Degrees of separation: After all, it is who you know

With government’s highly systematized pay grades, titles, ranks, organizational charts and job descriptions, you might think agency executives have a pretty good handle on how work gets done and who their key players are. But in many cases, you’d be wrong.

That’s because an org chart will not show that the department veteran who found her way through four reassignments, weathered a half-dozen regime changes and now sits in a small regional office is the person decision-makers at headquarters should consult when they need a long view and some institutional perspective. And it won’t show that the policy and compliance units barely interact — even though they are supposed to — because the two bosses can’t stand each other.

The desire to get at those inner truths is the reason why a growing number of agencies are following the cue of many in the private sector and turning to organizational network analysis (ONA), a technique for mapping and evaluating interpersonal relationships in the workplace — or the informal networks. It can also map decision-making processes and the way information flows within groups.

“It’s like an X-ray that helps reveal what’s going on in an organization under the surface,” said Valdis Krebs, chief scientist at Orgnet.com, an ONA consulting firm and software developer.

That kind of insight can be especially valuable for government agencies, which often lack the private sector’s ready-made metrics for evaluating performance, such as profits or the number of widgets produced. Instead, information is the lifeblood of many government operations, and collaboration is the circulatory system. But any number of problems can hinder an organization’s effectiveness.

“ONA is a good tool to understand the human factor in our communication networks so that we can improve productivity or understand why we fail in some areas,” said Leni Oman, director of the Office of Research and Library Services at the Washington State Department of Transportation, in describing her organization’s use of ONA.

However, as is often the case, the biggest investment with ONA is not necessarily the fees for consulting services, data gathering or software but the management commitment needed to convert the findings into action and an ongoing process of change, measurement and improvement.

Despite ONA’s promise, that attention and follow-through have often been in short supply in government projects, to the frustration of those who have seen firsthand just how valuable a workplace X-ray can be.

Why it matters

Government executives face a host of challenges that ONA can help them better understand and address. Some issues are long-standing, such as the organizational silos that can make collaboration across geographic, functional or jurisdictional lines difficult.

ONA can help leaders break down those silos by identifying key employees in different departments or offices who are brokers, those who have lots of connections to others, said Rob Cross, an associate professor of management at the University of Virginia, an ONA consultant and the author of two books on the subject. And better connecting those brokers by creating a semi-formal group called a community of practice can result in increased communication among all employees.

Interest in improving communication among multidisciplinary networks was what prompted Washington state transportation officials to consider ONA, Oman said.

Agencies can also benefit by recruiting brokers to be the lead adopters when they deploy a new collaborative tool, such as an internal social network application. “You can get a 30 percent or so better uptake by targeting them versus diffusing it blindly,” Cross said. “It’s much more efficient than just throwing a collaborative tool out there and saying, ‘Good luck with it.’”

In addition, ONA can help agencies address issues that have recently become more problematic, such as the brain drain resulting from baby boomer retirements. As agencies are discovering, the knowledge inside a person’s head is not the only thing that walks out the door on retirement day.

“More hurtful in many cases are the networks they disrupt, the way they hold things together,” Cross said.

The Defense Intelligence Agency conducted several ONA evaluations a few years ago and found a strong correlation between a person’s network diversity and higher performance.

By profiling what the brokers and high performers are doing and what their networks look like for a given role, agency leaders can use targeted mechanisms, such as rotating assignments and mentoring programs, to help up-and-comers replicate seasoned networks.

“Companies and agencies have done this forever on an expertise or competency basis, but they are just starting to pick this up on the network basis,” Cross said.

Government downsizing is creating new problems that ONA can help address. For example, as the remaining employees take on new responsibilities, they can become over-connected, and decision-making bottlenecks can crop up, Cross said. Also, well-functioning collaborative links between departments can be disrupted by the departure or reassignment of a few key employees.

The Virginia Department of Transportation, which saw its workforce shrink from 9,000 to 7,500 employees in the past two years, plans to do a comprehensive ONA this summer involving about 5,000 employees, said Maureen Hammer, the department’s knowledge management officer. Virginia officials have already used ONA for two smaller projects.

“We’ve got a lot of new people in a lot of new roles, so our leadership wants a snapshot to see where the problems are and where we need to prioritize our efforts for leadership training or setting up communities of practice,” Hammer said. “Then we’ll go back later and see if what we’re doing is working.”

The fundamentals

The data that is the basis for an ONA and its graphical network maps can be collected by conducting employee surveys; tracking the use of online tools such as e-mail, wikis and other collaboration systems; or a combination of those methods.

Transportation officials in Virginia and Washington worked with Cross on their ONA projects and used his group’s questionnaires and analysis and mapping software. Cross said an initial project can cost about $35,000 to $75,000, depending on the scope. Krebs’ firm Orgnet.com likewise offers consulting services and an analysis and visualization software tool called InFlow.

Once data about interactions is collected and used to create network maps, analysts must put the findings into perspective while taking into account the circumstances of the group under review. For example, in one of the two groups studied in Washington, most of the leaders had relatively few connections to others, which seemed odd given their roles. However, their responsibilities reached far beyond the group being studied, so their peripheral position in that context was not deemed a concern.

The results of an initial ONA are often eye-opening, Cross said. He routinely asks who leaders believe are the key players or top talent in their organizations and compares those answers to the people that the network analysis reveals to be the central players. He said the two lists typically have an overlap of less than 40 percent.

Ray Horoho made a similar discovery when he did a network analysis shortly after he was appointed director of administration and resources at the Army’s Office of the Deputy Chief of Staff for Operations, Plans and Training about three years ago. (He is now retired from the military and serving as senior human capital executive at the Center for Human Capital Innovation.) The previous director had a top-heavy leadership style, and Horoho wanted to understand who the key players were and how work got done before he made any changes.

“I wanted to find out in this very stovepiped organization who talks to who and who doesn’t talk to who,” Horoho said. “Also, when I go forward to make changes, who are the folks who have the ear and trust in the organization outside the line-chart hierarchy so I can get their help with buy-in.”

Surprisingly enough, the analysis revealed that one of the lowest ranking individuals in the organization was one of the most important people on the staff. Horoho worked with that employee to better understand the organization’s dynamics and help promote his agenda for change.

The hurdles

In an ideal scenario, an organization would use the findings from an ONA to craft a set of interventions, give those changes some time to work, then do a follow-up analysis to measure their effectiveness and fine-tune further action, said Shane Brown, an assistant professor at Washington State University who conducted the two ONAs for the state’s transportation department.

That approach takes a steady commitment in terms of management attention and funding, which is by no means easy in the public sector these days. Washington transportation officials were pleased with their ONAs and have used the results to make some organizational tweaks and take steps to improve communication. But unfortunately, time constraints and tight budgets have prevented them from following up further or analyzing other groups.

Interestingly, an issue identified in one of the ONAs eventually affected the group’s ability to do more with the study results. The analysis identified three people as critical to how information moved in that network. All three have since moved on, diminishing, as expected, that group’s communication effectiveness, Oman said.

After some early success, one of the two ONA projects at Virginia’s transportation department was similarly derailed by budget cuts. The other project was more of a test proof of ONA for a one-time project.

Horoho said ONA was extremely useful in helping him get his footing and make some needed changes in the office he took over, but he left his position before ONA could become a regular tool that others used to guide business process improvement in that office or elsewhere in the organization.

Another set of challenges relates to the surveys that workers fill out. Washington state officials expected their questionnaires to take about 15 to 20 minutes to complete, but they ended up taking closer to an hour, which hampered participation. In the future, the survey needs to be shorter, Oman said.

Employees might also resist participating or answering honestly if they fear how the results will be used. ONA surveys can include questions that ask respondents to rate co-workers’ effectiveness in providing the information they need to do their jobs. Horoho said that in hindsight he probably launched his ONA too soon after taking over, before he had employees’ trust.

Krebs said that because of the difficulties with surveys, his clients are using them less frequently and are instead analyzing traffic data from e-mail and other collaboration systems.

Executives are unlikely to order major reorganizations as a result of an ONA alone, Krebs said. But an ONA can be indispensable for spotting budding areas of concern and coming up with plans to keep them from becoming bigger problems later.