In this series of retail-focused articles, we explore opportunities offered to retailers by trends and innovations online and in tech.
Consumers have become accustomed to the concept of dynamic pricing – where the price of a product or service fluctuates on a day-by-day or hour-by-hour basis according to demand and supply.
This is old news.
Retailers are starting to adopt "personalised pricing" technology - systems that offer an individual a price for identical goods or services, based on the information the seller has obtained about the potential customer.
Online retailers have been able to observe characteristics of potential customers for many years, using data gathered about the customer via IP addresses, cookies and customer accounts.
The data sets that retailers can now obtain about an individual are vast, and machine learning in this field has become bewilderingly complex (for a short and fascinating introduction to this, watch "how machines learn"). AI is close to being able to determine (among other things) the maximum price an individual would be willing to pay for the particular good/service that they are viewing.
Offline counterparts have struggled to implement dynamic/personalised pricing models in bricks and mortar stores, mainly because the practicalities of changing prices in retail premises are onerous (i.e. employees manually substituting one price tag for another, across large stock ranges).
However, increased adoption of “smart shelving” (i.e. shelving with electronically displayed prices) has enabled offline retailers to manage shop-floor prices with greater flexibility. When combined with Bluetooth technology (e.g. Apple’s iBeacon), smart shelves can offer specific customers a price based on data transmitted from the customer's profile held on their device. B&Q trialed this technology as long ago as 2014.
What are some of the spanners in the works for adopting personalised pricing technology?
- Consumer perception: Consumers have not historically been receptive to, and reputations have been damaged by, price discrimination (e.g. when Orbitz used cookie data to adjust prices after it calculated that Apple Mac users would pay 20-30% more for hotel rooms than users of other devices). A lack of transparency may harm consumer trust in traders and their business practices. This might explain why some online retailers deny that they are using personalised pricing technology.
- Technological uncertainty: Retailers are collecting more and more data about their customers / potential customers, but there is a risk that their (and AI’s) understanding of that data is poor. Poorly understood data leads to poor execution and “artificial stupidity”, which will cause consumer frustration and less trust in machine learning.
- Regulatory uncertainty: Regulators have not commented meaningfully on personalised pricing since The OFT’s report in 2012 (which, in summary, said that the OFT will consider enforcement action if personalised pricing practices are found to be misleading or unfair, and that it is an area that they will keep under review).
The changes in law which will be introduced by the GDPR (in May this year) require retailers to obtain certain consents from each consumer in respect of the collection, transfer and processing of their data (with significant financial penalties for failing to do so).
There are also risks that, depending on how it is implemented, personalised pricing may contravene current privacy, unlawful discrimination, advertising and unfair trading laws.
Pricing is just one area of machine learning development that retailers are focusing on. Personalised pricing could make retail markets and businesses more efficient and enable businesses to build pricing strategies and brands around generating revenue through the lifetime of individual customers (rather than on short term, one-off purchases).
However, to avoid bad PR and the risk of regulatory intervention, businesses adopting personalised pricing strategies/technologies should:
- Be transparent with consumers about the data they are collecting and explain how that data is used to determine prices.
- Obtain consent to the collection, transfer and processing of each consumer's data and the use of that data for personalised pricing.
- Allow consumers to opt out.
- Reassure consumers that personalised pricing is focused on the consumer's benefit (i.e. to provide them with specific / relevant offers) and not just the benefit of the retailer's bottom line.