In the world of HR, the ability of HR practitioners to grapple with data and glean insights has been suspect by business leaders for a long time. There are a variety of reasons for this from the lack of sufficiently high data quality, to insufficient HR practitioner skills to analyze data, to the lack of funding to implement analytical technologies and capabilities. Whatever the reasons, HR has not proven to the business that it can deliver insights that have high impact for business leaders and the things they think about on a daily basis.
In HR’s pursuit to get better at catering to the business, it has pursued a variety of programs including implementing technologies that provide “insights out of the box” and governance programs that promise to clean data so that HR has more credibility. But these programs don’t increase HR’s ability to deliver an insight – they only allow HR to deliver something they should have had 10 years ago. THis presents one of the significant problems HR has with data – in the lack of having adequate capability, HR thinks that bigger is better and the shiny new toy will solve all problems.
Just as when 5 years ago “social” was the shiny new toy (that HR was afraid to tackle), today HR looks to big data to assist with capability shortcomings. HR reasons that if their organizational data is paired with external data, instant benchmarks can provide immediate and deep insights for leaders. They think that bringing in data from non-HR systems within the organizational firewall will help them identify human resource attributes like productivity that were previously off limits. These assumptions can be true, but only if HR staff has a capability already to grapple with data problems and create insights of of those problems. Shiny tools that help visualize make it easier to create insights, but the beginning and end to the problem is and always has been about data oriented people.
HR should not go big – in fact, HR should probably go smaller to the smallest segments of data that they already have but don’t use effectively. Rather than looking towards big data to solve problems and introduce new complexities that they are unready to deal with, HR should be taking the current HR data already in possession and looking at unused meta-data to tap into new insights. There are things going on within our organization that HR has not even thought about, but impact the future of the workplace and how employees get things done.
HR Meta-Data Use Case 1:
Within almost every good HR system is the ability to track data about each transaction. HR already reports on mundane transactions like job changes and trends about who goes where and how long the average person takes to get promoted. But embedded within each of these transactions is the meta-data that tells us when managers are performing these transactions (date and time stamp), and sometimes what type of device they are using. Purely hypothetical, an organization could find out that managers typically perform HR transactions outside of normal work hours, presumably because they are busy with their “regular” jobs between 8 to 5. Or that managers who have had tablets deployed to them and have a better user interface approve transactions faster than their counterparts. In today’s world, HR is consumed by administering the employee record rather than understanding how work gets done more productively. The meta-data can tell us how a transaction got done, not just what the transaction was.
HR has plenty of data, especially now that Talent Management has been in full swing for almost a decade. Meta-data by definition is “data about the data.” For years, HR has not cared about this type of data and has almost completely ignored it. Indeed, the purpose of meta-data seemed to be system oriented rather than business useful. As HR data sets have grown however, use cases have increased beyond just understanding how managers are doing their work, to increasing insights about the employees themselves, and how HR can improve their own programs.
HR Meta-Data Use Case 2:
Most employers provide some form of total compensation statements with the objective of increasing employee engagement and satisfaction. It’s hard to correlate the simple presence of a total compensation statement with satisfaction without meta-data. Combining meta-data with employee engagement survey data, HR can make real correlations between employees that log into total compensation statements to view them, and that populations’ engagement scores. Are people who look at their compensation all the time more or less engaged? Is checking total compensation a predecessor activity for voluntary terminations? Taking the meta-data to another level, the volume of logins per employee might )or might not) yield even greater impact to engagement. This activity could be extended to payroll and benefit portals as well. Given real data regarding correlation, HR might elect to develop marketing campaigns to increase traffic to benefits portals among older employees, total compensation among mid career employees, and learning portals among younger employees.
In the last several years, people have become more sensitive to the intrusion of privacy that governments can exercise. From the NSA tracking who we are talking to, to our cell phone providers tracking where we are at any time, meta-data has gotten a bad rap for being a tool of “big brother.” Indeed, HR has actually been using meta-data since the dawn of “personnel administration. Simple headcount reports are not about the HR data itself, but a count of the data. Turnover reports are similarly not about the data, but a count of a specific type of transactions. So looking deeper into meta-data than HR has in previous generations should not be a surprise.
HR has a wealth of data already at its disposal. In its attempt to find meaning within that data, expanding the data set and making it even harder for HR to craft new insightfulness would seem to be a mistake. HR already has more data than it can consume, and finding meaningful ways to use and interpret that data is a better bet than trying to create entirely new data sets while existing data sets stagnate even more.