414732-3a8d-4cc8-885d-93566f6defe6/audio

BRIEFING NOTE

Price personalization

Source:

Ennis, S., & Lam, W. (2021). Personalised Pricing and Disclosure. https://www.gov.uk/government/publications/personalised-pricing-and-disclosure

Executive Summary

This research paper takes an in-depth look at personalized pricing, where prices vary for the same product across buyers, without cost justification. It explores the ambiguous theoretical consequences for consumer welfare, highlighting both the potential benefits of expanded market access and the drawbacks related to behavioral bias and fairness. The report also considers the role of price transparency and disclosure requirements as possible policy responses, noting that while they can increase consumer awareness, they also carry risks of collusion and price standardization. Finally, it discusses various policy options, ranging from no regulation to prohibition, emphasizing the complexity of the issue and the need for solid empirical evidence before any intervention.

Key words

Personalized prices

Price transparency

Consumer welfare

Policy options

Introduction:

This research paper, prepared for the Department for Business, Energy and Industrial Strategy (BEIS), examines the complex relationship between personalized pricing and price transparency. It aims to synthesize relevant theoretical and empirical literature to develop a policy perspective on the impacts of disclosure in markets where personalized pricing is practiced. The question of the policy response to potential online price personalization is considered complex, particularly with regard to distributional consequences.

Definition of Custom Pricing:

Personalized pricing (or price discrimination) is defined as a pricing strategy where the same product is sold by the same seller to different buyers at different prices, without this difference being justified by variations in quality or costs. The report distinguishes several types of price discrimination:

Uniform price: All consumers pay the same price.

Menu pricing (second-degree discrimination): Companies offer a range of price-product options and consumers self-select.

Group pricing (third-degree discrimination): Different prices are offered to different groups of consumers based on indicators correlated with their willingness to pay (age, location, purchase history, etc.).

Perfect price discrimination (first-degree discrimination): An individualized price is offered to each consumer based on their online behavior.

The development of "Big Data" and "data mining" makes it increasingly possible to personalize prices at a very fine level, often without consumers being aware of it.

Theory: Consequences on Well-being:

The theoretical literature suggests that the overall welfare consequences of personalized pricing are ambiguous and depend on many factors, including market structure, the nature of competition, market dynamics, the quantity and quality of information held about consumers, the technology available to firms and consumers, and the sophistication of consumers.

Potential benefits: Personalized pricing can expand market access for consumers with low willingness to pay. It can also increase business revenue, enabling the provision of products that would otherwise be unviable. In some cases, it can increase competition and benefit consumers by reducing average prices.

"Personalized pricing may be beneficial in that it expands market access to consumers with a low willingness to pay."

Potential drawbacks: Certain consumer groups, particularly those prone to behavioral biases, may be disadvantaged. Concerns about fairness, loss aversion, and regret may make personalized pricing unprofitable for companies. It may also make price comparisons more difficult and raise privacy concerns.

"Some groups of consumers may be hurt under personalized pricing, especially those who suffer from behavioral biases such as optimism and naivety."

The report emphasizes that price variations do not necessarily mean price discrimination based on consumer information; they can result from price randomization strategies in the presence of search costs.

Role of Price Disclosure and Transparency:

Demand-side measures, such as disclosure requirements and price transparency, can improve consumer awareness and facilitate price comparisons. However, their welfare implications are not always positive due to distributional effects between knowledgeable and less knowledgeable consumers.

Disclosure of future prices could lessen competition and increase overall market prices if firms can commit to those prices.

Increased market transparency could facilitate tacit collusion, especially with the help of artificial intelligence.

◦

"Disclosure that improves market transparency and makes competitors more knowledgeable about actual transaction prices could facilitate less price personalization along with tacit collusion; this effect could be exacerbated by the high data availability on personal characteristics when combined with enhanced possibilities to analyze data via artificial intelligence techniques."

Empirical and Experimental Data:

The report notes that empirical evidence of widespread online price personalization is limited. Personalization of offers (the selection of products presented) appears to be much more common than price personalization itself.

The sectors where digital price discrimination appears to be most prevalent are airlines and hotels, although this may be linked to demand management programs and dynamic pricing. The magnitude of this personalization generally appears to be low.

Studies have found forms of price customization on some sites, but the sites tested were not chosen at random.

For essential services (postal services, water, energy, telecommunications), there is little evidence of price personalization.

Consumer perceptions of the frequency of personalized pricing are often higher than research suggests.

"The perceived incidence of online personalized pricing is that 42% believe some or nearly all websites use it... though these figures do not seem borne out by the main mystery shopper findings."

​

Informal examples suggest that some companies have experimented with or are practicing price customization, but public reaction can be negative, as Amazon's experience has shown.

​

Regarding the disclosure of price discrimination, empirical evidence is scarce. Disclosure could take different forms: informing about customization, explaining how it works, or showing how the offered price compares to others.

A study showed that increased transparency and simplification of the cookie deletion process led to a decrease in the rate at which participants switched platforms at the time of purchase confirmation.

Consumers have expressed concerns that prices could be higher than usual and that their data would be used to build an online profile.

Interestingly, potentially vulnerable customers showed a greater increase in awareness of price personalization when there was high transparency.

However, price transparency can also facilitate the formation of cartels and lead to higher prices.

“Transparency over pricing can support cartels.”

The report stresses that policy conclusions based on current evidence would be difficult to draw due to its limited nature and often modest effects.

Policy: Considerations and Options:

Faced with personalized pricing, perceived as unfair by many consumers, the question of a political response is being debated. There is a broader consensus on the value of personalized advertising and the personalized presentation of product options.

​

Policy considerations include the risk of making access to goods and services more expensive for low-income groups if price discrimination in their favor is prohibited. The loyalty penalty, where long-term customers pay more, is also a concern. Maintaining consumers' incentives to shop around for the best deals is crucial.

​

Policy options presented include:

No disclosure requirement: Avoids potential adverse effects and preserves incentives to seek low prices. Lack of compelling evidence of a specific problem could justify this default option.

Broadly mandatory disclosure: Requires specifying the nature of the disclosure (price difference, customization factors, price range). Enforcement can be difficult, and this could lead to price increases and standardization, or even support cartels.

Making personalized pricing illegal: Could lead to higher prices for the poor and impact business models based on price discrimination. Enforcement would remain a challenge.

Selective disclosure: Targeting products where customization results in higher prices for vulnerable groups. Requires a clear legal framework and could benefit from randomized controlled trials to test the effectiveness of disclosures. Encouraging "simplification of switching" between suppliers is also mentioned as a potential solution.

Conclusion:

The report concludes that the fundamental elements to consider in assessing the impact of personalized pricing and disclosure are the nature of demand and substitution between products, the impact of information on consumer decision-making, and the evolution of prices over time.

​

The current and future prevalence of personalized pricing remains a key question, given the limited empirical evidence to date. Identifying the true extent of price personalization is a major empirical challenge, potentially exacerbated by AI-influenced pricing.

Any disclosure policy should provide clear guidance to companies, enable effective enforcement, have clearly defined objectives and avoid unintended negative consequences.

Sharing data, in compliance with the GDPR, could be a way to strengthen competition and reduce price discrimination.

It is likely that personalized pricing could benefit financially disadvantaged consumers through more refined price discrimination. However, it is crucial to ensure that behavioral biases do not lead vulnerable groups to systematically pay more. Policymakers should be cautious before imposing uniform pricing, as this could harm the most vulnerable who currently benefit from lower personalized prices.

In summary, policy options range from no disclosure to broad or selective requirements, or even a ban on personalized pricing. Any policy decision should be based on solid evidence of harm, particularly to vulnerable consumers, and consider the risks of unintended consequences. Additional research and randomized controlled trials for specific products would be ideal before any policy implementation.

Paper Summary Initial Draft By NotebookLM