The return of inflation is placing retailers in a terrible quandary. Should they act now to raise prices and protect margins, at the risk of yielding business to cutthroat competitors? If so, by how much? Should they treat the current trend as a temporary blip, or consider it the permanent new normal?
Pricing decisions have long been among the most difficult aspects of retailing, more art than science. Price too high and you lose fickle customers; too low, and you end up selling products at below cost.
With inflation on the rise, raw material costs soaring, supply lines becoming ensnared and labor at a premium, the pricing conundrum today poses a greater challenge to retailers than ever before. The result is often a series of hastily imposed price actions that confuse shoppers and weaken brand loyalty.
The key for retailers seeking a coherent and considered approach to pricing is being able to achieve visibility into behavior at the end of the supply chain: the consumer. And that’s where analytics and modern-day predictive systems come into the picture.
The traditional process of price adjustment might begin with notification from a vendor that its costs are going up. The retailer is forced to pass the increase down the chain, but the precise amount and timing of that action can be devilishly tough to fix. It needs to take into account specific products, geographies, customer profiles and the possible response of other merchants. In retail sectors where margins are especially thin, the difference between success and failure can come down to a few cents.
Modern-day analytics, driven by artificial intelligence, can parse a flood of information, including which products can or cannot survive price increases, then strike the proper balance, says Matt Pavich, senior director of retail innovation with Revionics, a vendor of price-optimizations software. In certain cases that might even result in discounts that the retailer discovers it can afford to offer.
Predictive analytics allow merchants and their vendors to run scenarios that assess the potential impact of a set increase on particular products. The amount of data that AI takes into account “could be enormous,” Pavich says. “It’s very important to understand what consumers are willing to do, based on their shopping behavior — what they voted for in the past with their wallets.”
The AI engine can also help to ensure that the retailer isn’t making pricing decisions that unfairly impact some segments of the market — for example, women’s razors versus men’s. “You’re not only able to apply science to arrive at optimal pricing decisions by product, channel, customer and store,” says Pavich, “you can put in rules to make sure it’s working from a basic logic.”
The predictive aspect comes into play when retailers can anticipate shifts in demand before they fully impact the market. That allows time to model out various scenarios and their effect on consumer behavior. At the same time, retailers need to be aware of changes taking place upstream, such as factory shutdowns due to the pandemic, or rising raw materials cost.
Pricing can also be a tool for shifting demand from one product or category to another, based on supply and inventory levels, Pavich says.
Within this deck of observable events, inflation provides the wild card. “It’s one of those things where a lot of really smart people aren’t sure when it will end, or how high it will get,” says Pavich. For the past decade, merchandisers have enjoyed the luxury of not worrying about a significant rise in the cost of money. Now, they’re in need of sophisticated analytical tools to cope with a factor over which they have no control.
When the problem is so complex as to be impossible for humans to process all of the data and its implications, can science be trusted to make the right decisions? The modern-day retail industry appears to think so. Pavich cites a recent study by Gartner of 23 AI use cases for retailers. Leading them all in perceived value was demand forecasting, and how it impacts pricing and promotions.
The best AI tools for the job are those that learn with experience, and adjust their conclusions in accordance with receiving new information, such as pricing history. “The more price changes you take,” says Pavich, “the smarter the system gets.”
But even the “smartest” AI is a supplement to, not a replacement for, human intelligence. At least for the foreseeable future, it will be people making the final decision on pricing — equipped with tools of unprecedented sophistication and depth.
Timely, incisive articles delivered directly to your inbox.