Boundary Spanners: The Employees Holding Your Company Together Without Anyone Noticing

When Sita resigned from her job as a solutions consultant, her manager expected a small gap. Within two months, three engineering releases missed client requirements, the sales team stopped trusting delivery estimates, and the Kathmandu team felt cut off from the London office. Nobody had written down what Sita did. She was the person who translated between everyone else.

Who boundary spanners are

A boundary spanner is someone who links a group to other groups: they carry information, translate language and priorities, negotiate between conflicting goals, and represent one side to the other. They can sit in formal roles — account managers, product managers, HR business partners — but many of the most important spanners are informal: the engineer who understands finance, the support agent everyone in product calls, the bilingual coordinator between offices.

What the research shows

Why HR systems miss them

Boundary work is almost perfectly designed to be invisible to traditional performance management. Its results show up in other people's metrics: a smoother release, a renewed client, a faster decision. The spanner's own dashboard looks average. In calibration meetings, the person who made five teams better often loses to the person who delivered one visible project.

How to find them

1. Run a simple organisational network analysis: ask 'Who do you go to for information you can't get in your team?'
2. Look at who is copied, tagged, and invited across teams in collaboration tools — with privacy safeguards and aggregate reporting.
3. Ask managers in adjacent teams, not only the line manager, for input in reviews.
4. Review exit interviews for people whose departure caused cross-team slowdowns.

How to support and protect them

Limits and caveats

• Not every well-connected person adds value; some become bottlenecks or gatekeepers who restrict flow.
• Network data raises privacy concerns — use aggregate analysis and be transparent with employees.
• The strength of these effects varies by industry and organisational design.

What the research found

Michael Tushman (1977) studied research and development labs and found that information from outside a unit rarely reached it directly. It flowed through a small number of 'boundary-spanning' individuals who were well connected both inside and outside. Ancona and Caldwell (1992) later showed that product teams whose members actively managed external relationships — scouting information, representing the team, and coordinating with other groups — performed better than inward-focused teams.

The cost of this role is real. Cross, Rebele and Grant (2016), writing in Harvard Business Review, reported that in many organisations 20% to 35% of value-adding collaborations came from only 3% to 5% of employees. These people are often the most requested colleagues and the most at risk of burnout — a pattern the authors called collaborative overload.

Four kinds of spanners

How to find them

1. Ask a simple network question: 'Who do you go to for information you cannot get in your own team?' Names that appear across many teams are likely spanners.
2. Look at meeting and messaging patterns in aggregate, with privacy safeguards, to see who connects groups.
3. Review exit risk: check pay, promotion history, and workload for the people you identify.
4. Ask managers what would break if this person left for a month.

Protect and reward them

• Name the role: include cross-team work in job descriptions and objectives.
• Count it in reviews, so spanning is not unpaid extra work.
• Build backups: pair spanners with a second person in key relationships.
• Protect time: limit meeting load and help spanners say no to low-value requests.

• Key spanners in each function are identified.
• Their workload and retention risk are reviewed twice a year.
• Cross-team contribution appears in performance criteria.
• No critical relationship depends on only one person.

Worked case: the connector who was rated 'average'

A senior engineer spends a large part of her week helping sales understand what the product can do, explaining customer problems to engineers, and sorting out disagreements between two product teams. Her own code output is lower than her peers'. In reviews she is rated 'meets expectations'. When she moves to another company, sales deals take longer, two teams start building the same feature, and a key customer escalates.

Her real contribution was the connections between groups — which no one had measured. Research on networks, such as Cross and Parker's work on organisational network analysis, has repeatedly found that a small number of people carry a large share of cross-team collaboration, and that this work is often invisible in formal reviews.

How to find your boundary spanners

• Ask a simple survey question: 'Who do you go to when you need something from another team?' The same names will come up again and again.
• Look at who is invited to meetings across several departments.
• Ask managers who would be hardest to replace for reasons that are not in their job description.

Protecting them from burnout

Connectors are often overloaded, because every request finds its way to them. Research by Cross, Rebele and Grant (Harvard Business Review, 2016) described 'collaborative overload': a small group carrying a large share of collaboration work and burning out. The goal is not to stop people connecting, but to share the load and recognise it.

• Write cross-team work into their goals so it counts.
• Teach others to answer common requests so the connector is not the only route.
• Check their workload at least every quarter.
• Name them in succession plans — the connections leave with them.

Common mistakes

• Measuring only individual output in roles where connection is the value.
• Promoting connectors into roles that remove their cross-team contact.
• Assuming that because collaboration is 'everyone's job', no one needs to be credited for it.