Join our community of smart investors

AI’s hidden hand in rising prices

Consumers are getting hit by inflation at every turn and retailers have become increasingly proactive at mitigating their input costs
May 25, 2022

How much do you expect to pay for a pint of milk? This politician-vexing question is becoming trickier for consumers, too, as supermarkets raise their prices at the fastest rate in 13 years, and the Bank of England warns of more “apocalyptic” price spikes to come. 

If rampant cost inflation wasn't enough to disorient consumers on their own, the pricing equation has become more complex than ever with the rise of retailers using artificial intelligence (AI) to adjust prices of goods on a weekly, daily, or even hourly basis. They can incorporate a whole range of data into pricing decisions, including time of day, product expiry date, and weather, bringing to mind the seemingly eternal gambit of making it more expensive to buy an umbrella when it is raining. 

So-called ‘surge’ prices have existed for years in some sectors, encompassing Uber (US:UBER) journeys, hotel stays and restaurant ‘happy hours’. Data compiled for Investors' Chronicle by grocery price comparison site Trolley shows that the price of mundane everyday items like milk are now changing much more frequently than they used to - with UK supermarkets adjusting the price 61 times over the last week alone.

What began with Amazon (AMZN), which uses its swathes of consumer data to pioneer the trend in online shopping making an estimated 2.5mn price changes a day, or one change every 10 minutes per product, could be spreading. With cost inflation soaring, vendors of everything from clothing to second-hand cars are looking to pass on higher costs to consumers, and more of them are looking to AI to do it. 

“We call it the ‘inflation protector’,” said Pini Mandel, chief executive of dynamic pricing firm Quicklizard (IL:QLRD), whose software is used by the likes of Ikea, Philips, and Office Depot. “Retailers need the platform in order not to lose money, because costs are changing all the time on a daily basis, where in the past costs have changed only once every two to six months. Today, if you don't have an automation tool that helps you to manage that cost change, you can definitely be losing money,” said Mandel. 

He estimates that between 25 and 30 per cent of all retailers in Europe, including the UK, are already using dynamic pricing, with the largest uptake visible in sectors including consumer goods, grocers, consumer electronics, sporting goods, DIY and home improvement, and airlines.

Among grocers, cost inflation has already led to a rash of profit warnings from Sainsbury's (SBRY) and Tesco (TSCO) for the coming financial year, while US consumer bellwether Walmart (US:WMT) suffered its biggest one-day share price drop in 35 years after cutting earnings guidance last week. 

Dynamic pricing is not only being taken up in retail. Business secretary Kwasi Kwarteng has also called for “more continuous pricing, and more variation” for energy prices, charging households more for using electricity during periods of peak demand. 

Meanwhile, US-based pricing software maker Vendavo works with companies such as Dell (US:DELL), Boeing (US:BA) and Airbus (FR:AIR) on their business-to-business pricing, and vice president of product marketing Mitchell Lee said there had been a large increase in interest since the latest wave of cost inflation. “We have generations of managers and directors and even C-suite folks that have not really dealt with inflation of this magnitude before,” explained Lee.

In previous inflationary cycles, a company’s ability to raise prices has been paramount, noted Lee. “Price is your primary method of communicating with the market, and those companies that were able to manage that were markedly more successful than those that struggled with that.”

Lee is optimistic about the dynamic pricing’s standardisation over the next five years: “Looking around at CEOs and CFOs, [the comment will be] like ‘What do you mean, you have software for planning your warehouses and planning your production, but you don't have software for managing your prices? Did you miss something back in school?’”

Profit boost

Even in normal times, industry figures suggest that introducing dynamic pricing can increase retailers' revenues by up to 30 per cent, while driving profits up by 11 per cent.

Recent success stories include Levi’s (US:LEVI), which managed to raise its prices by an average of 10 per cent, without seeing a reduction in demand. The jeans brand credited AI, which it began using to determine promotions at the start of the pandemic, for helping to grow gross margins to 59.3 per cent at the end of 2021, compared with 55.7 per cent before the pandemic, while other fashion retailers such as Boohoo (BOO) and H&M have also implemented algorithmic pricing in their lines. 

Take-up is currently much higher in online retail than in physical stores. “We think that the gap will be closed in the next year or two,” said Mandel at Quicklizard, who calls the spread of dynamic pricing across traditional retailing “inevitable”.

This accords with retail strategy consultant Mark Pilkington’s view that traditional retailers will need to adapt to a future where business “equals data plus algorithms”. “There is a real question as to whether [traditional retailers] will be able to make the transition in time,” he wrote in his book Retail Therapy.

One of the issues is that it is much harder to implement frequent price changes in physical shops than it is online.

However, Marks and Spencer (MKS), Tesco, and Sainsbury’s have toyed with using electronic shelf labels in their physical stores in the past. These displays can change as quickly as online without incurring what economists call the "menu costs" associated with price changes – if a restaurant wishes to raise its prices, it will need to pay for the reprinting of all its menus.

It is especially significant for supermarkets that sell fresh products, since they already manage prices very carefully via promotions in order to make sure that they don't have to throw them out. Israel-based firm Wasteless has shown in a pilot study that its system of real-time price changes for fresh food based on expiry dates, rather than simply marking down the day before expiry, can result in a one-third reduction in food waste and 6.3 per cent higher revenues for supermarkets.

Traditional retailers’ experiments with dynamic pricing are likely to be kept under wraps for now, to avoid what one source described as a “PR nightmare”.

A quick search for dynamic pricing on social media turns up a litany of complaints against its most well-known proponents, ticketing giants and airline companies. Personalised pricing has proven to be one of the most controversial areas, with Amazon (US:AMZN) forced to issue an apology and ban per-customer price fluctuations in 2000 after an experiment with DVD sales kicked up a cloud of outrage.

“We don't see personalised pricing being served in the next five years," said Mandel. 

Meanwhile, the actual effect these algorithms are having is not fully clear. Research from Alexander MacKay, an assistant professor at Harvard Business School, found that dynamic pricing actually leads to less competition and higher product prices over time. This is because if an algorithm means that one retailer instantly matches its rivals’ price cuts, it takes away the incentive for a retailer to lower prices in the first place. The sectors in which dynamic pricing is already widespread – travel, entertainment, and transport – have been among the biggest contributors to inflation so far in 2022, according to the Office for National Statistics, though energy prices clearly bear the brunt of the blame here.

In any case, it is also not clear that dynamic pricing will always lift profits. One of the most baffling examples is in the used car marketplace where algorithmically-priced car dealer Cazoo (US:CZOO) has said that it could not "guarantee we will become profitable” in the future, despite the ongoing boom in demand for second-hand vehicles, where prices rose by 30 per cent last year. In May, it said that gross profit per unit almost halved since the previous quarter, to £124 per vehicle. UK second-hand car sellers such as Vertu Motors (VTU), which have more traditional sales systems, have seen profits soar on the back of higher prices. 

The need to change prices quickly has grown as cost inflation takes a bigger bite out of company margins, and AI algorithms can optimise the process of managing supply and demand. The question is whether finding the happy medium between margin protection and demand destruction needs a human touch. This new fraught retail environment should provide a clear answer.