A search on our blog host artsjournal.com re ‘big data’ yields a lot of hits, suggesting it is obviously something about which arts managers need to be aware of, and which citizens might have cause to fear as corporations and governments use data in hidden but nefarious ways. But is it really all that? A recent post from the Brookings Institution on big data and price discrimination highlights for me some of the unwarranted fears and speculations.
The post gives three examples of potential problems. First:
Companies could use Big Data to set different prices for customers. Economists call this price discrimination, which can benefit both retailers and consumers. For example offering a discount to seniors or students can generate more revenue for the seller and offers lower costs for the purchaser. Companies know that consumers in these groups are more likely to be price sensitive and offer them a cheaper product or service. Big Data enables companies to pursue this strategy with greater precision. One area of concern is that companies could target inexperienced customers with low sticker costs and then up charge them for extras. Big Data might make it easier for companies to use this “bait and switch” tactic.
A couple of things to keep in mind. First, companies have always used price discrimination to get more from the uninformed consumer. Deals are offered to careful shoppers, who are more price sensitive and willing to compare offers from the competition – that’s why grocery stores have coupons, electronics retailers offer rebates (only taken up by consumers with the energy and time to bother with the paperwork), and all stores have random sales more likely to be recognized by careful shoppers. It all predates big data. Further, ‘low sticker costs’ and ‘up charge’ for ‘extras’ is a commonplace method of two-part pricing, meant to discriminate between types of buyers. In these cases, it is generally the low-budget, low-use customer who gets the better deal: cheap desktop printers and expensive ink cartridges work to the benefit of skint college students, and to the cost of business customers.
Second:
A more worrisome issue is risk-based pricing. This approach popular with many insurers also has big benefits for consumers. For example auto-insurers charge more to people who are at fault for accidents. This can create disincentives for risky behavior. The potential problem is when customers are priced out of the market because of a factor they can’t control. If health insurance is prohibitively expensive for a person with a genetic disability then such a pricing scheme is unfair and possibly illegal. Big Data could also allow companies to uncover medical conditions that a consumer might not even know about themselves. Higher prices in this case could constitute a strange new type of privacy violation.
Again, this is not a ‘big data’ issue. Insurers have always differentiated on the basis of risk. Health insurers, if allowed, will charge higher premiums to those more likely to file future claims, but that is old news. Small data – e.g. a blood test – can reveal to a medical insurer that the customer has a condition that may lead to future woe, of which the customer was unaware. But that has been true for a long time, and has been insurer practice for a long time. The issue here is the flawed nature of a pure free-market for health insurance. That’s why good legislation prevents discrimination based on pre-existing conditions, if insurance must be obtained privately (and another reason why a public, single-payer universal system has a lot to be said for it, eh?).
Third:
The worst case scenario with Big Data pricing involves the treatment of protected classes. Federal law does not allow companies to treat people differently based on race, religion, and several other characteristics. Big Data pricing could inadvertently lead to higher prices for certain groups in a way that would violate American values and laws. For example an analytics platform could identify a large group of consumers who regularly buy evergreen trees in mid-December. A statistical model may not know these people were Christians celebrating a holiday, but it could still result in higher prices.
So, the guy who sells Christmas trees in the parking lot in December does some analytics and discovers a certain type of person is more likely to want to buy a tree, they are identified when they walk onto the lot by the computer chip implanted in their forearm containing data regarding personal characteristics and purchasing history, and it turns out Christians are more likely to have a higher willingness to pay for said Christmas trees, at which point … wait, what? This is so nonsensical I don’t know where to start. It concerns neither ‘big data’ nor ‘price discrimination’, it is simply ludicrous. If that counts as our ‘worst case scenario’ I think we are all going to be ok.
Arts managers: Small data is very useful. Audience surveys, and tracking patterns of sales for different events on different dates at different prices, can lead to very useful insights on pricing and price discrimination – what nights to offer cheap admission, how to scale the house. You can do it without having to hire a team of rocket surgeons to do the statistical analysis.
Arts consumers: Shop around, there are better deals for the careful shopper. If a firm is using sophisticated price discrimination techniques, remember it is generally to charge higher prices to the less price-conscious, the lazy shopper, the one with money to burn (Sprint even has a tv commercial to this effect). Don’t be one of those.
Leave a Reply