Should online platforms share market data with sellers?
Online platforms can maximise profits through the judicious sharing of information and offering incentives to set prices that match its goals
This article is republished with permission from China Business Knowledge, the knowledge platform of Chinese University of Hong Kong (CUHK) Business School. You may access the original article here.
We live in the midst of the digital revolution, one in which, increasingly, traditional ways of doing things are being complemented or outright supplanted by online platforms and marketplaces. Nowadays, instead of going to the travel agents or hopping down to the shops, it’s popular for people to book a room as part of their travel plans through Airbnb or buy everything from everyday goods to books through Amazon.
As these online marketplaces connect buyers and sellers together, their success hinges on their ability to generate revenue, typically in the form of fixed commissions on successful transactions.
In most cases, platforms let sellers decide prices for their products, and consumers decide whether to buy from them or not. By their very nature, these platforms are typically able to collect vast amounts of data that allow them to predict with a fairly high degree of accuracy how much customers are willing to pay for goods and services advertised by individual sellers. And because they operate on the fixed commission model, they stand to make higher profits if sellers set higher prices. On the other hand, the individual sellers themselves have little incentive to listen to pricing recommendations set by platforms.
Under these conditions, what should these online platforms do to maximise their profits? That’s the question that a group of researchers sought to answer in a new study titled To Interfere or Not To Interfere: Information Revelation and Price-Setting Incentives in a Multiagent Learning Environment, which was co-written by Chen Hongfan, Assistant Professor of the Department of Decision Sciences and Managerial Economics at The Chinese University of Hong Kong (CUHK) Business School, in collaboration with Prof. John Birge and Prof. Amy Ward from the University of Chicago, as well as Prof. Bora Keskin from Duke University.
Too risky or too conservative?
The study sought to look at three types of business models employed by online platforms. For example, Airbnb provides estimated demand information for a given region to its hosts to give them an idea of how to best set prices. Upwork, the U.S.-based white-collar freelance platform, makes available information about the popularity of different tasks. Then there’s Amazon, which does not share any of its information with its sellers – a model that researchers refer to as a “do-nothing” policy.
While past studies have focused on the poor performance of this do-nothing policy, Prof. Chen and his co-authors discovered, through their modeling, that it did not necessarily lead to negative results all the time.
For instance, if two hosts on Airbnb, in seeking to rent out their respective properties in a given area, were unaware of their customers’ preferences, they may set the same price; otherwise, they may cut prices to induce a purchase if the information they were provided indicated that customers are not satisfied with the location of the area or other condition. Because they lowered prices, the platforms stand to earn lower revenue even in the event of a successful transaction. Prof. Chen explains that under this do-nothing policy, there is a chance sellers would set a price in the absence of information that would allow them to earn a higher fixed commission, than they otherwise would if these sellers had better knowledge of market demand. “It’s possible they just happened to make a decision in the platform’s favour,” he says.
Read more: Are companies illegally collecting third-party data to profile you?
While simple to implement, the researchers note that this do-nothing policy is inherently risky, as alternatively, they can trade off their informational advantage to allow sellers to set more accurate prices that reflect market demand. Under this second model, known as a reveal-and-incentivise policy, platforms (such as Airbnb under its “Plus” program) right from the beginning share relevant market information with sellers and offer incentives for sellers to set higher prices. Prof. Chen notes that this model is known to help improve market transparency and minimises revenue losses associated for the platform. “Sellers will know all information about the market demands, which can help them to formulate a better price-setting strategy,” he says.
However, this second policy is not without its flaws, chief of which is that it is considered highly conservative. “When platforms disclose all the information they have about the market, they give up the huge advantage they hold,” Prof. Chen says, adding that this may lead to negative outcomes. As an example, he says that in a market where the sellers have full and intimate knowledge of customer demand, they may choose to engage in fierce competition against one another, driving down prices and causing platforms to take a hit on their commission earnings.
Finding the right balance
Prof. Chen says this leads to a dilemma that platforms often face: Should they withhold information on the chance that their sellers pick prices that benefit the platform, or should they share what they know about the market to incentivise sellers to set higher prices? “When sellers have no idea about the market, it’s difficult for platforms to provide incentives because they are not sure how these incentives will affect sellers’ decisions,” says Prof. Chen. “It is only when sellers know the consequence of their pricing choices that incentives become effective.”
With a do-nothing policy considered excessively risky and the reveal-and-incentivise policy as likewise excessively conservative, the researchers propose that platforms may consider adopting a third alternative, which they refer to as a strategic-reveal-and incentivise policy, that combines the best features of the two previous approaches.
Subscribe to BusinessThink for the latest research, analysis and insights from UNSW Business School
Under this model, platforms provide individual sellers with incentives (such as lower commissions on sales) to set prices that benefit the former. At the same time, after collecting and analysing sufficient information on customer demands, they can choose to share or not to share information with sellers at different times. “This policy gives platforms time to gather and analyse useful information, to consider their future development plan, and then decide whether they want to disclose market information or not,” he says.
Finally, Prof. Chen notes the study’s findings are also applicable to online platforms and marketplaces in China, some of which have grown to rival or even eclipse their global counterparts, as they are to the rest of the world. “Platforms in China have similar business models about information revelation, and our findings should apply to them as well,” he says, while also reiterating that the aim of the research is to provide guidance for online platforms on maximising revenues. “We want to let platforms know that information is valuable, and that making good use of this data can help them perform better.”