The widespread use of pricing algorithms is reshaping the nature of competition in online markets and potentially driving up the prices of retail goods, according to recent research.
These automated, price-adjusting software programs may also be catching the eye of government regulators and antitrust authorities, who fear they could ultimately harm consumers by raising prices above typical competitive levels.
It doesn’t seem too long ago when a price change was a major strategic decision for companies, requiring extensive data analysis, management consensus, coordination with advertising schedules, and other factors in an era when computers were helpful but not critical to pricing strategy. The result: A price change was more an annual or semiannual event. But these days, when companies can analyze consumer data and use technology to raise or lower prices in the blink of an eye, changes can be made not just once a year but multiple times daily.
“What we show, theoretically, is that (algorithmic competition) leads to higher profits for both firms.”
Enter the rise of pricing algorithms, where software monitors prices posted by competitors and makes adjustments using parameters developed by the company’s marketers and strategists.
“They want to react to changing demand and supply conditions,” says study author Alexander J. MacKay, an assistant professor of business administration at Harvard Business School who studies competition, including pricing, demand, and market structure.
“But it can also have the side effect of changing the nature of competition, especially when your rivals know you’re reacting to whatever price changes they might make.”
When rivals change prices
In their recent research paper Competition in Pricing Algorithms, MacKay and co-author Zach Brown, assistant professor of economics at the University of Michigan, demonstrate that adopting advanced pricing technology can allow a firm to seize an edge relative to its competitors. Algorithms that frequently update website prices based on competitors’ prices result in a softening of retail price competition.
The study focuses on algorithms that are predicated on rival pricing and does not examine those that target pricing for specific regions or consumer groups.
Online retail pricing algorithms, whether developed in-house or supplied by a third party, commonly use competitor prices—information that is freely available online—as an input for calculating price changes. In a given industry or category, the most technologically sophisticated company is able to continually undercut its rivals by updating prices more frequently than its competition, in some cases (e.g., Amazon) as often as every 15 minutes.
As a result, any rival has less of an incentive to lower its price, because it knows that the sophisticated company will quickly undercut any price change it might make. Further, rivals have less of an incentive to adopt high-frequency pricing technology, as matching the technology of the leading firm would be costly and could lead to both companies charging less, explains MacKay.
One surprise finding of the research was that firms with inferior pricing technology also benefit from this arrangement, so investments in pricing technology by large firms may turn out to benefit small businesses.
“What we show, theoretically, is that (algorithmic competition) leads to higher profits for both firms,” MacKay says. “It’s just that the smaller firm, or the inferior-technology firm, is at a persistent disadvantage relative to the one with superior technology. But as long as you’re okay being in second place, it’s a pretty good place to be.”
Consumers end up paying more
To study how pricing algorithms affect competition, MacKay and Brown collected detailed pricing data from five large, multicategory retailers selling the same over-the-counter allergy drugs online over a period of 18 months. Because of differences in operational infrastructure, the companies varied in how often they updated prices on their websites: two changed their prices hourly or faster, one updated its prices daily, and two updated prices weekly.
Through econometric analysis, the researchers estimated demand for the various allergy products, which included several brands and package quantities, while modeling algorithmic price competition by the firms.
When compared with a simulation in which each firm exhibited traditional (symmetric) price-setting behavior, the algorithmic prices were 5.2 percent higher on average and variable profits increased by 9.6 percent, according to the study. The model also predicted that algorithmic competition resulted in a modest decline of .9 percent in quantity purchased over the study period.
Extrapolating the results of their simulation from allergy medicine to all personal care products sold online by the five firms implies that customers spent an additional $300 million annually as a result of pricing algorithms, out of a total of about $6 billion in e-commerce revenue for the category.
Advice for managers
The study illustrates how emerging technology is affecting and, in some cases, upending traditional ways of thinking about and conducting business. Historically, says MacKay, economists have tended to assume that firms have equal capacity to set prices and can therefore change their prices at the same time.
“…Customers spent an additional $300 million annually as a result of pricing algorithms, out of a total of about $6 billion in e-commerce revenue for the category.”
“If we’re in a world where that’s not happening, what are the implications? One of the implications is that having better technology is going to give you an edge.”
For managers, gaining competitive advantage through advanced pricing technology is not as straightforward as it might seem.
“It’s important to consider your relative position in the competitive landscape,” MacKay cautions. “If you have a chance to get an edge on your rival in terms of pricing technology, that can be quite beneficial. But if your rivals already have highly sophisticated and fast pricing technology, then maybe there’s not much of a benefit, and maybe you should look at investing in other ways to build your competitive advantage.”
Implementing and maintaining advanced pricing technology requires substantial resources, both in terms of computing power and human capital. Even companies that contract with third-party providers to run pricing algorithms on their websites need significant data storage capacity, processing speeds, and expertise. In addition, the assumptions that underlie pricing algorithms require revisiting as conditions change.
To ensure that institutional knowledge is embedded in pricing algorithms, managers can also consider hybrid solutions that offer suggested prices, enabling quicker pricing updates than would otherwise be possible, but allowing the firm to retain more control, says MacKay.
“It might take an expert a week to redo their analysis and update prices as conditions change, but if you have the factors that go into that analysis codified as part of a pricing formula, that can be updated very, very quickly,” he says. “That’s why the frequency element is one of the things we focus on, because it is really fundamental.”
Should regulators step in?
As Google and Facebook face threats of antitrust action, issues related to the digital economy, including online pricing, are drawing increased attention. Algorithms will only become more prevalent, and their use will likely receive greater scrutiny as policymakers try to understand the implications for competition.
In their paper, MacKay and Brown suggest that one potential policy solution would be restricting the use of competitor pricing as an input for automated price-setting. Algorithms could still use external factors that measure demand and supply to help keep prices in line with the market, the researchers write.
Although this may seem like the best solution, “I don’t know if we’ll get there,” MacKay says. “That is the policy that would be suggested by our analysis, but it involves companies turning over and making transparent their code, which they’re not used to doing.”
“Establishing a digital regulator would allow enforcement to adapt more quickly to evolving technology, and it would have the benefit of likely being more predictable for firms.”
A related suggestion is to establish a regulatory body made up of digital market experts, similar to the Federal Trade Commission, who would be tasked with monitoring digital price competition and protecting consumers from the adverse effects of algorithms, MacKay says.
“We don’t want to take too much power away from firms, but it is important to consider strategic effects that may harm consumers,” he says.
Because of the black-box nature of some pricing algorithms, pursuing consumer protection through antitrust lawsuits would be difficult, MacKay says, and could take decades. Establishing a digital regulator would allow enforcement to adapt more quickly to evolving technology, and it would have the benefit of likely being more predictable for firms.
“Many economists think a digital regulator is necessary, not only for issues of pricing, but also privacy, data portability, and platform competition. There are all sorts of elements of the digital marketplace that our existing legal frameworks are not particularly well-suited to deal with.”
About the Author
Kristen Senz is a writer and social media editor for Harvard Business School Working Knowledge.
[Image: Kanoke_46]
Related Reading
- If Your Customers Don't Care What You Charge, What Should You Charge?
- Rethinking Business Strategy in the Age of AI
- Don't Turn Your Marketing Function Over to AI Just Yet
As a consumer, do you like frequent prices changes?
Share your insights below.