The Unexpected Payoffs of Employee “Eavesdropping”

“We’ve all played the telephone game before, right?” asks Paul Leonardi, the Reece Duca Professor of Technology Management at University of California, Santa Barbara. “As information moves across so many different people, it gets atrophied, and there’s a selection bias that occurs every time someone is listening and saying, ‘I think this is important, this is what I’m going to pass on.’”

Leonardi says that this kind of information sharing — typical for so many companies today — wastes a lot of talent.

“We end up in a situation,” he says, “where we have very complex organizations that are hiring smart people who have tremendous amounts of knowledge and insight in particular areas, who just don’t get the opportunity to share that with one another very often or directly.”

Leonardi is out to help companies avoid that. He is trying to figure out better ways for employees to share knowledge and information. The goal, he says, is to help employees know what and whom their colleagues know — so that they can move more quickly and smartly in all their work.

Leonardi’s research focuses on the use of social tools. “I devote my time to thinking about how managers can implement new technologies and leverage the informal social networks within their companies to make knowledge sharing happen more effectively,” he says.

In a conversation with David Kiron, executive editor for MIT Sloan Management Review‘s Big Idea Initiative, Leonardi details one of his projects — an experiment he did with employees at Discover Financial Services. That experiment showed an extraordinary upside to making conversations transparent by moving them away from emails and onto a Facebook-like internal social networking site.

Let’s talk about informal social networks and technology-enabled social networks, and what you’ve seen work for companies.

Let me start by making a distinction between networking, social networks and social networking sites. Because I think they get jumbled in people’s minds, and they’re key to what I want to talk to you about.

Networking, in my mind, is meeting people. It’s schmoozing. We go out and we meet people, and we figure out what it is we’ve got in common with them.

When we talk about social networks, we’re talking about the content and the structure of our connections with other people — and who the people that we’re connected to are connected to. If we want to trace how knowledge and information moves, we have to be able to trace who talks to whom about what kinds of topics. So networking is often the mechanism that we use to create our connections, and those are the individual links in a bigger social network.

Then there are social networking sites, which are technologies that help people to both network and to display their social networks. On a social networking site like Facebook, you often get a recommendation that you should be friends with this person because they’re friends with Joe, who you’re also friends with, or because you both, say, went to UC Santa Barbara together.

The information that sites like Facebook provides and collects gives people the fodder for networking.

Yes. But these sites also are really excellent at making social networks visible. If you and I were connected on Facebook, I would be able to see who your other connections were. I might talk to you on the phone (that’s networking), but after I hang up I have no sense of who else you talk to. But the social networking site Facebook begins to lay out for me what your social network looks like.

So when I think about social networking sites, what I think they really do is enable networking to happen in ways that couldn’t easily happen before, and they make visible people’s social networks that were often very difficult to see.

Let’s bring this to the business context and the transparency of employees’ social networks. You argue that there’s business value in making people’s social networks visible at work.

Absolutely. Let me give you a specific example. I did work with Discover about two years ago. You know Discover, big credit card company, headquartered just outside of Chicago. Discover’s got 15,000 employees, distributed across a number of sites throughout the world. They’re broken down into a lot of different divisions. The marketing division at Discover, for instance, has 13 different departments in it.

Discover wanted to try to provide an informal channel of communication for its 15,000 employees. So it decided to implement an enterprise social networking site called Jive. Jive looks like Facebook, for all intents and purposes: you’ve got a wall, you’ve got a profile, you’ve got newsfeeds. You can friend people — Jive calls them connections. You can send messages to people in private. You can post things to someone else’s wall that then anyone else can see.

Discover was wondering if and how these tools could be useful for communication, and they wanted to figure out how to get employees to use them. I began working with them to advise them on these issues, and the first thing I did was asked them to let me conduct an experiment.

Was your idea to make all these emails public somehow? To have people have those conversations on their social networking site walls?

Yes. What lots of other companies do to try to make what people are working on more visible is hold all-hands-on-deck meetings within particular divisions, where people talk about projects. That’s just not feasible for lots of reasons. Or they try to implement some knowledge-management tool where they ask people and maybe even require them, after they completed a specific assignment, to enter information into that tool about what they’ve completed. But people tend to not do a very good job with that. They don’t remember what they did, they fib a little bit because they don’t want to write out all the documentation.

My thought was, imagine if I could take all my work-related communications and just make them visible for everyone else to see. Now, anybody could go in and look at those emails and say, “oh, Paul talked about this; he must know something about that. So I’m going to go ask him that question.” Social networking sites allow for that, because when communication happens through them, that message becomes visible for other people to see.

Social networking sites as a way to increase transparency?

Yes. The other thing that those sites do is they make our networks translucent. So if you and I have a connection, Joe, he can see that you and I have a connection. He can see that we’re friends. He can see that we’ve exchanged a message. I say “translucent” because he can get a sense that there’s a connection between us, or a tie, but he doesn’t know the nature or the strength of that tie necessarily, right? He doesn’t know how good friends we are. He doesn’t know how long we’ve been in contact; that’s hard to tell.

But if messages become transparent and networks start to become translucent, then maybe he could start to develop a better awareness about who knows what within the organization, and who knows whom.

That was exactly the hypothesis that I talked Discover into letting me test.

So the hypothesis was that as messages became more visible and networks became translucent, people would have a better idea of who knows what and and who knows whom. How did you investigate this at Discover?

We picked two groups. One group was in the marketing division, the other group was in the operations division. The groups were both in a leadership rotation program, about 60 people each. They were in that program as a means of trying to develop a community of practice, and hopefully start sharing more information and knowledge across departments.

I picked these two groups because they were a relatively matched sample, same demographic characteristics, same years of work experience, all that sort of thing.

Discover let us give Jive to the marketing group, and not give Jive to the operations group, over the exact same time period.

Before Jive was implemented, we surveyed both sets to assess how accurately could they identify who knew what in their division. And then how accurately could they identify who knew whom in the division.

That “who knows whom” piece is important because a lot of times you find someone who has some kind of knowledge that you need, but you don’t know how to get access to it. It’s going to take a while for them to explain it to you, or, politically, it’s important for them to keep it — they don’t want to let it go. So having an entry point, someone to introduce you or give you access, can be really key. I call this “meta-knowledge”: the knowledge about who’s knowledgeable about what.

Both groups did about the same at the beginning — same level of accuracy within both groups.

So the marketing group gets Jive and uses it for six months. Over that same six months, operations didn’t use anything. And then we gave the exact same set of metrics and measures at the end to assess if there was a change over time in either of the two groups, and could it be attributed to Jive?

And what did you find? What was different six months later?

What we found was that, from before Jive to six months after using Jive, the marketing folks improved their accuracy at identifying who knows what by about 30%. And they improved their accuracy at identifying who knows whom by about 88%.

Wow! That’s quite an improvement.

Yes. And, over that same time period, there was no significant improvement in operations in people’s in ability to accurately identify who knows what or who knows whom.

That’s a pretty striking finding, I think. With Jive, people would post regularly or send messages to each other about their work, and other people could essentially eavesdrop on those communications. And in doing that, they could see, “Oh, this person knows this, and this person knows that, and this person talks to this person.” They basically built up a mental Rolodex of who knew what and who knows whom. And that mental Rolodex proved to be pretty accurate.

Those numbers are powerful — increases of 30% and 88%. Did you see a relationship between employee engagement and innovation that’s enabled by these social platforms?

Yes. This is the really exciting part, and where I think the biggest benefit is for companies. I would say it’s engagement that’s enabled by awareness. Employees are able to develop an awareness of what’s going on at their company outside of just their myopic view of the world, which is created because people specialize and staff employees into particular departments so we can manage them more efficiently. But it’s exactly this type of management and organization that makes it difficult for our employees to develop metaknowledge. That type of organization stops the informal flow of knowledge.

When daily communication becomes visible — messages become transparent and networks become translucent — through the use of social networking sites, employees can develop a better idea of what the lay of the land looks like: “These are the people that I might want to talk to, the people I might want to connect with, the people that can help me within my work, the people I can pool my resources with to do things together.”

In my experience, doing a decade’s worth of consulting with lots of big companies, this is the thing that people who are in the thick of the work are really good at doing. They’re really good at seeing how their project could be made better if it had this or that aspect to it. So, I think what these social tools can do is democratize some of that ability to make connections by pushing it down levels. People can see how their own work might be better informed by combining it with someone else’s work.

Now, there must be risks with greater transparency and translucency. What are the downsides to this path?

There are a couple risks. The first concern that many companies have is, “But what about proprietary information? What if that gets out?”

My answer to that is, number one, you don’t have to share everything on these tools. There’s still the appropriate channels for sharing things that would be dangerous to make public. And number two, I think it underestimates the power of a crowd. In the workplace, people get censured very quickly by their peers for saying things they shouldn’t be saying.

The second concern is about information overload and time waste. If I’m on there constantly reading what other people are saying to each other, am I not doing my work? Is it too much information, that I’m just going to be overloaded?

I think that’s really an open question. What was interesting with the people in the study that I did at Discover was that they were quite adamant when I interviewed them after the six months that they learned absolutely nothing from being on Jive.

Wait — the cohort who had used the online networking site thought they hadn’t learned anything by being on it?

That’s right. So if I said to them, “Did you learn anything about what other people knew while you were on Jive?” they would say “No.” Like, 100% of people said, “No, I didn’t learn anything.”

And then I would pull out the survey and I’d say, “But look, from time one to time two, your accuracy increased dramatically. What happened here? At the beginning, you didn’t know that Mark knew about customer acquisition, but at the end, you knew that Mark knew about customer acquisition. How did you learn that?”

And they would think about it for a while and they’d go, “Oh, well, you know, I remember I was on Jive one day kind of stumbling around, and I saw that Mark sent a message to Jane, and it said X, Y and Z. And so that kind of signaled to me that he knew something about market acquisition, and I guess I just must have remembered that.”

That is fascinating.

It’s really fascinating, right? This is the paradox of the system like this, I think, that people learned a lot by just becoming aware. They learned by proactively scanning the environment, without any idea that anything they were gleaning would be useful in the future.

That’s a different way to acquire knowledge than we typically are used to via search. Usually we search when we run into a problem, and we go looking for a solution, so we’re looking for something that can fill an immediate need.

These people were not doing that at all. They were slowly becoming aware of what and whom their coworkers knew and filing that metaknowledge away for later use. And so it didn’t register to them that they were actually learning stuff until the moment came when they were asked, carefully, in our survey, “What kinds of knowledge does so and so have?” and, “Who does so and so talk with regularly?” Then they could recall all of these little bits and pieces of information that added up to something pretty important.

People were learning without realizing they were learning.

Exactly. This gets to why it’s possible to overcome the question about information overload — it doesn’t actually take a whole lot of information processing to learn something.

Of course, if you think that you’re not learning anything, you may decide not to use the system at all. If you can’t recall an instance in which you actually learned something, you might say, “Being on a social networking site is a total waste of time,” and then you decide not to use it.

So I think one of the implications is that we’ve got to train people to a different mode of information acquisition. That there’s something to be said for just learning a very, very little bit about what is it that someone else might know. Because if you can just store that in memory, passively, there may come a point in time in the future where you run into a problem and you go, for instance, “If only I knew someone that could help me do semantic key encryption.” And then you think, “Oh, wait: I remember. John sent a message about this to somebody. He knows about this. I’m going to go talk to him.”

That’s a great example of how a network can be useful and have a big business payoff without being something that takes up too much time.

Yes. And what we found was this increased accuracy of metaknowledge — knowledge of “who knows what” and “who knows whom” was linked to reduced work duplication and increased innovation at the company. Employees stopped duplicating work that someone in another department had already done because they now had a more accurate sense of what projects others were working on. This began to save the company a lot of money. Also, because employees were able to see what things their coworkers knew, they were able to generate new ideas for products and services or process improvements by recombining knowledge. So, using these tools can, in some cases, help companies save money and make money by increasing innovation. That is very important.

I’d also add that probably the most important factor in keeping the risks of this kind of transparency down is the judicious use of that information by the people that are collecting it — the people who are essentially eavesdropping on each other.

For instance, if you and I were on this tool, I would learn that you know about how to publish good articles in MIT Sloan Management Review. If I now go and tell everyone in the world that I know that you know that, and you get bombarded with requests, then you’re probably not going to want to be sharing that kind of information again.

And so there really is a personal responsibility that needs to be placed on those people that are listening in, and some trust on your part that you’ve got a well-mannered audience out there that is not going to take advantage of what you put out there. How you create that is probably a cultural issue, is my guess. But how you create the kind of culture that has those kinds of norms, and that places value on those kinds of things? I’m not totally sure how to deal with that problem.

That is, in some respects, the $64,000 question.