What is personalised pricing and why is it used?

Personalised pricing involves a firm’s use of data gathered from customers (such as location, purchase and browsing history, and, potentially, demographic data) to set tailored prices according to the characteristics of a particular customer. This allows firms to personalise their pricing based on what customers are willing to pay, reducing prices for customers that would ordinarily not purchase a product or service and increasing prices for customers who are willing to pay more for the same product or service.

Customers generally perceive some unfairness in having different prices charged for the same products or services. But would it be a contravention of the South African Competition Act 89 of 1998 (Act)? While other provisions in the Act may also be relevant, we, for the purposes of this alert, consider whether personalised pricing constitutes price discrimination.

When can personalised pricing be price discrimination?

The Act prohibits a dominant firm from charging a different price to customers in relation to equivalent sales for products or services, in instances where such differentiated pricing will result in anti-competitive effects. A 2019 amendment to the Act also introduced a contravention where such differentiated pricing impedes the ability of small and medium businesses (SMEs) or firms controlled by historically disadvantaged persons (HDPs) to participate effectively in a market.

Prohibited price discrimination can only arise where it is undertaken by a dominant firm. In terms of the Act, a firm is dominant where it has more than 45% market share or if it has less than 45% market share and possesses market power, meaning that the firm has the ability to control prices, exclude competitors or behave to an appreciable extent independently of its competitors, customers or suppliers.

Establishing the correct market definition is a challenge in the context of personalised pricing, since the most common market definition tools, such as the SSNIP test, which analyses substitutability by considering which products consumers would go to in light of a small price increase, rely on an assumption that the price is the same for the same product or service. Using such tools could lead to narrowing the market to each individual customer. By the same token, using the traditional analysis of dominance and market power would result in the firm undertaking personalised pricing arguably being, ipso facto dominant, since it can control prices independently without customers refusing to purchase. On these bases every firm using personalised pricing could be dominant and meet the first requirement to establish a price discrimination contravention.  

Assuming this first hurdle is overcome, a dominant firm’s personalised pricing can amount to price discrimination in terms of the Act if all of the following criteria are met:

• the firm charges different prices; or provides different discounts, allowances, rebates or credit; or provides different services in respect of the products or services (e.g. free delivery) to different customers;
• the differential pricing is in relation to the sale of equivalent goods and services of like quality; and
• it has the effect of substantially preventing or lessening competition or impeding the ability of SMEs and HDPs to participate effectively in a market.

When the above criteria are considered, it seems likely that personalised pricing would meet the first two. It would then be necessary to determine whether the price would have an adverse effect on competition or on the ability of SMEs and HDPs to participate in a market.

When can personalised pricing be anticompetitive?

Economists refer to three types of price discrimination:

1. “First-degree” price discrimination is where each customer is charged a different price that perfectly matches their willingness to pay.
2. “Second-degree” price discrimination is where a supplier charges different prices based on the quantity purchased (e.g. quantity discounts, special offers to customers who buy in bulk, buy-two-get-one offers, loyalty rewards cards for frequent customers, etc).
3. “Third-degree” price discrimination occurs where prices are different to different groups of customers (such as students, pensioners, etc).

The type of price discrimination we are considering here is first-degree price discrimination – this is the holy grail for businesses because the business captures the full consumer surplus. A consumer surplus happens when the price that consumers pay for a product or service is less than the price they are willing to pay. It therefore measures the additional benefit that consumers receive because they’re paying less for something than what they were willing to pay. 

There are different ways of conceiving potential harm or anti-competitiveness arising from differential pricing. It may be exploitative (given the inability of consumers to benefit from the consumer surplus) or exclusionary (making it difficult for players in the market to compete). It is this latter harm which is captured by the Act’s price discrimination provision (exploitative harm arising from price discrimination could potentially find a home in other provisions of the Act, such as excessive pricing).

The price discrimination provision focuses in particular on exclusionary conduct which results in what US antitrust experts call “secondary-line injury”. This occurs where a seller gives better prices to some customers and not others, and it hurts competition among the customers who compete with each other downstream. This is different to primary-line injury, which happens when a seller’s low prices hurt their direct competitors.

As it stands, most personalised pricing currently happens in relation to end purchasers of products (accommodation, car hire, etc). It may be that the Act’s price discrimination provision is not currently well suited to challenge the most prevalent algorithmic pricing. However, there is a significant possibility that personalised pricing will be undertaken more regularly in relation to sales to businesses of inputs. In those circumstances, the current Competition Appeal Court (CAC) jurisprudence on price discrimination (putting aside the amendment relating to HDPs and SMEs) would require establishing whether there is a reasonable possibility that the seller’s pricing structure will have a substantial effect on competition in the downstream market. The CAC has emphasised that this analysis does not require proof of actual consumer harm (Sasol Oil (Pty) Ltd v Nationwide Poles CC [2006] 1 CPLR 37 (CAC)).

Although it may not be well received by consumers, personalised pricing may not necessarily harm consumer welfare overall. By charging lower prices a firm could increase the number of transactions it engages in and facilitate access to products or services by consumers who would have, but for the personalised pricing, been unable to afford such products or services. In an effects-based analysis, this access benefit would need to be weighed up against a reasonable possibility that purchasers who are charged more for the same product or service may not compete as effectively in the downstream market. It is possible that this access benefit may justify higher prices for some customers because customer welfare would benefit overall.

It is important to note that, post the 2019 amendments, the ability to prosecute personalised pricing to SME and HDP customers does not require the effects-based standard described. It appears, from the Price Discrimination Regulations and the Competition Commission’s current prosecutions, that price discrimination to SMEs and HDPs measures harm against firm-specific circumstances as opposed to the effect on the market as a whole. For instance, the Price Discrimination Regulations require analysis of the significance of the input in the purchaser’s own cost structure as well as the likelihood that the differential treatment would result in decreased demand for that purchaser’s product when determining whether SMEs or HDPs are impeded from participating in a market. Therefore, automated pricing algorithms that are influenced by criteria such as geographic location and purchase history may well end up biasing large purchasers over SMEs, or increasing prices to HDPs and thereby inadvertently contravene the Act when viewed against these lower thresholds for prosecution.

Conclusion

Automated personalised pricing favours certain customers over others and, particularly with the addition of a different standard for discriminatory pricing to SMEs and HDPs, there is scope for such pricing being found to contravene the price discrimination provision in the Act. That said, the existing provisions in the Act do not necessarily provide the framework for a proper analysis of the harms and benefits arising from such relatively new technologies and determining whether they are helpful or harmful in the South African context.