hayek's hamburger hell

sir, this is a wendy’s

The other week there was a lovely opportunity to observe the way in which economists inhabit a mental reality which is quite adjacent to, but often very different from, the economy.  An American hamburger chain announced (and then immediately rolled back when people yelled at it) that it was planning to use Artificial Intelligence to implement “dynamic pricing” to charge people more for their food at busy times of the day and less when things were quieter.

Economists, predictably, loved this idea, and many of them wrote opinion pieces and blogs explaining to the consumers that they were wrong.  Marketing people, also quite predictably, and possibly out of some dim memory of an old proverb about telling the customers they are wrong, hated it.

More specifically, someone who has come up through microeconomics is always going to like this sort of scheme, because when the parameters you have to work with are price, quantity and cost, anything which flexes the price to be closer to the marginal cost is going to show up as an efficiency improvement.  The people paying the higher price are the ones who value that particular “burger X at time T” combination the highest; other people who aren’t so stuck on a time slot get some of the benefit of cheaper production costs passed on.  Many of them pointed out that airlines use load-factor management, Uber uses surge pricing and bars have “happy hour” offers and lunchtime specials, and nobody complains about those; this is just a more sophisticated version of the same thing.

And some marketing people pointed out in return that a) people do complain about Ryanair and surge pricing, a hell of a lot, plus b) making it more sophisticated is likely to be worse for the customer.  Because – and this is where economics and marketing occupy parallel and adjacent realities – there is actually a lot more going on in the purchase decision than just quantity and price.

Originally, according to some historians, sticker prices were invented by the Quakers.  Allegedly (I have not checked this), they had a theological belief that it was dishonest to charge someone more than a “fair” price reflecting the cost and trouble of production, and also dishonest to offer someone less than goods were worth.  Consequently, Quaker merchants and industrialists set a price and refused to haggle.

People really liked this invention.  It helped that the Quakers had a reputation for honesty, but even when the practice spread to more pragmatic faiths, it’s a huge saver of time and effort.  A regular theme of this ‘stack has been that, in cognitive terms, “yes/no” is a much cheaper operation than trying to work out a reservation price, allow for strategic behaviour, consider the need for “schmuck insurance”, and so on.

In a modern industrial economy, there is a lot of cognitive demand on everyone at any given time.  One of the services that most companies provide is that of deciding on a simple pricing schedule which, on average, covers their variable and overhead costs, and presenting their customers with a “take it or leave it” proposition. 

Implementing any form of dynamic pricing is effectively pushing this cognitive load back onto the customer.  Rather than just thinking “a burger costs four quid, do I want one”, the customer is suddenly hit with a bunch of questions like “what do I think a burger is going to cost if I go into the shop in ten minutes?”, “how hungry am I, and how hungry will I be if I wait an hour?”, “is there a football match or something which might have significantly changed the normal pattern?”, “what’s my implicit price for waiting for lunch or being hungry later in the afternoon?”, “even though I can afford four quid, will I feel like a schmuck if the price goes down to three quid five minutes later?”.  It’s not exactly like this meme, but it’s not exactly not like it either:

Translating the marketing intuition into the language of economics, restaurants sell a bundled good-plus-service.  This is more apparent at the high end, but even fast food retail has a significant service component, of which immediacy – the ability to make the decision and conclude the transaction quickly and easily – is a significant component.  The implementation of dynamic pricing is a degradation of part of that service component; it makes the overall product offering worse because you have to spend more of your valuable brainpower thinking about it.  When you look at the cost of beef and bread compared to the price of a hamburger, you can see that the service component is a very significant part of the overall value of the good.  No wonder customers hated this idea.

A big chunk of my book (of course it’s relevant, it’s not like I’ve been thinking about anything else for the last three years) is about this problem in economics – having won the socialist calculation debate in the 1920s and 30s, economists went on to presume that Hayek said all there was to say about information, despite having written most of his key works before 1948 and Claude Shannon’s “Mathematical Theory of Communication”.  In fact, matching the variety and complexity of the world with the capacity to process it is every bit as important an economic phenomenon as matching supply and demand.