A number of famous Business Schools in the UK and US such as MIT Sloan, NYU Stern and Imperial College have launched new programmes in business analytics. These courses have been nicknamed ‘Big Data finishing school’. Why might qualifications in this area be highly valued by firms?
Employees who have the skills to collect and process Big Data might help firms to successfully implement a pricing strategy that approaches first-degree price discrimination.
First-degree price discrimination is where the seller of a product is able to charge each consumer the maximum price he or she is prepared to pay for each unit of the product. Successfully implementing this type of pricing strategy could enable a firm to make more revenue. It might also lead to an increase in economic efficiency. However, the strategy might be opposed on equity grounds.
In reality, perfect price discrimination is more of a theoretical benchmark than a viable pricing strategy. Discovering the maximum amount each of its customers is willing to pay is an impossible task for a firm.
It may be possible for some sellers to implement a person-specific pricing strategy that approaches first-degree price discrimination. Firms may not be able to charge each customer the maximum amount they are willing to pay but they may be able to charge different prices that reflect customers’ different valuations of the product.
How could a firm go about predicting how much each of its customers is willing to pay? Traditionally smaller sellers might try to ‘size up’ a customer through individual observation and negotiation. The clothes people wear, the cars they drive and their ethnicity/nationality might indicate something about their income. Second-hand car dealers and stall-holders often haggle with customers in an attempt to personalise pricing. The starting point of these negotiations will often be influenced by the visual observations made by the seller.
The problem with this approach is that observation and negotiation is a time-consuming process. The extra costs involved might be greater than the extra revenue generated. This might be especially true for firms that sell a large volume of products. Just imagine how long it would take to shop at a supermarket if each customer had to haggle with a member of staff over each item in their supermarket trolley!! There is also the problem of designing compensation contracts for sales staff that provide appropriate incentives.
However the rise of e-commerce may lead to a very different trading environment. Whenever people use their smart phones, laptops and tablets to purchase goods, they are providing huge amounts of information (perhaps unconsciously) to the seller. This is known as Big Data. If this information can be effectively collected and processed then it could be used by the seller to predict different customers’ willingness to pay.
Some of this Big Data provides information similar to that observed by sellers in traditional off-line transactions. However, instead of visual clues observed by a salesperson, the firm is able to collect and process far greater quantities of information from the devices that people use.
For example, the Internet Protocol (IP) address could be used to identify the geographical location of the customer: i.e. do they live in a relatively affluent or socially deprived area? The operating system and browser might also indicate something about a buyer’s income and willingness to pay. The travel website, Orbitz, found that Apple users were 40 per cent more likely to book four or five star hotel rooms than customers who used Windows.
Perhaps the most controversial element to Big Data is the large amount of individual-level information that exists about the behaviour of customers. In particular, browsing histories can be used to find out (a) what types of goods people have viewed (b) how long they typically spend on-line and (c) their previous purchase history. This behavioural information might accurately predict price sensitivity and was never available in off-line transactions.
Interestingly, there has been very little evidence to date that firms are implementing personalised pricing on the internet. One possible explanation is that effective techniques to process the mass of available information have not been fully developed. This would help to explain the growth in business analytics courses offered by universities. PricewaterhouseCoopers recently announced its aim to recruit one thousand more data scientists over the next two years.
Another possible explanation is that firms fear a backlash from customers who are deeply opposed to this type of pricing. In a widely cited survey of consumers, 91% of the respondents believed that first-degree price discrimination was unfair.
Big data is coming for your purchase history – to charge you more money The Guardian, Anna Bernasek and DT Mongan (29/5/15)
Big data is an economic justice issue, not just a Privacy Problem The Huffington Post, Nathan Newman (16/5/15)
MIT’s $75,000 Big Data finishing school (and its many rivals) Financial Times, Adam Jones (20/3/16)
The Government’s consumer data watchdog New York Times, Natasha Singer (23/5/2015)
The economics of big data and differential pricing The Whitehouse blog, Jason Furman, Tim Simcoe (6/2/2015)
- Explain the difference between first- and third-degree price discrimination.
- Using an appropriate diagram, explain why perfect price discrimination might result in an economically more efficient outcome than uniform pricing.
- Draw a diagram to illustrate how a policy of first-degree price discrimination could lead to greater revenue but lower profits for a firm.
- Why would it be so difficult for a firm to discover the maximum amount each of its customers was willing to pay?
- Explain how the large amount of information on the individual behaviour of customers (so-called Big Data) could be used to predict differences in their willingness to pay.
- What factors might prevent a firm from successfully implementing a policy of personalised pricing?