An Unusual Application Using Organizational Network Data

A case study on determining contribution and compensation

Since I first learned about organizational networks, I’ve been fascinated by their potential applications. While an org chart can be quite informative, they don’t capture the full picture of how work gets done in organizations. They don’t tell you the dotted line and informal relationships between different employees.

I’ve followed the work of people like Rob Cross and Michael Arena in organizational network analysis (ONA). I am fully convinced of the value of their social capital view of organizations rather than a more limited, human capital view.

A human capital view sees employees as entities to be optimized, in the vein of Frederick Taylor’s principles of scientific management or what Douglas MacGregor called Theory X. Employees are simply a composite of skills, attributes, and competencies without regard to their relationships with others.

A social capital view recognizes that humans are more than just isolated individuals; that the relationships and connections we have with each other allow for greater collective success as an organization beyond the sum of individual contributions. In short, a human capital view is about what you know while a social capital view also includes who you know.

Organizational network analysis uses network data to understand and improve organizations. Network data can represent communication channels, relationships, and connections within an organization and can be collected in a variety of ways.

Let me share the coolest application of organizational network analysis that I’ve been able to work with. More details can be found in this presentation I gave on the topic a while back.

What was the business problem we were trying to solve? In this situation, we were working with a peer-to-peer evaluation system that was an input into compensation decisions for employees. Rather than relying solely on a boss to determine the compensation, we gathered input from teammates and others on someone’s performance. The original process of collecting this input was cumbersome and time-consuming. Some employees would spend days doing these evaluations and participating in meetings to discuss the evaluations.

The search for a more efficient solution started with a simple idea. What if we could allow the employees to tell us those who know the value of their work? Rather than relying on leaders with a more limited view to decide, why not give the individual that ability? Each employee could select between 5-20 other people that they wanted to evaluate them. We used these selections to generate a full network map across an entire organization of more than 10,000 people. As you can imagine, the participation rate and coverage of the organization are quite high, much higher than you would get in a voluntary survey. You can see the network in this map, which I generated using Gephi open-source software. Each dot represents an employee and the lines represent a selection of another employee to evaluate them.

Network map of more than 9,000 employees and their 118,000 connections. Areas of black lines represent more dense clusters of employees and colors represent major business units.

From this network map, we used an algorithm to generate more than 250,000 pairs of employees to be evaluated against each other. So if Joe and Jane selected me to evaluate them, I would see them in a pair. In that pair, I’m asked to select who made the biggest contribution in the past year, using a 5-point scale. I also have the option to rank them equally. The time required to complete the evaluation was reduced from hours to minutes as I can easily make those side-by-side comparisons in my given pairs. We capped the number of pairs to 40 or 50 to avoid survey fatigue and for the time savings we were seeking.

From the data on the pairs, we used a statistical model to analyze the data and generate rankings which then fed the process of determining compensation for those who made the biggest contributions. Bosses or managers did not make this assessment of contribution, rather it was made by those who worked most closely with the person.

The key insight of a network view is recognizing that everyone has a unique network of people they work with that is not limited by geography, division, team, function, or business. The vast number of connections allows us to compare anyone to each other from the newest employee to the CEO. With a custom selection made by each employee, you leverage that employee’s network to get a holistic view of their contribution.

A network lens is a powerful way to view organizations. People are more than just individuals. They are humans with relationships and connections. Those relationships and connections are valuable in today’s knowledge economy and ONA is an incredible way to tap into insights about networks.