The HR profession has been bitten by a bug; using big data, analytics and numbers to predict organisational and individual behaviour.
The HR profession has been bitten by a bug; using big data, analytics and numbers to predict organisational and individual behaviour. The theory of big data is to have no theory; you just gather huge amounts of information, observe patterns, and estimate probabilities. As Viktor Mayer-Schönberger and Kenneth Cukier wrote in their book “Big Data”, “this movement values correlation over causation because correlations not only offer insights, but also because the insights they offer are relatively clear. These insights often get obscured when we bring causality into the picture.” While big data is useful in flagging what to focus your attention on, doing something what you are paying attention to takes you back to the worlds of narratives, judgement, and causality.
Over the last few years many surveys suggest that more than 50% of companies reported higher spending on HR analytics. During the same period, two important HR areas—performance management and employee engagement—saw the beginning of the end of quantitative methods, the bell curve and the yearly employee engagement survey, employed in them. While much has been said on how mathematising HR will aid better decision making and a lot has been spent in attempting to do so, there is a gap between executives’ perception of importance of key HR areas and HR’s capability to deliver on them. Paradoxes, such as these, can be solved only if HR professionals exercise caution when mathematising HR, by being selective in areas of application of analytics and clearly differentiating between co-relation and causation.
To test our hypothesis, we looked at changes in transactional, traditional and transformative aspects of HR practices in managing people. We grouped HR process that have the highest leverage in impacting an organisation as transformative—leadership development, organisation design, M&A, workforce planning. Traditional HR processes include typical employee life cycle management processes such as hiring, training, deployment, performance management, learning and development, employee engagement; transactional processes would include all HR processes that are high frequency and high volume such as payroll and query management. We observed how these processes have moved from correlation to causation.
Hiring has moved from rejection for lack of quantitative abilities to selection based on competency-based structured interviews. Performance management has moved from grouping people using bell curves where they compete against each other to being focused on continuous feedback and coaching to enable collaborating teams in high performance workgroups. Training is no longer based on quantitative measures of knowledge and skills but on developing competencies where motives and attitudes matter more for differentiating performance. Compensation philosophies have moved from focusing heavily on transactional pay elements, percentiles and compa-ratios to carefully managing relational elements such as need, faith, trust and hope. The grand annual employee engagement survey followed by co-relation analysis has given way to quick pulse surveys that lend themselves to immediate and pointed corrective action. Retention programs are wiser now to have moved from throwing money at problems identified through co-relation analyses to addressing root causes through causative analyses. Even in something as machine oriented as HR information systems that address transactional elements like employee queries, the focus has shifted from building systems of record and systems of transactions to using empathy and appreciative inquiry to build systems of engagement that focus on user happiness.
Transformative areas of HR operate more in the realm of vision and probability than in those of math and co-relations. Leadership development is more about crucial conversations and diversity in experiences than quantitative succession planning. Organisational design has moved from just addressing elements of structure, process and rewards to creating self-managing terms that work on a series of transformative projects. Quantitative workforce planning is partially irrelevant in a volatile, uncertain, complex and ambiguous world; what matters is intuitive development of long term organisational capabilities. IPO and M&A valuations, today, give as much weightage to the entrepreneur’s motives, management aspiration and company vision as much as they do to EBIDTA multiples.
New York Times columnists David Brooks suggests three problems with Big Data and a correlation view of the world. First many things can be correlated based on how you structure the data and what you compare; to discern meaningful correlations from meaningless ones you have to rely on some causal hypothesis. Second, unlike physical objects, humans are discontinuous; we have multiple selves, we are ambiguous, we are ambivalent, we get bored, and we self-deceive. Finally the world is error prone and dynamic; people often misinterpret reality and those misinterpretations sometimes create self-reinforcing feedback loops.
We do not argue against the use of Big Data or analytics in HR; we are just flagging caution with “the mathematising of the subjective”. The current wave of mathematising has led to crores spent without an appreciable improvement in perception of what has been delivered by HR. The world of quality changed when reliance on solely quantitative QC methods gave way to TQM methods, where quality was dealt with as a state of mind and the physics envy of economics gave rise to a mathiness and financial market quants that jointly contributed to the 2008 global financial crisis. Great HR practices and high performing organisations have always been hard to create; mathifying them gives the illusion of them being easy. Careful with valuing analysis over synthesis.
The writers Manish Sbharwal and Santosh Thangavellu are with Teamlease Services. Views are personal