Implementing data-driven variable pricing is challenging, but Corus has an answer.
When shopping online, most consumers are at least vaguely aware that dynamic pricing is a ubiquitous part of the experience – and not just for airfare, but for all manner of goods. When the earliest accounts of the practice first came to light at Amazon in the early 2000s, it was sufficiently shocking that the company felt compelled to issue a public apology. Today, of course, this is the stuff of everyday press releases, and no longer the exclusive preserve of Amazon.
Physical retailers, in their long-running (and steadily losing) battle against online sellers, rightly point to this as an Achilles heel. Rapid price testing and optimization technologies are as easy to embed in a digital storefront as they are difficult – and in many cases, impossible – to apply in a physical setting.
However, even though stores, restaurants, and other offline consumer businesses cannot feasibly conduct thousands or millions of price experiments each day, there are still powerful means available to implement data-driven variable pricing.
At Corus we’ve pioneered one such approach, which departs from the traditional use of price elasticity modeling as the fundamental basis of price optimization. This decision was partly driven by necessity: most companies’ transaction data simply doesn’t include a wide enough range of price points in order to establish a demand curve that’s useful for this purpose. Unlike Amazon, say, which can easily test extreme highs and lows of price (and everything in between), physical stores capture a much narrower set of historical price observations. This is a poor basis for extrapolating the effect of significant price changes.
To address this data deficiency, our method is grounded in economic demand modeling rather than price elasticity modeling. This approach essentially inverts the focus of analysis by asking, “what is the sales volume we should expect for any given product?” This might sound like an indirect path to price optimization, but it has the virtue of contextualizing price among many other variables that holistically explain demand. For instance, to what extent does the surrounding demography of a store affecting buying behavior? What about the presence of nearby competitors? How much can “micro” decisions, like store layout, cause sales to outperform macro industry trends?
Sales volumes are obviously a function of a much broader swath of factors than price alone. Because physical sellers aren’t able to adjust prices as rapidly as their counterparts in the digital realm, it’s all the more important that they optimize price on the basis of the fullest possible picture of everything that might be driving customer behavior. This can often include quantitative research and buyer feedback, which our team is expert at collecting and tailoring for this very purpose.
In the aggregate, the tools and methods we’ve developed can serve as an equalizer for some of the intrinsic advantages long enjoyed by online channels.