Many years ago (we're talking about the late 1980s) I spent a year and a half as a shop manager. Well, that and a retail pharmacist running a pharmacy: but in addition to dispensing prescriptions, a chunk of retail management came into the picture. (The 24-year-old Charlie really sucked as a retail manager. I would not hire him. Luckily both stores were parts of small local chains with competent management backup—even if one of them was owned outright by a very happy junkie—and in any event made up most of their turnover via prescriptions. At which I merely sucked somewhat.)
Walking around various British cities over the past couple of years I've noticed an increasing number of vacant shop fronts (some in prime retail situations). I've also noticed a disturbing loss of diversity in our high streets, as quirky local shops give way to cookie-cutter national chains. I have, like most people, had the frustrating experience of trying to work out whether my mobile phone contract or the airline flight I'm been booking is actually the cheapest one that meets my needs, or whether I'm being gouged by a computer somewhere. And so I'm trying to put the pieces of the jigsaw together because I'm interested in guessing what our retail experience is going to look like in 10 years' time—the traditional "if this goes on ..." exercise beloved of science fiction writers.
Which leads me via the following chain of logic to a hypothesis about the future of retail. Starting in this blog entry with some abstract propositions about the forces that are going to shape retail over the next decade. (In the next blog entry: what this means for your retail environment.)
1. The internet is a communication tool that tends to disintermediate supply and demand.
—That is, it makes it easy for consumers to find whatever they're looking for.
2. Search tools exist that permit direct comparisons between competing offerings.
—First-generation search/comparison tools look for numerical data to base comparisons on. Price is the most accessible numerical data. So price is what gets presented to consumers first.
—Available evidence suggests that the majority of consumers consider price to be the most important aspect of any buying decision (if the goods on offer are equivalent, or there is economic stress). Note for example the 23% drop in organic food sales since 2008 in the UK, corresponding to the banking crisis and subsequent recession.
—More recently we've seen comparison sites with customer ratings to supply metadata about, for example, supplier reliability: but these tread dangerously close to defamation in some cases: see for example the class action lawsuit against Yelp (alleging extortion: dismissed, but shows the shape of things to come). Price comparisons are relatively safe.
3. It follows that, in the long term, the internet tends to induce pure price-driven competition between rival suppliers.
—This reduces profit margins both to intermediaries (the supply chain) and to producers.
4. Price fixing cartels (and government-mandated fixed prices) are illegal, and this prohibition has been globalized via international treaty law.
—So suppliers are trapped in a race to the bottom, unless (a) they can find some clear value-added proposition to attract consumers other than low price, or (b) they can work out a way to reduce price transparency in order to impair the accuracy of consumers' pricing decisions.
—luxury goods and designer brands adopt strategy (a). Some other well-known organizations go this way, too: Apple is an obvious example in the consumer electronics sector. (Underlying Apple's value proposition is simplicity: a simple, easy to understand product range, pared-back industrial design, "easy to use". Also underlying their success is "easy to fix": if it goes wrong you take it to the Genius Bar and it will be fixed. Or, for a fee, they'll uplift it and fix it for you.) But what about strategy (b)?
5. Algorithmic pricing of airline tickets has been around for decades. Flights are scheduled up to a year in advance, between known end-points, with a well-known number of seats in each class. So why are there so many airline ticket types, and why do prices for a given ticket type on a given route vary depending on when you book them? The answer is that airlines have an incentive to ensure that there is a paying fare in every seat, but they also have an incentive to not discount seats if they can possibly sell them at a higher fare. So they've developed complex pricing algorithms that balance supply and demand internally and offer potential customers the highest spot price that the airline thinks the market will bear. More recently, mobile phone companies, cable TV companies, and other service suppliers offer us the "freedom" to build our own package of services, but the underlying pricing information is highly opaque.
—For example: cell tariffs include a set number of minutes per month to certain other types of phone. Non-inclusive minutes and text messages, and internet data are charged at a rate that varies depending on the price of the tariff selected, so that to identify which tariff is cheapest for a given customer it is necessary to work out their average inclusive and non-inclusive usage pattern and then use a spreadsheet. And comparisons between different cellcos are even harder because the basic components of a package may be defined differently.
—Alternatively, cable TV packages usually contain a mixture of channels that cut horizontally across different specialities, so that if you have a specific interest you will need to pay for the most expensive package on offer if you want to see all the channels covering that area.
—Because of the difficulty of navigating such complex pricing schemes, the majority of customers end up over-paying for services such as phone tariffs, airline tickets, and utilities.
6. Because of price competition, internet shopping is eating into the retail sector. High street retailers need to maintain expensive high street retail outlets; internet suppliers do not. In general the cost of delivery/fulfilment can be outsourced to the customer, or offset against the (higher) cost of operating retail premises. (Tesco et al wouldn't offer internet shopping and delivery services if it wasn't profitable to do so.)
—Price competition in retail is therefore an incentive for retailers to look for other means of sustaining profits.
7. How far can algorithmic pricing be pushed into the retail sector?
—Use of loyalty cards gives retailers huge volumes of data about individual customer purchasing habits. And CPU cycles are cheap. The bait on the hook is discounts on some products, or special offers available only to cardholders. These offers aren't free: if you use a loyalty card, the offers are your payment for surrendering a chunk of your privacy.
—Of course, online retailers like amazon get this information by default. (And compliance with local data protection laws often leads to a privacy "race to the bottom" in consequence.)
—I note that many personal computers are solid on a build-to-order basis: retail stores only supply a basic range (retail floor space is expensive) but offer the option of choosing accessories and spec on a machine that will then be delivered direct to the customer, or for collection. (This is similar to the way automobiles are traditionally sold, but pushes down the price floor for such a system by 1-2 orders of magnitude.)
—Cost of laptops may be dependent on the spot price of various components purchased in bulk (for example, RAM, hard disk, screen). Is it reasonable for consumer electronics vendors to therefore offer algorithmic pricing on their products? (For example, "the spot price of wholesale DRAM is trending down this week, so if you order your new laptop before Friday we'll give you a £7.52 credit towards the optional 4Gb memory upgrade! Hurry now! Offer ends Friday!")
—Other physical goods' cost of production may vary with other dependencies, e.g. fuel costs, plagues affecting crops, civil unrest/industrial action, floods hitting industrial parks in Thailand, etc.
8. I expect that we're probably going to see algorithmic pricing extended down the retail chain to include products that would normally only see that kind of trading on futures markets. Is it reasonable for a supermarket to offer customers futures deals? "We notice you bought 76Kg of potatoes last year. Currently potatoes cost £1/Kg, but the cost of fuel may cause this to rise, or there might be a famine. Would you like to pay us £80 by direct debit at 80p/week instead for a guarantee of up to 80Kg of potatoes with nothing extra to pay for the next year? For an extra 20p/week, we'll insure you against a potato famine driving prices up to £3/Kg for the next 100Kg you buy!"
—We've already seen some signs of this on amazon.com (with prices on offer to different customers varying for the same product, presumably on the basis of the customer's willingness to pay more for goods in prior transactions)
—Legal justification from the retailers would be "we're simply making an offer to sell at a unique price to this particular customer—what's your problem with it?" (The difference being that by putting a price sign on a rack of oranges in a supermarket, the supermarket is making the same offer to sell to everyone who walks past the rack.)
—The flip side is: it'd mean the end of transparent retail pricing, and a whole raft of new predatory practices coming in. (If you earn 50% more than average you can expect your grocery prices to begin creeping up, because your suppliers can infer what's in your wallet).
—How far can algorithmic pricing be extended? And what are the long-term implications if, in 10 years' time, the only way you can check the price of an item in a store is to point your camerphone at it—and the price will be tailored specifically to you and not available to anyone else?
(Part two of this prognostication will follow, when I feel like writing it up.)
 Obviously not all "goods" are equivalent. Books, for example, aren't always substitutable; a novel by Hannu Rajaniemi or Karl Schroeder might be an acceptable substitute for one by C. Stross, but a novel by Nora Roberts or a cookbook by Mrs Beeton most certainly isn't. And some superficially different goods may be substitutable depending on context; a return flight from EDI-AMS for a business meeting isn't substitutable for a return ticket EDI-CDG, but if the goal is a weekend vacation in a European capital city with tourist attractions it might be.