Factory workers at the Western Electric Company, based just outside 1920s Chicago, were found to be more motivated and productive because of social factors than monetary incentives and good working conditions. These factors included choosing their own co-workers, working as a group, being treated as ‘special’, and having a sympathetic supervisor.
Fast forward almost 100 years after the Hawthorne Studies, as the experiments above were known, the art of managing people is making another significant leap towards becoming a science. Enter People Analytics (or HR Analytics).
The 3 ‘V’s
Is this just hype? Maybe. But there are three drivers that have been identified that suggest this revolution is real.
1) The volume of data being collected has been increasing at a frenetic pace. In the time that it takes you to read this sentence, more data has crossed the internet than was stored in the entire worldwide web as recently as 1996. The cost of data collection and storage has been falling to support this trend, as is projected to continue falling.
This makes it easier than ever before to warehouse huge repositories of people data, which holds the potential of yielding never-before-considered insight on managing people.
2) The velocity of data creation and transfer has allowed companies to operate with real-time or near real-time information, improving business agility. The demand for such data immediacy is getting stronger – almost 50% of companies want information within an hour of a business event. With team and task changes moving at this pace, does an annual or bi-annual employee engagement survey still make sense? Or for that matter, having the reports from those surveys delivered 6-8 weeks after the responses have been given, in service of a ‘cascade’ of communications that has become legacy in many companies.
3) The variety of data sources has also allowed for creative analytics to further mitigate the omitted variable bias in using traditional HR data. GPS data from mobile phones, near-field communication chips in employee ID tags, facial recognition, social media sentiment tracking, information from wearables, and other data sources have made up what Kevin Ashton has coined the “Internet of Things” at the workplace. Deloitte, an accounting firm, has experimented with a wearable gadget that tracks the movement and conversations of every employee with consent, to draw conclusions about the behaviour of their best performers, granular understanding of productivity, and speech patterns of the best managers.
The potential for mashing up this data with traditional metrics like key performance indicators, employee demographic data, resume entries, job descriptions, and others, is tantalising to the average Joe data scientist.
Here’s the rub
PwC reports that 86% of business leaders indicate that creating or maturing their people analytics function is a strategic priority for the next 1 to 3 years. However, there are two main problems with implementing this. Any company intending to use analytics to make better people decisions needs to have action plans that address both.
Problem #1: Data Collection and Silos
While ever-increasing amounts of data are being collected at ever-increasing speeds, not all of this data is easily tractable for analysis. This part of an excellent infographic explains the problem succinctly:
The larger a company is and the longer it has been in operations, the greater the degree to which data silos and incompatible information management systems complicate attempts at making the data useful.
This is such a massive problem that the market for enterprise resource planning software (which promises to maintain common standards for data collection and analysis across an organisation) is expected to reach $41.69 billion by 2020, registering an annual growth rate of 7.2% from 2014-2020.
Problem #2: HR’s analytical skill level
Traditionally HR professionals excel at qualitative analysis: understanding conceptual models of human behaviour, building frameworks for competencies and process areas to support the business, and intuitively assessing reactions to change initiatives.
When it comes to quantitative analytical chops, on the other hand, we as a profession may be deluded. Oracle reports that while 63% of HR leaders think they draw insight from data, only about a fifth of their non-HR counterparts agree.
If it looks like a duck, and quacks like a duck…
Interestingly, for all the pontificating on HR analytics and People Analytics, there’s very little consensus as to what this field actually is, and what it can do.
The camp adopting the ‘narrow’ definition would only qualify any hard-core data science initiative with multiple new and traditional sources of HR data, sophisticated data modelling and predictive analytics. This group is familiar with algorithms in R, Python programming, and machine learning, and when these tools are applied with the right business acumen as guidance, and grounding in HR for background, the results can be truly insightful.
While this approach maintains fidelity to the broader data science movement, the reality is that it may be too advanced for many HR departments and companies. This makes the understanding, and subsequently implementation, of the recommendations weak.
The other group, adopting the ‘loose’ definition of HR analytics, trumpets bread-and-butter analytical methods as the latest and greatest in HR. Regressions, cluster and factor analyses, correlation monitoring, and data segmentation are used to generate workforce insights based on a relatively creative degree of data sourcing (correlating employee and customer engagement data for example).
It may be tempting to dismiss that camp as overhyped marketers, but it seems that even that level of application of analytical tools can yield low-hanging fruit. In a recent survey, 28% of companies admitted to not tracking the cost per hire, and almost 10% said they had no idea how much each hire was costing them at all.
Both camps are important
Actually, it is clear that moving towards either camp would take the HR profession forward. Using Bersin’s talent analytics maturity model, 86% of companies have their HR stuck in a reporting function, without yet contributing significantly to executive decision-making.
Why should we care?
Thomas Davenport, Jeanne Harris, and Jeremy Shapiro identified six main areas in which people analytics can be helpful:
- Human-capital facts are HR data points regarding individual performance and enterprise-level data such as headcount, contingent labour use, and recruiting. The choice of the right metrics can provide key indicators to an organisation’s overall health.
- Analytical HR collects or segments HR data to gain insight into specific groups of employees. This allows targeted intervention activity to be applied to the right groups, at the right time.
- Human-capital Investment Analysis help prioritise actions that have the greatest impact on the business, for example investing in the retention of key front-line service staff who provide customer service and build customer relationships.
- Workforce Forecasting allows companies to model turnover, succession planning, and business opportunity data to identify potential shortfalls or excesses of key capabilities in advance, to take pre-emptive action to staff up or cut back.
- Talent Value Modelling crunches data to identify the retention levers of key talent, to help managers design personalised performance incentives, adjust pay, or decide when to promote someone.
- The Talent Supply Chain helps companies to make real-time decisions about talent-related matters such as optimising a store’s next-day work schedule on the basis of predicted receipts and individual sales performance patterns. This requires high quality data, rigorous analysis, and cross-platform data integration.
Although these observations were made in 2010, they remain very prescient. In the five years since then, while there have been pockets of companies that have executed on these well (Google and ConAgra are frequently cited), most are still grappling with transcending the operational and transactional demands of the HR function.
In short, developments in the Volume, Velocity and Variety of data processing, the potential for HR to become valued partners in the C-suite is greater than ever. Some companies have already done this, and have generated 30% higher stock returns than the S&P 500 over the last 3 years. The rewards to those who catch up look to be significant, as long as the two bugbears in implementation around data collection and silos and the capabilities within HR are addressed.
As William Gibson observed, “the future has already arrived, it is just not evenly distributed yet.”
[This post was originally published here.]
Author: Chee Tung
CheeTung is the CEO of EngageRocket, an HR tech startup that analyses employee feedback in real-time to advise you on how to build a better culture, one team at a time.