In honor of Pi Day (March 14), I decided this is a good time to channel Marie Antoinette and take stock of the current state of the HR analytics revolution. Just like with any revolution, the current state of HR analytics appears to have its share of zealous converts, soldiers and civilians, active resisters, and PR that obscures what’s really happening on the front lines. And perhaps the biggest question is whether it’s a genuine revolution that’s truly advancing the “national interests” of the business or a distraction that’s drawing attention away from where the real battles should be waged. As the risk of the analytics hoards storming the castle in protest, here are some personal reflections.
Revolution or evolution? Rome wasn’t built in a day, and no revolution ever happened over night. Currently we are seeing some amazing and rapid advances in the tools of HR analytics available to frontline business leaders and HR professionals, so it feels like the revolution is moving quickly. But the availability of those tools and the ability to leverage them to solve the most pressing business challenges are two different things. Getting to the stage of deep insights is going to take a lot more evolution in how both the business and HR address analytics, an evolution in directions that are often off the beaten paths of where the soldiers are marching today.
One the one hand, the heavy artillery being sold and installed by the leading software vendors really is a leap forward in providing ready access to data that otherwise could take enormous hours to collect and analyze. Providing data in ready-to-use formats that can be analyzed at the click of a mouse and displayed in graphical ways that make the data come alive to non-statisticians really are advances worth celebrating. These new software systems are saving countless hours that HR business partners and analysts otherwise would have to spend painstakingly pulling data just to produce regular HR reports on turnover, headcount, career progression, performance management, and much, much more.
At the same time, the largest companies are building entire HR analytics teams, while smaller companies have greater and greater access to an army of consultants ready to help analyze their data. The current arms race in HR analytics is just that: a race to get staffed up, either internally or externally, with the arms, legs and brains needed to put all the newly accessible data to good use. In the early days when the internet first launched, HTML programmers and web site builders where the hottest commodity that companies couldn’t hire fast enough; these days it’s data scientists and big data analysts.
Yet just like the Cold War that stockpiled weapons that were never deployed for their intended use, the current HR analytics arms race raises questions about what the benefits are of all this extra spending. No software system has ever provided the kinds of insights most salespeople claim it will. These software suites can contribute a foundation of reliable data that helps inform better data-based decision making. But they don’t address the huge gaps that still exist: the software suites alone are incapable of addressing the fundamental drivers of behavior and motivation that are the barriers to strategy execution and organizational effectiveness.
Is your HR analyst (or consultant) friend or foe? One of the big challenges facing the business and HR these days is understanding who to turn to, and who to trust, when it comes to creating insights about human capital development, talent management and data-based decision making. There is an ever-growing army of analysts and consultants who are available to help crunch your data and provide assistance. But are they doing the right things?
I was intrigued recently by a blog post that talked about “real” research scientists working in HR analytics versus the more “ordinary” folk who are just good with data without having lots of formal social science or natural science training. The implication of the post was that good analytics can be done by regular people. While it would be great to have a “real” research scientist at your beck and call, the conclusion was that you can get by with ordinary folk – even though you should want the services of the research scientist if you can find (and afford) them.
Speaking as a Ph.D. economist who has worked closely with some of the leading Ph.D. IO psychologists in the field, I have to take issue with that conclusion, though for different reasons than you likely expect to hear from someone like me. The issue is that classically trained social scientists are really good at a fairly narrow set of issues and aren’t typically challenged in their training or day-to-day work to tackle the big messy problems of strategy execution and organizational effectiveness.
If you want a scientifically valid screening procedure, or competency assessment, or 360 evaluation, etc. then there is no substitute for Ph.D. expertise. But as soon as you turn your attention to issues like “why aren’t we getting the productivity we need out of our work groups” or “why don’t we have alignment across our functions and work streams” or “how can we squeeze greater efficiency out of our go to market processes without sacrificing quality”, for those types of questions classical social science training does not confer any competitive advantage. In fact, it can do the opposite because the social scientist isn’t trained to deal with those kinds of big messy org issues.
Which is not to say that non-social scientists are necessarily any better than the social scientists at tackling those big messy issues. In my years of working with companies I have been struck by how few people take a true systems perspective when diagnosing the sources of performance problems. If you don’t look at the complete system – the entire battlefield – you can easily lose sight of critical bottlenecks that can stop your troops dead in their tracks.
This is where the promise of HR analytics gets confronted with the cold hard reality of the breadth of perspectives and knowledge needed to properly diagnose barriers to strategy execution and organizational effectiveness. A successful militia consists of soldiers who play many different roles, drawing on a wide array of competencies and experiences, as well as leaders with the vision to see the bigger picture and ultimate end game to win the war. The team you need to conduct proper systems diagnoses of the barriers to performance in your organization needs to be similarly diverse in the analyses it conducts.
Don’t fire until you see the whites of their eyes. Like soldiers on the front lines waiting for the battle to start, HR professionals today are armed with so many different ways to access and analyze data. The barriers to doing data analysis are being lowered so dramatically, soon just about anyone with access to a computer will be able to manipulate HR data to potentially derive some interesting insights. Yet providing the tools to access and manipulate existing data in your IT systems is very different from providing the training and perspective needed to know what questions to ask of the data, and when you need to look elsewhere for answers.
Back in the time of the Revolutionary War, the soldiers were ordered to “not fire until you see the whites of their eyes” when engaging in battle. This was done to help preserve scarce ammunition and to account for the inaccuracy of weapons that were not totally reliable. The instructions to wait were part of the training that made them better soldiers.
The situation today in HR analytics is similar. The inaccurate tools are the analyses that rely only on archival data on headcount, compensation, career progression, turnover, spans of control, and more. The timing issue speaks to the tendency to report out preliminary analyses that look like they provide insights without ensuring that you’ve dug deep enough to get to the root of the problem. The training that’s needed – and where people have to train themselves – is to take the time to make sure you’re taking a comprehensive enough look at the issue.
The “whites of their eyes” in this case are the models that fully describe the motivation and behaviors you’re trying to explain. The data needed to provide the right answers almost always are a combination of what’s sitting in your systems, what’s in your annual employee survey (if you have one), and a big dose of data from other sources: stakeholder interviews on the problems with organizational alignment, on the barriers to effective team performance, and on the aspects of your culture that enable versus impede effective strategy execution; targeted surveys that measure very specific issues, not the broad brush items in a typical annual employee survey; and so on. In fact, a good, thorough analysis sometimes requires only interviews without crunching lots of numbers in your data warehouse or fielding a new survey. This is an entirely different type of weapon and ammunition than is typically deployed in the practice of HR analytics today.
Don’t forget the drawbridge so you can easily get in and out of the castle. A big risk today is that we are building great walls around HR analytics processes in the effort to ensure we are doing the right thing. Enormous time and energy is devoted to cleaning up the data so we can get to the “truth” about headcount and turnover. Hours and hours of meetings are devoted to determining who should have access to which elements of the HR IT system so that confidentiality is maintained and employees are protected. Standardized reports are created so managers can do self-service – and come up with requests for more and more additional reports. All of these are important steps in the analytics evolution, but they are not the building blocks for a real revolution. The risk is that we won’t spend enough time building bridges to the rest of the organization to ensure that the analytics are aligned with and directly focused the underlying business issues.
Those bridges start with making sure the analysts and business partners who are using the data work hand in hand with the business and the rest of HR. Together they must define the issues to be addressed, get stakeholders on board to support the analysis, and prepare the organization to act on the results. The drawbridge in this case is the information that gets shared across functional lines, and the decisions that are made jointly so that HR analytics is not conducted in a vacuum.
The deepest, most useful insights always start with an integrated process that brings together subject matter experts from the business and HR, along with the analysts, to jointly identify the issues to be addressed. From there the team needs to define the analyses to be conducted, including collecting additional qualitative information (through interviews) and data as needed to do a thorough systems analysis. The team has to be closely involved as the analytics are conducted, sharing in the initial discovery and helping to make sense of the patterns that emerge, and jointly deciding what tweaks or additions to the analysis are needed. That way, by the time the analysis is completed, members from both the analytics castle and the walls outside will be fully aligned and supportive of the results, and in prime positions to drive change.
The HR analytics revolution in many ways is in pretty early stages. The key to maintaining momentum is to recognize just how far we have yet to go, being realistic with ourselves, our clients, and our stakeholders about what can and can’t be said through the application of current state HR analytics. That’s not as exciting as championing the dawning of a new golden age of understanding human capital and the drivers of productivity and profitability. Yet realism is what’s needed to keep the troops moving forward in the right direction so you can win not just the battles immediately in front of you but also the long game of the entire war.
