People Analytics Rising.

Patterns of human communication at workplace reveals so much more than we thought possible.

Tata Consultancy Services (TCS) in a recent survey found that only 5% of all data science investments currently is routed to HR. An insignificant percentage to enhance and maintain the company's greatest resource — its people.

A recent study published by Harvard Business Review states that 70% companies state people analytics to be a top priority, which is in stark contrast to studies by Tata Consultancy Services. The clue might lie in another study by Deloitte which exposed that only 9% of companies despite of interest in people analytics has any clue regarding metrics that drive performance within their companies.

The difference between the companies that benefit the most from people analytics and those that struggle, including a vast majority of Fortune 500 companies might be a little heard offshoot of people analytics called relational analytics.

People analytics beyond attribute analytics

While most companies spend time trying to find motivation and metrics of individual employees, companies that succeed greatly at employing people analytics are busy solving the puzzle of team dynamics — how interplay of relationships within an organization drives performance and long term gains for business as well as employees.

While attributes — both static (color of skin, race, gender) and dynamic (education, age, commute time, days of absence and the likes) are important, it can only shed light on insights that when combined with the relational aspects of each employee.

Decades of research convincingly show that the studying relationships employees have with one another —relational analytics, combined with attributes such as racial, gender diversity and education can explain employee performance to interesting accuracy.

The raw material for relational analytics is derived from the emails, work chat and other online communication between people. The digital exhaust of a company.

Data that is simply being overlooked can be mined for precious gems that shed light on what makes people (and teams and organizations) work!

According to a study by Paul Leonardi and Noshi Contractor, the key to successful relational analytics is finding “Structural signatures”: Patterns of data that correlate to varying levels of performance. Their study goes onto state six signatures of Relational Analytics.

1. Ideation

This signature predicts who is more likely to come up with good ideas.

By studying patterns of communication within an organization you can identify employees who communicate the most outside of his core group. These employees are statistically more likely to come up with better ideas than employees who only communicate within their core groups.

The yellow turban clad member has a network that spans multiple teams.

Those people with multiple links can be said as ‘brokers’: people who have knowledge about the company on a wider level and can use findings from one corner of the company to problems faced at another. These people combine good ideas from two or more groups to create even better ideas.

There is a measure developed by sociologists to indicate who is more likely to come up with better ideas known as constraints.

Constraints captures how limited a person is when gathering unique information. Study after study of people from diverse fields such as lawyers, software engineers and financial analysts have gone on to show that people with low-constraints, that is, people who aren’t bound by a tight, close knight network in communication is more likely to come up with better and more novel ideas valuable to the organization.

No single attribute came even close to having such strong correlation for predicting ideation compared to having a low constraint.

This study has been taken to heart by many organizations. Infact, the whole idea behind open plan offices is to improve collaboration among employees from various departments. Google has taken to designing their offices from ground up providing for “casual collision”.

Here is a snippet of a conversation between David Radcliffe, a Google civil engineer and Paul Goldberger:

The layout of bent rectangles, then, emerged out of the company’s insistence on a floor plan that would maximize what Radcliffe called “casual collisions of the work force.” No employee in the 1.1-million-square-foot complex [Googleplex] will be more than a two-and-a-half-minute walk from any other, according to Radcliffe. “You can’t schedule innovation,” he said. “We want to create opportunities for people to have ideas and be able to turn to others right there and say, ‘What do you think of this?’”
An office designed for ‘casual collision’ Image: smartfurniture.com

2. Influence

The influence signature predicts which employees are in better position influence change in the organization.

Research shows that people with higher influence are not people with most number of connections or even the senior leadership of the company, but people whose connections are better connected. Studies show people with stronger (not larger) networks help spread ideas fart her (instead of merely fastest). These individuals are said to show higher aggregate prominence.

The yellow turban clad team member has the strongest influence as his networks are more strongly interconnected than that of his peers.

People with higher aggregate prominence have strong connections, who in turn have strong connections of themselves. This is infact how google’s search algorithm PageRank works.

A study at a medical device company by Paul Leonardi, Professor of Technological Management, University of California, shows that it was not people who were rated ‘most influential’ by employees who were able to drive greater compliance among employees of a new policy, but those who were identified with relational analytics to have higher aggregate prominence.

While the ‘most influential’ succeeded to influence 15% of employees to adopt the new policy, those who who were identified by relational analytics to be highly influential was able to influence more than 75% compliance.

3. Efficiency

The efficiency signature unlike the previous individual signatures is a team signature and predicts which team is best poised to get a job done efficiently. And studies show it is not about identifying people with relevant skills and experience (superior attributes).

The efficiency signature is formed when teams exhibit two individual characteristics: Internal density — how closely connected the individual members in a group are and 2) external range — showing how wide the external connections of the team members are, giving it access to an even greater array of outside resources.

The team in yellow turban have high internal chemistry and great external networks. Making for a potentially high efficiency team.

Relational analytics measures efficiency, in short, by the chemistry between the team members and its ability to draw on outside resources and expertise.

The internal density is critical for building trust, taking risks and reaching consensus on important issues, quickly. While external range, the second social variable enables a teams access to a large network of experts and positions of power that will provide for much needed information and resources that will allow deadlines to be hit.

4. Innovation

Innovation Signature predicts which teams are more likely to come up with innovations that sell. This is predicted by identifying teams with low internal density, but high external range.

The folks in blue turban have low internal density but high external range. Potentially a prime innovative team.

Low internal density translates to higher possibility of diverse ideas and more productive debates. While a wide external range means networks that will help them with allocation of resources and buy-in of their innovations.

It is also vital to note here that innovation signature doesn’t equal ideation signature, for one ideation signature identifies individuals instead of teams as in innovation signature. More importantly creating a team of ideators doesn’t create for an innovative team, infact research shows performance levels dipping when attempted.

5. Siloes

This, unlike previously discussed individual and team signatures is an organizational signature.

A siloed organization have multiple departments with low external range.

Neither of the 3 teams in the given structural signature has good external range. While team white is completely excluded, team red and green are linked by single nodes.

Low external range translates to redundancy of work and expenditures, weaker employee morale and lower rates of innovation. This is highly detrimental to an organization and relational analytics enables companies to identify silos within the company to take corrective measures.

We can assess the degree to which an organization is soiled by measuring its modularity. Modularity put simply is the ratio between internal internal density to external range. Researchers have pegged 5:1 as the modularity for organizations that are detrimentally siloed.

Siloed organization with modularity upwards of 5:1 can even take corrective measures and improve collaboration and corporation with corrective measures such as redesigning of office, creating shared events and more.

A non profit organization with modularity with 13:1 was able to break down on siloes quick by appointing certain employees from each department as liaisons and creating themed meeting once a week, forcing even junior level employees whose work related to the theme to come and merge into the wider organization.

6. Vulnerability

Relational Analytics help identify vulnerabilities in the structural signature of an organization with the help of vulnerability signature.

Vulnerability signature identifies individuals who act as lone connections to a key external stakeholder.

The crucial supplier is connected by a single point of contact, the yellow turban clad.

A single one of these employees can cost the company big time. It is therefore crucial to identify the signature and retain the employee until the company can put a succession plan in place which will ensure robustness of the networks.

Networks are robust if you can maintain connections even when you randomly remove certain nodes.

The trouble here is that in most cases these integral nodes are junior level/average employees. The company wouldn’t know what hit them until the lone node quit or took a long vacation. It is not that they aren’t irreplaceable, they aren’t just backed up.

Relational analytics shed light on vulnerabilities within the company in order to maintain robustness and prevent nasty surprises.

The Challenges

Once you identify the structural signatures in your organization you can act on it. The actionables as we saw are never complex, infact, it is usually simple-managerial level decisions such as conducting cross departmental meetings, retaining a key employee or assigning a certain task to a certain team.

Why, then, hasn’t most companies ventured out deep enough into relational analytics for performance management? One of the problems is that network analysis currently do not portrait patterns that predict performance.

Secondly, organizations do not have information systems in place that can capture and identify structural signatures in relational data. But all companies do have access to their digital exhaust: email trails, slack messages, project assignments on trello, teams created on microsoft teams, where the platforms record the interactions. This data can be analyzed to unearth structural signatures which can then be acted upon.

Another direct challenge is privacy, and this needs to be handled carefully by the H.R. Consent letters allowing for analyzing of their digital trials or even content of communication (in extreme cases to pair relational analytics with machine learning) needs to be obtained for a friction free operation of relational analysis.

Way Forward

A tested helpful strategy to counter privacy concern is to share the data collected about employees on a regular basis with them. This helps create trust and transparency — number one hedge against privacy concerns.

It is also to be noted that studies have shown surveys don’t work too well in unearthing structural signatures. People don’t really know or they lie naming people they should be communicating with instead of who they are really communicating with.

The best way is to observe digital trails on regular intervals owing to the dynamic nature of human connections and pair it with machine learning — in order to combine the relational information gathered from multiple sources and figure out patterns emerging with the help of individual attribute data.

To get the best results off people analytics in your organization, you need to go up and beyond just attribute analysis and bring in relational analytics to figure out structural signatures present in the organization.

In conclusion, when you augment your decision making process with relational data to assign tasks, predict possibility of organization achieving its goals and plan succession, you will achieve a more satisfied, healthier and more productive workforce.


If you enjoyed reading please give some claps so that others can find it too.👏👏👏

Subscribe to get latest posts on feed. 🌏

This story is published in The Startup, Medium’s largest entrepreneurship publication followed by +394,714 people.

Subscribe to receive our top stories here.