People analytics – under a variety of names – has been a hot topic for several years now, to the point that most HR leaders we talk to have included it somewhere in their functional strategy. There are impactful case studies – predictive models preempting (some) attrition with recommended job, salary or development actions; retail workforce optimization models that increase top-line performance and decrease safety incidents; talent modeling to build higher quality leadership pipelines. Yet despite big investments in cloud HR and new talent, many organizations have shifted their sights from stat-heavy predictive modeling to less rigorous (though still impactful) dashboards and interactive data visualizations.
So, what’s stopping HR from realizing the predictive analytics dream? From our work with companies across industries and sizes, we’ve identified four barriers that leaders can act on immediately.
Hype versus problem-solving
The most significant barrier we see is the excitement for predictive analytics without a corresponding big problem to solve. After all, people analytics is a set of tools, not an outcome in and of itself. To have impact, these tools need to be aimed at compelling business problems or opportunities, or else they risk being a hammer without a nail. Leaders who frame their analytics goals as business outcomes rather than abstract HR capabilities are more likely to find a greater purpose and buy-in within their teams and among their C-suite peers.
Pushing through this barrier is about achieving high-visibility wins to generate demand and engaging cross-functionally with peers in finance and operational functions to identify, prioritize and solve enterprise problems.
Be careful what you look for…
The second barrier is that analytics often point to patterns and root causes that are intractable, whether technically, financially or more often culturally, so improvement efforts fizzle in the face of the monumental changes that would be required. Our analysis may help us better understand where glass ceilings are formed in our talent processes or job levels, however breaking them requires a complex set of interventions with iterations over time. Or we may find that we can predict attrition only to find we don’t have the budget to mitigate it or the fortitude to exit the managers who disproportionally cause it. Plus, the wide range of interventions for any gnarly talent issue requires an equally wide range of HR diagnostics and remedies, including talent assessment, talent brokering, coaching, mediation, workforce planning and so many others. For our analytics efforts to succeed, these HR capabilities must be strong enough to overcome whatever root causes the analytics have revealed. It comes down to the question Jim Malone (Sean Connery) asked Elliot Ness (Kevin Costner) in The Untouchables: “What are you prepared to do?”
Big stakes analytics are more successful where there is leadership alignment up front, agreement to follow the facts and a commitment to pragmatic prioritization of solutions.
Data management is relegated to nerd-dom
Many of us have had the unfortunate experience of revealing an analytical masterpiece in a leadership meeting (with a flourish of our slide rule), only to have the conversation descend into minutia when the underlying data is questioned. For people analytics to inform business strategy, the curation of the underlying data needs to be treated as a strategic imperative. With the proliferation of point solutions in the HR tech landscape, our data is traversing a complex network of platforms, apps and vendors in our own house, raising the stakes for getting the master data right. Beyond this, HR data is also relied upon outside of the function for analytics, workflow, security, etc. The bottom line is, without high quality data, we don’t have a shared view of reality that we and our stakeholders can rely on, yet it’s too often treated as a tactical box to check. We are the curators of this asset and should treat it as foundational – to the smooth operation of our business and the insights that guide our most important decisions.
Organizations that can serve up analytics without fear of data integrity breakdowns do so by maintaining a structured rigor amongst designated data owners and ensuring that key data elements are visible and well-trafficked by employees.
Diminishing returns for data science versus easier alternatives
Analytical projects can be difficult and labor intensive to run as full end-to-end experiments. Newly identified problems might require non-existent data to be captured or existing data to be repurposed. Data science can be iterative and thereby challenging for tight deadlines. Correlation does not always equate to causality. It’s easy to see why many HR teams find more immediate benefit from less statistically intense projects and instead focus on visually describing population differences, process metrics and other straightforward measures of people-related outcomes. There’s a case to be made for getting the basics right and selectively investing in data science where it’s the right tool for a high-value job.
HR teams who succeed with analytics balance the more accessible tools of dashboards and other visualizations with high-value, statistical projects, relying on established methodologies to “rinse and repeat” with accelerating efficiency, while providing their data science team latitude to get curious.
Whether your priority in the coming year is better understanding your current workforce or finding hidden cause and effect relationships to improve future outcomes, step back and consider which of these roadblocks (or others) are most likely to get in your way. Take stock of your talent, technology, data and most importantly, the “why” for your analytics journey, then allow for experimentation along the way. And as always, reach out to your colleagues at herronpalmer for help as you go.
