��«����Ƶ

Online dating platforms such Tinder, match.com, and Bumble are a source of rich study in developing a successful matching platform. In new research, Amit Basu and Sree Bhaskaran of ��«����Ƶ Cox’s ITOM Department, with coauthor Rajiv Mukherjee, break down what features and decisions make an online matching platform work. From ten years of research, the paper offers insights and an actionable path for building effective matching platforms.

The three colleagues started researching online matching platforms around 2015 after they observed some interesting patterns on some of these platforms. “I had been looking at this issue of authentication in online settings for many years,” Basu added, thinking it seemed important in matching platforms too. The performance of matching platforms was already apparent when they looked at online dating platforms. 

Matching platforms differ from transaction platforms like Etsy and Amazon Marketplaces as a successful match requires acceptance by both parties. In many matching platforms for business or personal relationships, the platform does not make matches itself (as in organ donor platforms, for instance). Instead, it presents potential matches, and people make choices. This is where things are tricky. “Given that information on the platform is provided voluntarily, mostly by participants: how do you know who you're actually dealing with?” asks Basu. 

The researchers examined whether authentication, or verification, is always valuable on these platforms. "Our research revealed that it's not always optimal for a matching platform to offer authentication," says Basu. While authentication can reduce uncertainty, it may require the disclosure of significant personal information. They consulted with Match.com's management during their investigation. Match.com did not offer much authentication then, and its executives argued that this was because user information disclosure for authentication would create friction that could scare away some potential users. "Through our models, we found that this reasoning makes sense under certain conditions, which we were able to identify," Basu explained, and this led to their first paper in Management Science.

Even with authentication, compatibility between parties isn't guaranteed. "Unlike authentication, where you can discuss a match in objective terms, compatibility is subjective and works differently," Basu says. It requires mutual alignment—each side must be compatible with the other. "Unlike the authentication issue, which was a vertically differentiated dimension, compatibility represents a horizontally differentiated dimension," Basu adds. The tradeoffs that arise when platforms also guide users in identifying compatible matches led to a second paper in Management Science.

Having a broad set of results and insights in the context of matching platforms, the authors wanted to offer business practitioners the benefits of their findings, the current paper.

Making matches

In this paper, the authors highlight what distinguishes online matching platforms from traditional platforms, and reveal insights from their learnings and the evolution of online dating platforms. “We focus on the decentralized nature of matching in such platforms, which is a key distinctive factor,” says Bhaskaran. And the bilateral nature of matching, successful only if there is mutual acceptance by both match-seekers, is also key. This is especially true in dating markets, where compatibility and authentication are important, says Bhaskaran.

Although Match.com, eHarmony, Tinder and Bumble are all dating platforms, each of them pursued very different paths in attracting users and developing services offerings. For example, Match.com focuses on the casual dating space, and eHarmony serves the long-term relationship market. User priorities on these platforms are very different, and this has significant implications for platform design. Bhaskaran notes, “By analyzing online dating platforms, practitioners can understand these implications—learning how their features relate to how businesses transact with each other.” 

For a long while, Tinder, cofounded by ��«����Ƶ alumna Whitney Herd, had been on a tear with respect to growth and scale. The goal of platform-oriented companies is to achieve scale. The faster you can scale, the more network effects created, and the greater the value you provide. “This endogenously generates more of the scale you are seeking,” notes Bhaskaran. He continued, “But as they scaled up quickly, the quality of the matches and reliability of user profiles on Tinder became less certain: you just didn't know whom you were dealing with.”

Bumble, also founded by Whitney Herd, was formed from the lessons learned from Tinder.  Herd understood why women hesitate to be on online dating platforms. For women, security and authentication are important issues. “She turned the tables completely,” Bhaskaran says. “She shifted it to a very different kind of decentralized matching, where the onus of accepting or rejecting a match was completely on the woman's side.” Authentication features became paramount. “As time went by, and the quality of matches on improved, the network effects started kicking in,” says Bhaskaran. “The quality of the matches started attracting more users.”

Supply chain and business platforms

Global sourcing and global supply chains are a reality today, says Basu. Given that companies have to consider sources outside their local geography, the value of matching platforms becomes very significant. “But that value proposition will depend upon what kind of product you're trying to source,” explains Basu. “Our research provides a decision framework to decide when services such as search, counseling and authentication make sense, and how to factor in one’s level of capability in providing these services.”

Given the twin challenges of authentication and compatibility, authentication can be difficult and complex. However, authentication is typically “a one-time thing,” says Basu, used when first joining. However, assessment of compatibility is relevant each time a potential match is considered. 

In a supply chain setting, in sourcing commodities for example, compatibility is not a major concern. The key challenge is authentication of the supplier’s ability to reliably deliver the resource. Thus, a matching platform in the commodities space can support fast matches between authenticated participants/firms relatively easily. On the other hand, a platform operating in a complex product space may be better off focusing on rich compatibility counseling capabilities based on a deep understanding of industry; the technology features can provide high-value matching services even at relatively low scale, particularly if it also offers an authentication service.  

Advances in AI and algorithms can help with the tasks of authentication and compatibility assessment. Basu notes that the implementation of those services effectively is where algorithms can achieve a high level of quality. “Ultimately, we're addressing whether to offer certain levels of services or not on a platform,” says Basu. That problem is not solved by AI because it's an economic problem, a strategic choice problem, he concludes. 

Today, developing a viable matching platform that attracts significant funding requires a more deliberate approach. "We believe our insights and research will enable these platforms to succeed not through trial and error, but with sound design principles to guide them," says Basu. Maybe even some matches made in heaven.


The paper “Strategic Choices for Matching Platforms,” by Amit Basu, Chair of ITOM, ��«����Ƶ Methodist University’s Cox School of Business; Sreekumar Bhaskaran of ��«����Ƶ Methodist University’s Cox School of Business; and Rajiv Mukherjee of Texas A&M University of Texas, is forthcoming in Management and Business Review.

Written by Jennifer Warren.