Big data is a relatively simple concept. It’s a lot of accumulated data, and stockpiled within its depths are the answers to all the trends, patterns and associations you need to accurately predict the future. The question then, is how can a business extract and use this knowledge. One way is through predictive analytics, like a doctor examining your medical history to determine your future health and threat of disease. As businesses continue to better leverage their workforce, predictive analytics can present a valuable opportunity for SMEs to effectively allocate limited resources and predict and plan for future personnel turbulence.
The HR competitive advantage
Predictive analytics is relatively new to HR, and within business circles it was a tool primarily used by the marketing department. However, organisations are discovering that it has the potential to help gain a better understanding of talent models and how to predict future performance based on current and past behaviour.
This presents an opportunity for businesses to gain the competitive edge through:
- Prediction of future job vacancies and leadership needs
- Turnover prediction and analysis
- Data-based risk management
- Prediction of future performance and pre-hire flight risk.
By leveraging the mass repositories of data, organisations are able to improve every facet of their HR analytics strategy, saving companies money on turnover and upskilling. Though the potential is huge, there are a number of major issues that must be considered before jumping on the bandwagon.
Current problems with predictive analytics for HR
With any predictive model the key issue is being able to validate your predictions. Tech insight company Visier suggest that predictive capabilities require a ‘minimum of 2-3 years of data for analysis to be valid,’ since it is impossible to validate predictions if it can’t be compared against real outcomes.
Another major issue is breaking down human decision making into simple factors. It is ‘data with feelings’, as Visier describes it. For instance, large companies such as Hewlett-Packard are able to use predictive analytics to determine the correlation between promotions, salary bumps and turnover rate. However, datafication of workplace aspects such as changing emotion, sentiment and interpersonal relationships - and what impact they have on decision making - is more complex. Deriving valuable insights from such data would require analysis across many sources of information. The first issue is the availability of such information, and if that is not the issue, then the insights themselves become the arduous task.
And finally, like all predictive analytics, the analysis is only as good as the quality of the available data. A model based off a competitor's data set or even your own company’s previous year data has the potential to provide very little value. Once these issues are considered and dealt with, companies must then differentiate between business goals and HR goals to ensure the organisation continues to recognise its value.
Business goals vs. HR goals
Big data is only of use if business goals are aligned.
Businesses that apply a piecemeal approach may uncover, through the exploration of HR data, a bevy of interesting, albeit unactionable facts. For example your analytics may tell you that an increase in performance reviews directly correlates with employee job satisfaction. This is an interesting insight, the question is how can you take that information and transform it into a valuable outcome?
Part of the problem businesses can face is that HR aren’t focused on business outcomes, but their own goals. For example, if the business goal is to reduce costs, HR’s objective should pivot from boosting employee morale to reducing turnover and increasing retention. So the challenge is for HR to demonstrate how these insights can provide real value for the business not just themselves. Interestingly the way marketing analytics have been used are paving the way for HR. In the early days of predictive analytics, the marketing teams mined their data to discover trends relevant to their purposes:
- Garnering higher click rates
- How many whitepapers would be needed to double the lead base
- Which channels to use to increase brand awareness.
Ultimately these metrics meant very little to the business as a whole. As the technology became more sophisticated, marketing were able to shift their focus and began identifying the correlation between marketing events with sales and conversions. Amazon is a famous example of this predictive capability. Recommendations on products and services are made to customers based on past behaviour, with reports that claim they account for up to 30% of total sales. Netflix places so much value on predictive analytics that in 2009 they offered one million dollars to a team that could improve its algorithm. The winners improved the movie recommendations by 10%. They incorporated a temporal dimension to the algorithm which accounted for changing opinions based on mood at the time of rating and comparisons with other recently seen movies.
It’s not until marketing demonstrated the capability of predictive modelling (through demonstrable ROI) against the context of business outcomes that resources were readily assigned. HR needs to learn the same lesson. They must translate their findings into actionable information relevant to the business.
Seeing the future
Predictive analytics may never reach the point of taking the ‘human’ out of human resources. However, it can be a powerful tool to leverage the masses of employee data that your company has stored in its servers. It can be a means of optimising your workforce, improve its longevity and effectively allocate available resources. HR can ensure their investment is worthwhile by breaking down departmental silos and always keeping the goals of the business as the focal point of their design.