Pricing Strategy and Algorithm Aversion

The idea of implementing dynamic pricing—that is, trusting an algorithm to make pricing decisions—can be intimidating. Most people understand that pricing is a powerful lever that can be used to increase profitability, and understand that “with great power comes great responsibility.” Abdicating pricing strategy to a computer, then, might feel risky.

However, as a recent study published by the University of Pennsylvania’s Wharton School notes, “Forecasts made by evidence-based algorithms are more accurate than forecasts made by humans. This empirical regularity, documented by decades of research, has been observed in many different domains…when choosing between the judgments of an evidence-based algorithm and a human, it is wise to opt for the algorithm.” In their paper titled “Overcoming Algorithm Aversion,” study authors Dietvorst, Simmons, and Massey write, “Algorithm aversion represents a major challenge for any organization interested in making accurate forecasts and good decisions.”

Fortunately, their research points to a solution: “we found that letting people adjust an algorithm’s forecasts increases their likelihood of using the algorithm, improves their forecasting performance, heightens their tolerance of errors, and increases their confidence in the algorithm.”

Here at Digonex, the ability to adjust our algorithm’s dynamic pricing recommendations is built into our SEATS pricing portal. Users are encouraged to accept our prices, but also have the ability to override proposed price changes, either leaving current prices in place or shifting prices by a different amount than suggested by the pricing algorithm. Additionally, minimum and maximum prices can be entered for each product being priced, constraining the algorithm to working only within boundaries defined by the user.

Algorithms and humans may each make mistakes, but generally, humans trust their own forecasting over an algorithm’s, even when the algorithm can be shown to be more accurate. Allowing humans to sometimes substitute their own judgments when they think an algorithm might be making an error—even if the human’s judgment turns out to be wrong—ultimately leads to trusting the algorithm more, resulting in a more profitable pricing strategy.