Algorithms Can Save Networking from Being Business Card Roulette

Networking can be something of a crapshoot. Though there may be someone in the room you really should meet — someone whose acquaintance might help you out professionally — it’s often impossible to determine who precisely that is. And so we go to industry conferences and share business cards like shotgun spray, hoping that with a little luck we’ll make a useful connection. In large organizations, the same thing goes for networking at company functions; surely there’s someone in another department it would be useful to know, but when scanning the room at the holiday party it’s not always clear who that is.

If this seems like a trivial issue, it’s actually anything but. Innovation, a critical component to success for today’s companies, is about coming up with new ideas. And generating new ideas involves recombining old ones in new and interesting ways. In practice, that means mixing people together, so that the information they share can be recombined into something new. That’s why companies go to great lengths to design offices that encourage spontaneous encounters, and why dense cities, where people can bump into one another on the street, are more innovative than sparser suburbs.

It’s also, at a theoretical level, why networking matters, and why a new academic paper on networking at conferences is so interesting. Researchers from the UK, Japan, and Italy set out to improve how networking happened at a scientific conference they were organizing. They asked all attendees to share which of the other attendees they knew, their own area of expertise, and subject areas or methods they were interested in learning. With this information, the researchers ran two rounds of “speed-dating” using algorithms designed to maximize the formation of new relationships. In the first round, participants met with others who were “far” from them in the conference social network, as well as “far” in terms of expertise. In the second, they looked for those who were “far” from each other in the network and who had expertise that the other was interested in learning.

This algorithmic speed dating ensured that the scientists were recombined in new ways, strengthening ties in the network that wouldn’t have existed and increasing the odds that each participant would make useful connections. It’s too early to tell if this approach yields quantifiable benefits, but the post-conference survey response was extremely positive. Notably, more than half the respondents indicated that “potential new collaborations were emerging from discussions at the meeting.”

“Of course, we have no objective baseline,” said Rafael Carazo Salas of University of Cambridge, one of the authors. But that doesn’t mean there couldn’t be one in the future. When I spoke with Carazo Salas he mentioned using Twitter or LinkedIn to quantify how the algorithmic matching compares to traditional networking in creating new connections. “In the case of science,” he continued, “you could go to grants co-submitted [or] patent applications filed,” to measure not just the effects on the network but the eventual impact on collaboration.

If such a measure of success could be found, it’s likely that this algorithmic approach to meeting people could transform networking both between and within businesses, much like it has for the online dating world. When you sign up for an Eventbrite event, an algorithm could suggest the people you’d most benefit from meeting there. Something similar could happen inside companies, with algorithms scanning email, intranets, and project management tools to suggest collaborators who might improve a project before it even gets off the ground.

No doubt some people will react negatively to the idea of algorithms deciding whom we should meet and collaborate with — not least the business development professionals whose networking acumen would potentially be less valuable. But society has grown used to it in the dating realm, and the conference goers in Carazo Salas’ experiment actually reacted more enthusiastically when they were told their pairings were the result of computation. Moreover, this kind of algorithmic matching already shapes the way we connect in the online world.

It’s still early days for this algorithmic approach to matching people, at least face-to-face and for the purposes of business and scientific productivity. But Carazo Salas and his colleagues are exploring the commercial applications of their experiment. “It’s all about innovation,” he told me. “By mixing things that are different, probably new things are going to come out.”

This entry was posted in Leadership. Bookmark the permalink.

Comments are closed.