Part of any job’s responsibilities nowadays entails spending a great deal of time sending emails, creating documents and generating other digital data that leaves a unique trail of information exhaust that is now starting to be analyzed by what we call “employee analytics.” Not far from our headquarters in Silicon Valley, there is an innovative small startup called Knack that has made its mission to analyze a small portion of this exhaust. In a nutshell, Knack quantifies the world’s talent and potential based on video game performance. According to Guy Halfteck, Knack founder and CEO, 20 minutes of play generates several megabytes of data about the prospective employee – more than your regular SAT, IQ or Myers-Briggs test. Other similar companies, like Lumosity, have seemingly sprung out of nowhere to capitalize on the human desire to gamify their skills and become “smarter” with only a few minutes of play each day. And employers are interested.
It can’t really be that easy, can it? The science is inconclusive at best, with some research even suggesting it’s essentially useless. But in an analytics era with a “no data left behind” philosophy, businesses aren’t exactly quick to dismiss the potential value of such tools. Maybe we’re focusing on the wrong things, though. If 20 minutes of gaming can provide such purported insight on a candidate, the amount of knowledge behind an employee’s unstructured data in an organization is enormous. Access to analytics that can capture, manage and interpret great volumes of cumulative data then becomes the Holy Grail –- not just for recruitment, but also for employee management, retention and evaluation from cradle to grave (figuratively speaking, of course).
Existing unstructured data, like documents and work emails, are an often-overlooked source of insight to employee skills. Pre-hire assessments are often speculative at best, and frequently rely on suspected correlations that are later proven to be spurious or nonexistent. Just ask Google how well their infamous brain teasers correlated with subsequent job performance (spoiler alert: they didn’t). Of course HR is very interested in the potential of finding the miracle data that can predict an employee’s potential before an offer is given, but in the fever pitch of doing so, they are frequently overlooking the treasure trove of existing data that we DO know has correlation to performance: existing work product. Granted, such data can only be analyzed after an employee has spent some time in their role, but to ignore it would be to ignore a vast wealth of information. In fact, the work product and patterns of existing employees can actually be used to shape more accurate recruitment processes, creating a feedback loop that links internal work metrics with screenings for new workers.
To analyze the benefits that can potentially come from employee analytics, let’s take a generic recruitment process. Say a company knows the main qualities of the people it is searching for (e.g. agile, practical, fast-learning candidates), or the traits and accomplishments of the best workers in a given department (e.g. fast reaction time, previous sales experience). These preferred traits can be identified by looking at current employees, their past work experience, personalities and corporate data exhaust. The company could then attempt to design a game or quiz to get a rough measure these qualities and administered to prospective employees to detect those that best fit the requirements. Although some detractors of these recruitment methods claim that the traditional face-to-face interviews are more personable, they certainly stray recruiters away from making quick judgments based on gender, race, having common interests with the prospect, and other subjective assessments that are not indicative of future performance.
Once an employee has been hired, the information gathered about him could be compared against the information about current employees in order to optimize the placement and training of the individual: in which department would their skills be most useful? What kind of people is this individual more likely to learn from? What kind of training materials should they be provided with? Are they visual learners? Auditory learners? Kinesthetic learners?
Eventually, a company might want to evaluate the employee’s performance, which could be done via the unstructured data that the employee has generated while at the company. As a matter of fact, all companies could do this on their employees based on current data: how many emails have they sent? How many items are in their calendar? How many documents have they created? And most importantly, how many of these are work related? Similarly, we could evaluate productivity based on time spent on different online applications: how long do they browse the web for? How long do they spend on their email communications? But this analysis can be even more complex, as we can analyze the employee’s main relationships within the company: who do they communicate with the most? What is the tone during these exchanges? Are they key people in the organization? This could help during the sales cycle as well, as having an organized repository with all corporate communications would allow employees to trace back who in the company communicated with a client at each stage. Finally, this knowledge will be helpful when the employee retires: what does the replacement have to be familiar with? Which applications? Which relationships are vital in this position?
The possibilities that come with this type of analysis are endless, and in the palm of your hand if you capture and store emails, files, IMs and other types of unstructured data. Some professionals have called this “the Moneyball approach to talent management,” as it could even be possible to predict future behavior with existing data, such as whether the employee is likely to leave the organization by looking at his time at the company, number of promotions and engagement with coworkers. Doing so could have effects in employee happiness, as companies would have the ability to anticipate employee needs.
Of course, some aspects of this type of analytics can make people feel anxious about loss of privacy, and it would be wise to limit access to this unstructured information to respect employee expectations of privacy. It would ultimately be up to the company to exercise discretion and draw the line at a logical, ethical point. But I would argue that employee analytics can bring more good than bad to the corporate world. Today we analyze people’s careers based on test scores, education, interview performance and work experience… with varying success. But do brain teasers measure job performance? Does attending a certain school ensure salesmanship or innovation? Or does GPA matter two years out of school? These factors have been the determinants of people’s success for years, to the point that universities have become the gatekeepers to a prosperous future. Employee analytics could refine the fit between company tasks and employee skills, to try to capture pools of prospective employees that have those skills even if their resumes are less than impressive by traditional measures. This practice could arguably make for a fairer market place, giving employees more options and the power to find jobs that they will more likely enjoy.
Human capital is arguably the most valuable asset in an organization, and consequently one of the main drivers of economic development and performance. Under this premise, it is only logical that people management techniques become more suitable to fuel an economy as sophisticated as ours. The interesting premise, however, is that big data could very likely be at the core of such drastic, yet positive change in the labor force.