Organizational network analysis: overcoming common pitfalls

Experts argue how organizational network analysis has fallen short of its promise, but an organization's approach makes a big difference.

Link to article

In a 2019 article in Forbes, Dr. Andrés Cardona and Dr. Laura Weis, two experts in business psychology and social sciences, give a critical perspective of the potential of network analysis in organizations. While they believe there is value in network analytics to increase engagement, productivity, creativity, and innovation, they believe it has fallen short of it’s promise. They argue that 1) digital traces of collaboration such as email, chat, or comments provide only a proxy for social connectivity; they miss interpersonal trust and emotional predictability, conditions necessary for psychological safety, 2) technologies to overcome these limitations such as sociometric badges or sensors invade employee privacy, 3) perceptions of relationships may be more relevant than actual relationships, and 4) employees may not welcome the insights and feel threatened by them. 

While psychological safety may be a better benchmark for collaboration than simple digital communication, it’s not clear whether emails and chat frequency is actually an output measure for psychological safety. People who feel a level of trust and safety are able to communicate with each other in a state of flow without feeling like they must proofread every word. Even if that’s not the case, sentiment analysis of the communication data can be used to bridge this question of trust. 

In a different approach to gathering trust data in a more scientific way, the company Humanyze, founded in 2010 out of MIT, has created ID sensors that track employees throughout the day. These sensors can pick up what they say and how they interact with each other. While this brings up privacy issues, Humayze says it uses only metadata, or patterns about the data, rather than the actual content (1). Additionally, the company comes into the office of a client, explains how it works, asks employees to opt in, and has employees sign a disclaimer. Humanyze generally gets 90 percent opt-in participation (2). The company worked with several branches of a European retail bank, and using their technology, helped the bank better integrate new hires and make physical workspaces more conducive to networking. This resulted in improved loan sales by 11% versus a control group (3)

As Dr. Cardona and Dr. Weis point out, humans are known to have major deficits when it comes to accurately perceiving and recalling social relations around them. Thus, if perceptions, rather than realities predict how we feel and act, they question whether digital traces are useful to understand employee connections. This is why it’s important to use survey data of one’s feelings, attitudes, and perceptions along with behavioral data. Neither alone can provide a complete picture of the network. 

Finally, and maybe most importantly, it’s understandable that employees may be hesitant to adopt these network tools, technology, and practices and accept their insights. What this point brings up is the importance of an organization’s approach to organizational network analysis and people analytics in general. Network analysis can offer valuable, actionable insights, but employees must be treated as partners rather than pieces of data. 

Employers should apply network analytics only if their goal is to improve the organization and its people. The goal should be to help its employees create more and better ideas, further their influence, and feel more efficient in their work. Organizations should realize that employees may feel betrayed by the data, and thus, they must contextualize the entire initiative properly with a lot of transparency along the way. People analytics professionals should empower managers and employees to review their own data and be incentivized to make decisions that will further their own career objectives while helping the company. With this frame and the continued advancement of technology to discover the nuances in our communication, network analysis will eventually become a crucial element to how organizations optimize their structures to support their employees as well as their customers. 

Sources:

  1. “Humanyze Elements Platform.” Humanyze. Accessed April 12, 2020. https://www.humanyze.com/elements/.
  2. Carey, Scott. “US Biometrics Startup Humanyze Is Bringing Its Employee Tracking Badges to the UK.” Techworld, February 3, 2017. https://www.techworld.com/startups/us-startup-humanyze-is-bringing-its-employee-tracking-badges-uk-3653590/.
  3. “Re:Work – Mapping Employee Chitchat Can Reveal Information Blockages.” Google. Google. Accessed April 12, 2020. https://rework.withgoogle.com/blog/mapping-employee-interactions-reveals-networks/.

Previous:

Building Bridges with AI

Next:

A New Opportunity for Cost Cutting through People Analytics: Employee Retention

1 thought on “Organizational network analysis: overcoming common pitfalls

  1. Interesting post about the adoption and implementation of network analysis. I wonder what other tools management has at its disposal to encourage adoption of these invasive platforms. In addition to what you mentioned about transparency, I think a top down approach would speed up implementation. If top level management wears badges and gets feedback (up to the CEO getting feedback from the board), I think an organization is less likely to lose trust with employees and get buy in. Pairing promotions with higher data collection requirements could be seen as a reasonable tradeoff (i.e. once you reach a senior vice president level, you wear a badge). This emphasis on top level management could also help with reputation issues (like what we see with Amazon tracking data on warehouse workers).

Leave a comment