When do organisations do well? When the people who work in them do well. You would be hard pressed to find anyone with a vehement objection to this answer.
Yet, the small matter of managing people and talent in an organisation is often left to ‘the traditional methods of personal relationships, decision making based on experience, and risk avoidance’. This is according to Professors Massey, Haas and Bidwell from the Wharton School at the University of Pennsylvania where they run a course on People Analytics.
People or Talent Analytics is a fast emerging branch of Data Analytics, itself an area of statistical investigation applied to real world problems. In recent years, words like Big Data, Machine Learning and Neural Networks have intimated and confused in equal measure, obscuring somewhat, the vast promise of data analytics to transform organisational outcomes. For some, the term People Analytics conjures up images of white coated lab assistants attaching probes and sensors to rows of hapless individuals strapped to metal gurneys. For others, it means the abdication of the Human Resources profession to clever mathematicians and statisticians who communicate in barely intelligible one’s and zero’s, when they communicate at all. Both premises are ridden with anxiety and it’s important that we understand what happens when data analytics and people issues meet.
Hunting for buried treasure As a city that was once a frontier mining town, Johannesburg has a fascinating history. It drew adventurers from around the world who arrived on the Highveld with little else but the hope that there was something precious lurking in the ground beneath their feet. Organisations today sit on something much more precious than gold. They have data, and lots of it.
For too many organisations however, data is like that perennially scruffy relative who lives in the basement, accumulating odds and ends in a dimly lit warren that only a few IT colleagues, claim to understand. Organisations know intuitively that their data can be immensely insightful but it takes too much effort to clean it up, coax it upstairs, and make sense of what it has to say. Enter the data scientist, a new breed of prospector, who driven by an innate curiosity, uses statistical and other techniques to find useful patterns in data. In doing this, data scientists help organisations perform better by helping them make much better decisions. Nowadays, data analytics is routinely applied to a vast range of real world problems including fraud detection, medical diagnoses, insurance underwriting, shopping recommendations, system failure predictions, reward programs design, public policy and the allocation of marketing budgets, to name a few.
Relative to other areas of business and public administration, data analytics is only just starting to touch upon people issues in organisations and this is long overdue. Whether it is recruiting, performance evaluation, retention, promotion, job design, or compensation[1], data analytics can help HR practitioners make better decisions by uncovering the insights in their data.
The right one As winter draws to an end and the wedding circuit in South Africa gathers pace, it is interesting to note how similar the process of finding the right partner is to hiring the right candidate. Getting it wrong can cost a great deal of money in direct expenses and in lost opportunities, not to mention the potential for reputational damage. Businesses spend 50-60% of their total revenues on payroll and yet this large expense is rarely well analyzed[2]. If the job of hiring is based on predicting performance, then it’s surprising how few organisations look to predictive analytics to help them get it right. While there is a place for screening CVs and traditional interviews, the changing demands of a Millennial workforce suggest a dollop of science would not be amiss.
Just as some online lenders draw inferences from how up to date the software on your computer is to make decisions about your propensity to repay debt, hiring decisions can be enhanced by casting the data net wide enough to include unconventional attributes that correlate well to success in a particular role. These may be as counter intuitive as the type of car a candidate drives or whether she took a gap year before university or even how many siblings she has.
Please don’t go We’ve often heard many large organisations bemoan the fact that certain demographics – black women for example – leave their organisations in disproportionately large numbers. The wringing of hands in executive boardrooms is usually followed by focus groups set up to understand the issues better and the appointment of a high powered executive to deal with the matter. While these initiatives are all well and good, is anyone looking at the data?
Hewlett Packard applied the principles of data science to employee retention and assigned each of its more than 330,000 employees a flight risk score[3]. The HP data analysts looked for patterns across a wide range of attributes to determine the combination of factors that indicated an employee might leave the organisation. If a South African organisation were to conduct a similar exercise they might discover that:
Black women, between the ages of 29 and 35, with two or less children, who received a better than average rating at their last performance review, and whose managers have been in the organisation for longer than three years, … have a high propensity to leave the organisation.
While the above scenario is fictitious, it illustrates the power of data analytics to cast light on seemingly intractable problems and to do so at large scale. Knowing who in their workforce fits the description of flight risk, managers can put contingency plans in place and intervene before the employee might even have thought of leaving.
[1] People Analytics. coursera.org
[2] The Geeks Arrive In HR: People Analytics Is Here. Forbes Magazine, February 2015. Josh Bersin
[3] Predictive Analytics – The power to predict who will click, buy, lie or die. By Eric Siegel