Leaders in People Analytics: Capital One’s ongoing quest to advance HR research
We met with Guru Sethupathy, head of People Analytics at Capital One. He’s been with Capital One for over a year and has been working in human capital, analytics, and economics for over a decade. He likes to stay on top of the latest social science news, research, and books by following a wide variety of social scientists on Twitter and keeping frequently in touch with his former academic colleagues in economics and other social sciences.
RE:WORK EDITORS: What is an example of a business challenge you think is worth solving by using people analytics?
SETHUPATHY: There are a hundred things we could be doing. The key is to stay close to our customers – our human resource business partners and business leaders – to make sure we are doing the most impactful things. Right now we are working on a suite of products that can provide an “org health” view across a wide array of people and organizational metrics, and are planning to include future predictive and prescriptive insights using machine learning.
One example of how we leveraged people analytics at Capital One was through an internal study we conducted on our people managers. Like Google, we wanted to unpack the specific behaviors common amongst our most highly-rated managers. We surveyed our associates on thirteen behaviors, culled from research, and landed on two key manager behaviors: highly-rated people managers support their associates (especially when they need it most) and demonstrate consistency between what they say and do. We also unsurprisingly found that improving our lowest-rated people managers would likely have a high return on investment.
We’ve incorporated these findings into our management trainings, and have recently created a people leadership fundamentals development program that we’ll be rolling out to key target populations, including new people managers as well as low- and medium-rated people managers.
What's an insight from social science that you've found to be especially influential in the workplace?
SETHUPATHY: There are a few areas from the different social science fields that we are reviewing and incorporating into our work. From the field of industrial-organizational psychology, we are looking into the role that hidden networks can play. Rob Cross’s pioneering work on organizational networks (ONA) has been influential to my team’s work.
Most companies are organized by “reporting to” structures — “who reports to who”. Yet, research from Rob Cross and others show that these structures aren’t fully telling of how a company is organized or how work gets done. Cross-functional, collaborative, agile networks are usually “hidden”, meaning they are hard to observe and document, but can be more relevant to understanding how work really gets done as well as what productivity, bottlenecks, and work flows really look like.
For instance, positive or negative news can create a shock to different pockets of an organization. Looking at organizational networks can allow us to understand how these shocks reverberate throughout the organization so that leaders can manage and mitigate appropriately. When it comes to Diversity & Inclusion efforts, organizational networks can provide us greater insights into how different groups of employees are interacting with one another, how included they’re feeling, and how D&I commitments are being promoted within different parts of the business. We are also looking at organizational networks to determine how to improve productivity and enablement.
What's a particularly useful finding you read about recently? How has that informed your work?
SETHUPATHY: Companies often manage their talent according to a normal distribution, but Herman Aguinas and others put forth evidence that there is an underlying power law distribution that is closer to a log-normal distribution, depending on the job type and industry. A major implication of this performance distribution shows that top performers, or “stars”, can produce 5x, 10x, or even higher multiples of value compared to median performers, thereby incentivizing many companies to focus efforts on chasing star talent.
But while finding great talent is important, we also see that’s it’s not the end-all solution. In the book, Chasing Stars, Boris Groyberg argues that individuals who are stars at one company may not be at a different company, if they are bereft of the right kind of support.
Performance is likely to be a function of capabilities, context, and luck. Companies often over-index on capabilities and underestimate the role context can have in helping others succeed. Company culture, role, and manager fit, as well as other factors can significantly contribute to the performance of all employees, even those who score high on capabilities.
So while the importance of recruiting and selecting top external talent should and will always be a focus at Capital One, we’re investing in more targeted ways that we can develop people post-hire, so that all employees can have the ingredients they need to learn, grow, and perform.
What is the hardest part about straddling the research/practice divide? What strategies have you and your team used to overcome it?
SETHUPATHY: In many large organizations, one challenge research teams face is the limitations of their influence. Although such teams can yield critical insights, they’re often not the final decision makers. So how do we, at Capital One People Analytics, hope to ensure that our insights lead to impact?
We have three approaches. First, we’ve encoded insights into descriptive and predictive products and placed them in the hands of our decision-makers to inform their talent-related decisions. Going forward, we’ll keep tracking our products’ broader use — so far early signs look promising!
Second, our team works to influence decisions by building relationships with our leaders not only in HR (though it’s important for me to stay connected to our CHRO) but also leaders across the business. This allows us to stay in the loop on business priorities so we can find opportunities to proactively offer data-driven solutions. We’re able to use what we know ahead of time about the challenges each organization faces as we assemble our data, results, and recommendations. This way, we’re able to engage earlier and can offer more holistic solutions.
Lastly, we’ve started to build and deploy products directly for our employees. Currently, these products and dashboards have customized views for an associate (e.g., their skills profile) and serve both to inform our associates as well as to nudge their behaviors (think calorie displays at a restaurant). This path enables us to help our associates directly, and to impact skills and culture change, one associate at a time.
In this series, we interview leaders in People Analytics from across different industries to learn more about how organizations have applied research and data to their work and people processes.