It’s no longer enough for companies to manage and improve employee performance; in order to maximize business impact, today’s HR leaders must optimize the environment in which employees work. 

With Zugata Insights, we're using natural language processing (NLP) and machine learning (ML) to surface powerful insights about an organization’s health, a leading indicator of business success. For example, HR leaders now can: measure and manage their company culture, surface and mitigate unconscious bias, and understand the skills and attributes of the organization.

We learned a lot from our customers’ data that I am eager to share with the broader community. I’ll be sharing a 3-part blog series digging into our key findings. In this first post, I’ll share our learnings around company culture: how to define it and how it may not look as you’d expect.

How to uncover your “culture in practice”

I believe it was Peter Drucker who famously said, “Culture eats strategy for breakfast.” A strong, well-defined culture isn’t just a nice to have—it’s the key to business success and a huge competitive advantage. According to Deloitte, 82% of business leaders believe that culture is a potential competitive advantage, but only 28% of survey respondents believe they understand their culture well, and only 19% believe they have the “right culture.” So, while many company leaders understand the importance of company culture, they haven’t had a reliable way to measure it.

Coming up with the cultural values and sharing them broadly is the easy part, but where companies struggle is with figuring out if they are actually “walking the walk.” It’s one thing to say that a core value is to “be inclusive,” but are employees across the organization actually being inclusive?

Regardless of what you say it is, your actual culture is the set of employee behaviors that you reward. Where does this data exist? In your performance reviews.

In our product, we have collected thousands of pieces of feedback written about employees— what they are good at, what they need to work on, and so on. We’ve built algorithms to extract attributes and behaviors used to describe employees and their work. We then filter them based on a performance rating to come up with a list of highly-valued attributes. A detailed understanding of why performance is being rewarded reveals a company’s “culture in practice.”

Mind the culture gap: A tale of two tech companies

After training our algorithms with data from over a hundred companies, we ran this experiment for two of our customers. Our analysis shows that these two companies, located less than a mile from each other, have slightly different attributes and reward different behaviors.

With “Company X,” the most commonly cited attributes of high performers were (simplified for readability): willingness to help others (collaboration), transparency, job knowledge, willingness to go beyond the call of duty, and being results-driven.  With “Company Y,” the most commonly cited attributes of high performers were: growth mindset, leadership, creative problem solving, positive and can-do attitude, work output, and job knowledge.

Based on this data, you can deduce what the culture is at both these companies—both value collaboration and producing excellent results, for example. But what’s more interesting is looking at how this compares to what both companies want their culture to be, i.e., the values they document on their website.

We then created a “Culture Gap Index” (CGI)  to measure the gap between what a company’s culture actually is and what the company thinks it is. In order to find the CGI for each of these two companies, we categorized the most commonly used attributes in the performance data based on each of their cultural values. In an ideal world, where the CGI is zero, these attributes should map to all their published cultural values. For example, if a company’s published values are “transparency,” “inclusion,” “collaboration,” and “results,” performance data should also reflect these four values.

With Company X, we found that all the attributes found in the data were relevant to two of the published cultural values. Two other cultural values they had published on their website did not match up to a single attribute of high performers. In this case, Company X’s CGI is 0.5.

With Company Y, we found that all of the performance attributes in the data fit into one of five (out of the six total published) cultural values. There was one value that no one had described as an attribute of high performers. In this case, Company Y’s CGI is 0.16. This shows that Company Y isn’t too far off from its stated culture while Company X may have more work to do around alignment.

What does this mean for you?

Our analysis of these companies taught us an important lesson: Chances are that your stated cultural values don’t line up precisely with what you reward. If you’re aware of this discrepancy, you have two options: You can prevent your culture from moving too far in a different direction, or you can adapt your stated values to more accurately reflect your organization today.

What might this look like in practice? In order to prevent your culture from moving too far in the wrong direction, you can take a more data-driven approach to performance reviews so that you’re actually assessing and rewarding employees against your published cultural values.

If you decide that you’d like to adapt your values, look at the attributes that describe your top performers (that aren’t currently reflected in your values) and decide which ones you’d like to prioritize instead. This type of shift is not uncommon, especially when companies experience rapid growth, acquisitions, or other inflection points. For example, a younger company might encourage risk-taking as a core value, but as it matures, it may change its degree of risk tolerance.

Keeping your CGI as close to 0 as possible doesn’t just help leaders and managers stay aligned—it creates consistency and fairness for employees, too. When employees know which values they’ll be rewarded for, they’ll be more motivated and enthusiastic about their work. On the other hand, when rewards are inconsistent, it creates confusion and disengagement (and sometimes chaos and unpleasant workplace politics).

Your culture is not static—it’s constantly evolving. That’s why we recommend going through this culture measurement exercise at least once or twice a year. This way you can ensure that your CGI doesn’t widen too much, and keep your company accountable to creating the culture you want.

Learn more about Zugata Insights here and see the orginal article posted on LinkedIn here.