This column is part of Globe Careers' Leadership Lab series, where executives and experts share their views and advice about leadership and management. Follow us at @Globe_Careers. Find all Leadership Lab stories at tgam.ca/leadershiplab.
People are the most valuable asset of any organization. Being able to anticipate when (and why) star players will be at risk of leaving is paramount to success.
With the ever-increasing variety and volume of today's workforce data, the application of predictive models to talent retention is still nascent, but it is becoming the new norm.
Business leaders understand that the ultimate goal of predictive models is to anticipate the outcome of an event. For instance, helping to predict who will buy? Who will click? Who will vote? Who will lie? And anticipating those employees who may be looking for a new job leverages the same methodology and algorithm.
Traditionally, predictive models are built leveraging customer transactional history and profile data from systems that encompass the customer's buying behaviour. This information is then analyzed to determine those individuals who are most likely to engage, as well as how to best engage with them. The same principles apply to predictive modeling for talent retention.
Two major sources of data that can be used for talent retention include internal data from a Human Resources Information System (HRIS) and publicly available data from social media channels and niche sites.
For most companies without data scientists, the only insight into employee attrition is achieved during exit interviews. But for many reasons these interviews do not always provide useful information. That is why forward-looking companies are turning to Big Data to determine why.
Common employee attrition drivers from traditional HRIS data analytics include: lack of engagement, which is measured via engagement sentiment surveys; lack of leadership and clear vision from the executive team; lack of opportunities to grow internally; tenure, age and gender; years of experience; commute time to work; work-life balance; and performance reviews, assessments and rewards.
Attrition reasons differ from one company to another – however, mining this data will help to anticipate who is at risk of leaving, thereby giving an organization the opportunity to take appropriate proactive actions.
One concern is potential litigation around privacy issues and legal compliance, but the precedent for applying predictive modeling to ensure a more satisfied workforce exists. Companies such as HP scored their entire workforce in order to anticipate attrition for their near 34K employee-base, and Capital One and Deloitte followed suit. Google, one of the pioneers in this space, was able to build an algorithm that helped the company anticipate their employees' behaviour, even before those same employees had given thought to leaving.
HRIS data has done a great job in helping companies better understand employee behaviour, but it is not enough. Thanks to social media, and niche and community sites harnessing digital footprints is the new way to analyze changes in an employee's online digital profile and activity.
Publicly available data is made up of three sets of information that include:
Labour market data: Statistics Canada data, such as the unemployment rate by education level, region, industry, GDP, company payroll, university graduation data, completion data, job vacancies, and labour turn over, leave/quit/termination trends, wage and payroll data.
Company data: Company financial performance, organizational structure, reorganization history, brand, and online reputation.
Individual Data: Profile updates and changes (for instance on LinkedIn and other social media networks), as well as work activity on social community sites like Stackoverflow and GitHub.
This publicly available data is a gold mine for workforce analytics to address some of the primary workforce challenges, including identifying employees who are at risk of leaving, when and why they might be at risk to leave and how they can be retained.
Data science applied to publicly available data is helping organizations to compensate for the limitation of internal data, but more importantly, it provides actionable insights on what will happen with today's workforce. It also helps to avoid some privacy issues that mining HR data could pose.
Predictive analytics is a mandatory component of the new generation of strategic workforce and talent management. In today's competitive labour market, experience, data scientists, and a combination of publicly available data and HRIS data will help companies proactively retain and protect their most valuable asset: people.
Jean-Paul Isson is global vice president for predictive analytics and business intelligence at Monster Worldwide, Inc. His second book is, "People Analytics in the Era of Big Data: Changing the Way You Attract, Acquire, Develop, and Retain Talent."