A recent Harvard Business Review blog claimed that most HR data are bad data because we are terrible at assessing the performance of those we work with. Here’s another reason: whereas most HR data are collected about individuals, it’s the performance of their networks that really matters.
Sometimes the emphasis can be dangerously shallow, like when Yahoo came to grief with “stack ranking”. Under the policy, every manager forces their team members into performance “buckets”, where the bottom 15% of the team must, by definition, miss some or all of their goals, regardless of their performance. The result? To avoid competition, good performers don’t want to work with other good performers.
If you take an interest in innovation, you’ll share my reaction: insane.
Organisations should actively concern themselves with network-centric questions. Who shares information effectively? Who can connect diverse ideas? Who can see and promote opportunities? Who has strong relationships with key external stakeholders? Who engages and energises others? How can bottlenecks of information and decision-making be alleviated?
It turns out we’re terrible at answering these questions too when we rely on our perceptions. When it comes to identifying top connectors in an organisation, the overlap between management’s perceptions and a formal network analysis is as little as 30-40%.
Which makes organizational network analysis all the more useful in performance management.
A brief tutorial on organizational network analysis
1. The survey method. Employees are invited to complete a survey about their relationships with other employees, along dimensions such as information, trust and energy. An example question might be: “Please indicate the degree to which you typically turn to each person below for information to get your work done.”
Rob Cross and colleagues (including researchers at IBM) have been analysing network patterns for over 10 years. Here is their description of high performing teams:
On a network pattern level it turns out to be very important that teams not evolve into “cliques” or subgroups, that they not over-rely on a given person (either expert or leader), and that they manage a balance of external ties bridging to key stakeholders. (Cross et al, California Management Review, Summer 2008)
2. The electronic sensor method. MIT Media Lab developed electronic badges, worn around the neck, that could automatically collect data on an individual’s communication behaviour: their tone of voice, body language, who they talked to, how much, etc.
Alex Pentland and colleagues found communication patterns to be the most important predictor of team success, and:
Not only that, but they are as significant as all the other factors—individual intelligence, personality, skill, and the substance of discussions—combined. (Pentland, Harvard Business Review, April 2012)
That communication finding alone should be enough to question the HR data we produce, and how we think about organisational performance.
