How to Eliminate Information Silos

(by Toby E. Stuart, Sarah Jane Gilbert – Harvard Business School)

Working Knowledge, Q&A

Although many companies aspire to promote easy interaction and coordination across departments, office locations, and pay scales, the “boundaryless” organization – like the paperless office – hasn’t materialized.

The corporate silo is alive and well.

This observation is supported in a recent working paper, “Communication (and Coordination?) in a Modern, Complex Organization,” written by Harvard Business School postdoctoral fellow Adam M. Kleinbaum, and professors Toby E. Stuart and Michael L. Tushman.

In an unnamed company with over 100,000 employees, the team analyzed over 100 million e-mails and 60 million electronic calendar entries over a three-month period. The results provide an unmatched look inside the “black box” that hides what Stuart calls the “soft wiring” of previously invisible social networks.

In this Q&A, Stuart says the team was taken aback by the lack of communication across the organization. In short, most people tended to communicate with others in their own group or with peers. Among the exceptions: women.

“We were surprised by how little interaction occurs across three major boundaries: the strategic business unit, the organizational function, and the geographic office location,” Stuart says.

Although the research doesn’t try to answer why corporate silos are so difficult to tear down, Stuart hopes the data will help managers understand, pinpoint, and remove bottlenecks within their own organizations.

Sarah Jane Gilbert: What led you to study communications in corporations? Where did the idea come from?

Toby Stuart: For the first time in history, it is possible to actually analyze the “soft wiring” or largely complete social networks inside organizations. This is feasible because e-mail and other forms of electronic interaction have become the dominant means of communication within companies.

Our particular interest is in how organizations coordinate across their many internal divisions (e.g., functions, business units, geographic offices, pay grades, job ranks). Studying electronic communication networks enables new insight into this fundamental question.

Q: Your research examines a massive dataset that includes more than 100 million e-mails and 60 million electronic calendar entries over a three-month period. How was all this organized?

A: This particular study is of a very large company of more than 100,000 employees, but the process is not as daunting as it seems. Of course, the numbers indicate the richness of these data; they truly offer an unprecedented window into coordination and interaction patterns inside firms. For many organizations, these data are readily available on server logs. On our end, a dataset this size poses no major obstacles to analysis. In fact, its size enables us to be more confident in our findings.

Q: With the increased use of electronic communication, is there an impact on communication structures when we start to lose the human element of interaction?

A: Of course this is possible, but one of the intriguing findings from our analysis is that electronic and face-to-face interactions are near perfect complements. In other words, people talk to the very same people they e-mail. As electronic collaboration technologies further develop, this may change. For now, e-mail interactions seem to reinforce human relations.

Q: What influence do organizational structure, management hierarchy, and physical space have on an organization’s internal communication network?

A: Huge influences! Although every leader at some point in their careers becomes frustrated with the difficulties of coordinating behavior across organizational units, we were surprised by how little interaction occurs across three major boundaries: the strategic business unit, the organizational function, and the geographic office location.

Our analysis indicates that two people who are in the same SBU, function, and office interact about 1,000 times more frequently than two people at the company who are in different business units, functions, and offices, but are otherwise similar. Practically speaking, this means that there is very little interaction across these boundaries.

Q: Did you find gender differences?

A: Yes. We found that people exhibit a slight tendency to communicate with others of the same gender, but this effect is smaller than we would have guessed. However, we found that within-gender interaction patterns are more pronounced among executives. Surprisingly, we also found that women were more likely than men to play a key “boundary spanning” role in the company.

One can think of it like this. If we were to randomly remove an employee from this company’s communication network, the odds that this action would cause a communication breakdown between two units of the company are higher if the employee happened to be female rather than male. Women are more likely than men to link otherwise non-communicating groups of people.

Q: Were there any unexpected findings in the research?

A: We were surprised by a number of things:

First, that organizational units were such potent directors of internal communication.

Second, in this company at least, junior executives, women, and members of the salesforce were the key actors in bridging the silos in the firm.

Third, communication patterns were extremely hierarchical: Executives, middle managers, and rank-and-file employees all communicated extensively within their own levels, but there were virtually no cross-pay-grade interactions in the firm (other than mass e-mails and between administrative assistants and the managers they support).

Q: What can managers take away from your study and apply to business communication and coordination in their companies?

A: Potentially, many things. First, the type of analysis we conduct can offer a great deal of insight into the soft wiring of the firm. It can identify the key personnel, functions, divisions, and so forth that bridge the islands of interaction in the firm. It uncovers pairs of organizational units that fail to interact in a meaningful way, or isolated units that are not integrated into the social fabric of the broader organization. And it can illustrate whether lateral coordinating mechanisms put into place by senior management are taking hold in the organization.

Moreover, as we replicate analyses like this one in a variety of different settings, we will establish benchmarks that can be used to compare one organization with another.

Q: What are you working on now?

A: A variety of extensions of this project. We have two other research sites that have provided complete electronic communication data, one at a leading biopharmaceuticals company and a second at a global media and professional services firm. One of these follows senior leaders through a period of reorganization, and the other contains detailed information on the spatial layouts of buildings, the reporting structure inside the firm, and job performance measures.

These projects will allow us to address questions surrounding how individual managers’ (and scientists’) positions in the communication stream influence their job performance, as well as to begin to look at how communication patterns influence the performance of organizational units.

Additionally, we want to better understand the finding that junior executives, women, and members of the salesforce were the key actors in bridging the silos. We’re looking now at individual career histories and the ways in which it shapes people’s communication patterns to determine why those people – and not others – play such important bridging roles. We think this work will give insight into how individuals can plot the course for more successful careers and how firms can guide the career paths of their employees to create a better-connected social structure that will be better able to implement their strategies.

About the author

Sarah Jane Gilbert is a product manager at Harvard Business School.