What Emails Reveal About Your Performance At Work

We all spend an inordinate amount of time at work communicating with others. And we each have our little tricks. Some of us respond so fast it’s like a real-time system; others wait and do it when they have time. Some need “nudges” to respond, while others respond on their own.

What does all this mean? Does this behavior correlate with our performance at work?

Well, the answer is yes. Genpact, one of the largest and most successful global professional services firm (78,000 employees generating over $2.6 Billion in revenue), has figured this out.

After analyzing months of communication patterns using messaging metadata (data about the messages, not the messages themselves), the company can now statistically prove that certain types of communication behavior directly correlates to business performance. In fact, using employee communication data with a Deep Learning Model, Genpact can predict “Rockstar” performers with 74% accuracy. (This process works for emails, slack messages, skype messages, etc.)

So yes, your communication patterns are strongly correlated with your performance. What did we learn?

Some Background:  What is Organizational Network Analysis?

ONA is a slightly technical domain where consultants identify “who is talking with who” to figure out how organizations work. Tools for ONA include surveys, systems that monitor email traffic, and more complex applications that look at all types of communication (email, text messages, feedback, conversations). 

There are hundreds of tools that do this and consulting firms like Deloitte, McKinsey, and others have become experts. Thanks to lots of new products in the market today (Microsoft Workplace Analytics, Trustsphere, Keencorp, Yva.ai, and others) you can do this yourself. The discipline has been studied for many years and most social networking companies like LinkedIn, Facebook, Twitter, and the New York Times do this type of analysis to understand how different affinity groups communicate, cluster, and compare with each other. 

Here, for example, is a view of an ONA analysis to look at the impact of gender on decision-making. One can see, for example, that women are less involved in strategic decision-making but are actively involved in idea-sharing and emotional support. (Not a huge surprise.)

In the corporate world, the discipline is growing like crazy. As organizations become flatter and more interconnected, executives want to increase innovation, understand how teams work, identify top leaders, and look for ways to improve sales, service, and operational performance. 

One company, for example, found out that sales teams that operate in a more hierarchical way (management is involved in all deals) underperform those which are more empowered. So they changed the management structure.

Microsoft’s tool, MyAnalytics, can give you personal tips on productivity (you spend 25% more time in meetings than your peers, etc). This helps you decide how to spend your time better. 

One company, for example, told me that their ONA study showed that the highest performing engineers have many more meetings and relationships than others. This led the HR team to rearrange the engineering building and move the cafeteria across the campus. This forced the engineers to walk around more and meet others in the company.

ONA Vendors are learning to use this data in many interesting ways. One, Trustsphere, can go as far as identifying trust (how many people forward or listen to this person), impact (how quickly do people respond to this person), collaboration (how much traffic there is within a team), and engagement (how an individual’s communication compares to others). These analyses are all based on metadata: if you add sensing of mood and tone of emails themselves, these products can become highly valuable sensing systems to understand how an organization works. 

Note how the following analysis shows indicates how one employee is onboarding much faster than others, and another is well behind.

Now let’s get back to the topic of your own email performance: what did Genpact learn?

The Genpact Study

Praful Tickoo, the head of people analytics at Genpact, has been working with MIT to study the communication patterns of the company’s top 650 leaders. His findings were astounding: a 74% statistical correlation between communication patterns and the highest levels of individual performance (using a 9-box performance process). 

What did they find? The highest performing leaders use simpler words to communicate, they respond faster, and they communicate more often. In other words, they are more engaged, more efficient, and more action-oriented.

How did they do this? Genpact used the tool Condor from MIT, which converts raw communication data into ONA signals. The team then used Open Source Machine Learning / Deep Learning tools such as Python, Tensorflow to look for correlations, and used Machine Learning / Deep Learning to create predictive models. Interestingly, the distinction between the “Rockstar” group and the rest was mainly on parameters such as network, reach, responsiveness and need for follow-ups.  (Note: Genpact does not read emails, they only look at metadata and communication patterns.)

I asked Praful is this “cause” or “effect?” In other words, perhaps some of this behavior is because they are more senior?  He assured me that all these people are at equivalent senior levels, so these characteristics highly differentiate leaders.

How Top Performing Leaders Simplify Work

Many of these characteristics are logical and easy to understand.  In a company like Genpact, which is very dynamic, high performing leaders must build strong relationships all over the company. The more relationships they have, the higher performer they were. (Many studies now show that your “connections” define your success in most companies.)

They also found that these leaders use simpler words, they have more direct connections, and their emails and communications travel further and wider among peers. In other words, they are more focused, influential, and respected.

Praful’s team also discovered that these high performers react more quickly. They are better at time management so they are more responsive, which in turn gives them greater influence and organizational impact. And as you may expect, others respond more quickly to their messages, again indicating impact and reputation.

One of the other implications of this research is that high performers send emails with subject lines which use simpler words.


Listen, simplifying the subject line can be a big deal. Throughout my career I have been frustrated with the irritating process of people forwarding emails with the entire subject line intact, and “FW:” added over and over. This lazy behavior forces to read the entire thread to understand what’s important. While it’s easy for the sender, it wastes enormous amounts of time for the recipients.

In my case, I learned to “delete the subject line” and add a small new one that clearly tells the recipient what I want them to know. It turns out that this type of practice (which I developed during my years at Deloitte) is a good rule for high performers. Simplify your message and make it clear what you’re asking others to do. This, in turn, will elicit a faster response to your communication. (Note the Genpact research did not explicitly identify subject line length as a driver.)

Not only does this clarify your position, directive, or recommendation – it simply saves people time. 

ONA Data Accurately Predicts Flight Risk

The second study Genpact completed was identifying signals that predict a regrettable loss. Praful studied attrition using the ONA tools and found that they could predict loss six months in advance! 

Why? As much of my research indicates, when people leave a company it is usually the result of many months (sometimes years) of frustration. So they show symptoms of unhappiness early.

In the Genpact analysis, people who left the company were significantly less engaged in their communications up to six months in advance. And as Praful’s team did more statistical analysis, they realized they could pick up signals many months prior to the loss.

This points out an important point about retention I’ve learned over the years. Rarely does someone quit for one particular reason. It is often a whole series of “straws that broke the camel’s back” that eventually encourage someone to resign. So if we can identify these issues early we can open up communication and try to repair a problem before it turns into a regrettable loss. ONA facilitates this discovery.

How Can Genpact And Others Use This Data

ONA data like this can be used to coach high performers, identify leadership candidates, and better pre-empt departures by talking with people well before they leave. And this kind of work pays off: if this analysis prevents the loss of only one senior leader at Genpact it could easily pay for itself. Vendors like Trustsphere, Keencorp, and Microsoft now make this easier than ever.

In today’s highly interconnected companies, relationships and communication are often the most important skills for success. ONA, a toolset that has been on the back burner for many years, is now becoming essential and gives HR departments important new insights to truly make the organization perform better.

What does it mean for you? 

  1. Take time to simplify your emails, don’t waste other people’s time.
  2. Communicate rapidly and briefly when you can, it improves the productivity of others.
  3. Clarify your subject line, so people can scan your emails quickly.
  4. Reach out to others, develop relationships as often as possible.
  5. Forward important news to others, it improves your own reputation.

These may seem like small things, but data clearly shows they matter. In today’s world of relentless messaging and email, we can all learn how to make ourselves more effective.