People analytics is currently having a moment in the HR space. Fueled by the growth of data science, more and more companies and organizations are trying to drive greater insight into their people and talent, and more and more vendors are offering technology which claims to do so.
But what exactly is people analytics? Looking across a number of research projects conducted by us and others, spanning numerous organizations, practitioners, academics, industries and sectors, we conclude that people analytics, like data science, is currently a catch-all term. We’ve encountered the term “people analytics” being used for anything ranging from basic data reporting on talent processes to genuinely advanced, predictive insights on talent based on multi-year longitudinal studies.
Our belief based on the research is that, despite the sudden boost in people analytics and the number of organizations setting up dedicated people analytics groups, only a few organizations have a genuinely sophisticated operation, which has driven true insight and value to how they operate.
The number of CHROs focused on harnessing analytics and data to drive better insight and decision-making around talent is ever-growing. Jumping on the people analytics bandwagon is no longer an innovative move – it is table stakes. To stay innovative and ahead of the pack, CHROs should take time to understand what it takes to build a function which offers more than basic reporting and can generate genuine insight to support their talent strategy.
The current people analytics landscape
Although there has been scientific and analytic inquiry into people at work for over a century (Salas, Kozlowski & Chen, 2017), people analytics only really hit the mainstream in 2010 when a HBR article by Tom Davenport, Jeanne Harris and Jeremy Shapiro, “Competing on Talent Analytics,” was published.
Surveys of senior executives consistently show an increased interest in and expectation around the use of data to inform people decisions. It is now much more common for a large organization to have a people analytics team than to be without one, and as a core part of HR, it is starting to get its metaphorical footing. Notably, Bersin by Deloitte’s 2017 industry study of people analytics practices showed that only 14% of organizations were still mired in a fragmented and unsupported approach to people analytics. While that same study revealed that a mere 2% of organizations were at the fully integrated level of maturity, 69% of those responding had achieved sufficient infrastructure and expertise to deliver solid HR reporting and consistent use of HR data. Even with that progress, a 2017 study by Levenson and Pillans (restricted access) showed that over half (54%) of organizations were very limited or worse in their ability to use talent data to predict and improve business outcomes. Rather than declare people analytics a red herring based on those disappointing results, a recent i4cp report revealed that 70% of organizations (with at least 1,000 employees) are dedicating resources to people analytics. The train is not stopping any time soon.
Analytics still limited to basics
While it is exciting to see progress, in many organizations the term “people analytics” equates to basic HR reporting, surveys around fundamentals like engagement, and, where outcomes are included, they are often ROI estimates reminiscent of the utility studies of the 70s.
Much of this is a data literacy issue. People analytics can only live up to its full potential in an organization which embraces and understands data. We are now starting to see organizations increasingly focusing on data fundamentals, including accuracy, governance, privacy, and access. Additionally, some organizations, such as Chevron and JP Morgan Chase, are investing in broad data literacy across their HR population. Perhaps unsurprisingly, the organizations who did this groundwork early are also the ones with the most sophisticated people analytics functions and the closest link between people analytics and business outcomes.
Beyond the basics, such as reporting, engagement surveys, and, more recently, retention prediction, people analytics diverges broadly. People analytics teams vary massively in size. While some boast more than 50 individuals, most tend to be fairly small and the work of gathering, cleaning, organizing and analyzing data remains cumbersome. Levenson and Pillans (2017) highlight the wide range of projects in people analytics. The chart lists some examples. Click it for a larger version.
Moving to people analytics
One thing is clear from the research. The nature and purpose of people analytics – particularly how it differs from HR reporting and metrics – is not well understood among most organizations.
HR reporting and metrics is concerned with measuring the most obvious and critical people related phenomena in an organization, usually in isolation. Headcount, hires, terminations, utilization are all elements of HR reporting and metrics. The focus is current or historic, and is mostly internally facing. The data usually support operational decision-making, and are essential for that purpose. Mathematical methods are relatively straightforward and the most critical requirements are effective systems for tracking people data. As such, HR reporting and metrics generates facts and operational insights. Examples include: “Which recruiting sources have the greatest yield for us?” and “Why has our sales unit grown so much this year?”
People analytics is a much broader and encompassing field. Most importantly, people analytics drives strategic insight and influence by addressing future trends or decisions, and therefore can be a major differentiator in ensuring HR plays a role in shaping strategy and organizational development. The complexity of the questions addressed in people analytics are an order of magnitude greater, require a research-based approach, and demand skills across multiple disciplines to address. Developed people analytics teams may include data scientists, data engineers, psychologists, economists, even epidemiologists and anthropologists, as well as translators who can understand the business problem and oversee a multi-disciplinary approach to solving it. Examples of people analytics work include: “What will our most critical talent sources be in the next 5-10 years?” and “Can we reduce attrition by predicting which individuals are at risk of leaving?”
Building a people analytics function
In corporate enterprises, people analytics is often born out of HR reporting and metrics. Breaking out of the operational data needs of the business to dedicate substantial capacity to complex, longitudinal people questions is not easy – our research showed this to be a very common frustration among leaders in this space. Our belief is that there are several steps to achieving a truly world class people analytics operation.
Step 1: Automate HR reporting and metrics
In most HR reporting functions today, numerous individuals are fully utilized in repetitive, straightforward tasks often entered into spreadsheets like Microsoft Excel. The same process is followed weekly, monthly or quarterly with the same data and result. The output is rarely acted on, the process itself becomes dull and uninspiring for the individual responsible and no valuable skills are learned from performing the same task over and over again. With the rapid development of technology over the past 5 years, many of these processes can now be automated to execute with little or no human effort. Greater levels of automation in data analytics and delivery, can create precious capacity, and resources can be released and deployed against more advanced and strategic work. HR analytics groups hiring their first data scientist or data engineer should immediately deploy them to increase automation wherever possible.
Article Continues Below
Contingent Workforce Strategy Survey With ERE and Aptitude Research
If your company currently leverages contingent workers, please share your views in our brief survey.
Step 2: Build advanced approach around strategic priorities
As teams move into advanced analytics, groups should focus initially around a small number of use cases with a high degree of expected impact. As highlighted by the examples in the chart, these use cases can be driven by data currently in the digital exhaust, by external research, by new internal research, or a combination of all of these. By focusing this way, the group can keep the workload manageable and the chances of building initial success stories and internal branding are much higher.
Step 3: Adapt people measures for agility
Agility is a critical factor in the success of a people analytics function. Ongoing study of priority use cases will usually highlight phenomena that were previously unknown and will require active measurement and tracking moving forward. Tools and methods must be put in place to actively measure these new phenomena for the purposes of interventions or longitudinal analysis. Further, it is important to periodically test whether tools and methods identified by research have exhausted their usefulness and should be evolved.
Step 4: Build longitudinal insights into ‘bottom line’ people outcomes
A long-term goal for any aspirational people analytics function is to be able to understand the drivers of outcomes within their organizations. Within recruiting an outcome could be a successful hire. Within people development, an outcome may be promotion, successful performance in a substantially different role, or unwanted attrition. Building predictive capabilities that allow an organization to predict these outcomes at both the aggregate and individual level is a long-term goal for most teams, as it can add a new layer of intelligence to workforce planning and to individual development interventions. Almost certainly this level of sophistication cannot be achieved without sufficient progress on Steps 1-3. No organization can currently claim to be operating entirely at this level as of yet, although some companies have made progress in developing predictive capabilities in the areas of recruiting and retention.
Where the analytics team fits in
From interviews and research, it is abundantly clear that many of the frustrations in getting a people analytics group operating effectively were organizationally driven. How such a group embeds itself within the organization, how it prioritizes its work and how it interfaces with the rest of the organization arose as three critical factors in enabling success.
The most successful people analytics groups have consistent visibility to the CHRO or Chief People Officer. In most successful cases, the people analytics leader reported directly to the CHRO. Frustrations around impact were expressed in situations where the group was embedded in illogical parts of the organization, of where it had been “passed around” due to a lack of buy-in as to its role and value. In one case, the group was split between IT and HR in its reporting structure. In another it had moved from HR, to finance, to IT in the space of a three-year period.
Many people analytics groups also struggled to extricate themselves from the ongoing day-to-day demand for data from the business, and were not finding dedicated time to work on more advanced or strategic projects. Successful groups had either resolved this through high levels of resourcing, or in some cases through separating their people analytics group away from the team that does reporting, with separate managers and reporting structures, allowing the people analytics group to focus entirely on advanced analytics and research.
How the groups interfaced with the business and determined analytical priorities varied greatly, but the most successful groups were able to clearly articulate the logic for their particular configuration. Some groups were strongly CHRO-centric, reflecting a top down talent strategy in their organization. Others directly faced their key internal clients or business units, often with a key account manager for each. This reflected a more decentralized culture, but also a potential role for the people analytics group in feeding back to the CHRO the key concerns and priorities of the different business units.
People analytics has seen some development and progress in the years since its conception, but movement has been slow. Only through a more crisp understanding of the nature and value of people analytics and a more thoughtful approach to building the capability can CHROs benefit from its impact in their organizations. Those that have taken the time to do it properly will end up with a lot more to show for it.