Much like a dashboard, which indicates to drivers how their vehicles are performing, workforce analytics can function as a proactive “dashboard” for CFOs and others in the business to drive talent management – and thus generate more business value.
At least that’s the contention of Amel Arhab, senior manager at Deloitte US Consulting LLP. Her own office uses it. At Deloitte US, workforce analytics allows not only the measurement of employee performance. It also predicts employee behavior, such as someone getting ready to resign, thus allowing managers to take proactive measures.
Many companies are looking at workforce analytics to help them understand why employees decide to stay or leave an organization
A lot of Deloitte’s learning in America should be applicable in Asia as well, where the problems of job switching, recruitment and retention hurdles, and a shallow finance talent pool are acute particularly in China, Hong Kong, Malaysia and Singapore.
“The approach and techniques should be highly transferable, with some needed customization on implementation and messaging,” reckons Arhab. “For example, the process of collecting data and building a model follows the same approach.”
“However, we may find that travels impact US offices more than Asia offices, given that practices are different [in the two regions].”
Talent and techniques
Many companies are looking at workforce analytics to help them understand why employees decide to stay or leave an organization, says Arhab. The most effective models, she believes, balance capability-building and point solutions, meaning that companies develop both the talent that conduct analytics and the use of tools and techniques.
Scenario analysis (i.e. ‘worst-case scenario’) and simulation were once considered overblown, but these techniques are standard today in predicting and responding to future events. They are also useful in workforce analytics.
Arhab also suggests integrating analytics with the ERP system and Human Resources management. In this way, raw data are turned into powerful insights that anticipate problems before they happen. There’s a lot of internal data to work with, she points out. The company’s ERP and other systems would have generated huge amounts of information over the years.
The sheer volume of data could be a problem, says Arhab, and moving from basic operational reporting to advanced predictive analytics takes time. “Start small and make small wins,” she counsels. “Be practical about the solution.”
Where Are You In Analytics?
Bersin by Deloitte Consulting, a provider of advisory services in human resources, defines analytics as “the systematic discovery of meaningful patterns in data to support decision-making.” As applied to talent management, it “helps managers and executives make decisions about their employees or workforce.”
Bersin has developed what it calls the Bersin Talent Analytics Maturity Model (see chart below) to serve as a general guideline for organizations “in setting initial goals for using analytics before acquiring or developing more advanced capabilities.”
The Bersin Talent Analytics Maturity Model
Most companies already use analytics in one form or another. For example, your company probably analyzes performance results before and after employee training. This is a good start, says Bersin, but analytics should eventually go beyond just evaluation because “evaluation results are not actionable information.”
Deloitte US has already reached Level 4, predictive analysis. But it took years for it to reach that ultimate stage. As Arhab says, the best approach is to start small and make small wins.
Companies should take a step-by-step approach to workforce analytics, rather than go all out and acquire the most sophisticated tools to jump into predictive analytics. Bersin details the activities and goals of the first two levels below:
Embarking on the Journey
“The focus at Levels 1 and 2 of analytics is on developing familiarity with data, and generating reports, charts, and digestible data accessible for managers and executives,” Wang-Audia explains.
In Level 1, the chief activities are discovering data, asking relevant questions and setting objectives. It is helpful to build what Wang-Audia calls a “data dictionary,” a resource that records where data items are specifically sourced and how they are identified and defined.
The objective of this data dictionary is to “build a track record of validation for reliable analyses down the line.” The data that is mapped should not be limited to just HR or finance, but should include all the other different sectors of the organization.
Next comes asking the right questions. Data by itself is amorphous and sprawling. “To avoid getting lost in the sea of data, ask the right questions to set a direction or strategy,” Wang-Audia advises.
She helpfully provides some examples of business-relevant analytics questions to ask:
- What is the relationship between engagement, rewards, tenure, and skills with revenue?
- Why is turnover high in some areas (for example, in finance)?
- Why is there fraud in some offices?
- What will our talent gaps be next year based on this year’s retirement rates?
Based on the questions, companies should then collect the relevant data, referencing the data dictionary to know what is available and where. “Collect data by areas (e.g., recruiting, performance, training),” Wang-Audia recommends. In focusing collection this way, specific practice areas can be tied to business metrics.
“Building strong metrics on those areas first can set the foundation to cross-examine for patterns, once strong analytics skills and knowledge of data have been established,” she adds. At this point, measurement tools should be chosen, based on their capability to aggregate data and generate dashboards and customizable reports “with dynamic presentation options.”
Wang-Audia stresses the importance of getting management buy-in for the project, to help build an understanding of how different business units work, make sure the most relevant information is submitted and win IT’s cooperation to advise on methods of data collection and most appropriate tools (Excel? business intelligence software), for example.
On to Predictive Analytics
Level 2 (Advanced Reporting) in this progression focuses on analyses and identification of patterns and trends, while Level 3 (Strategic Analytics) tackles statistical analyses, development of “People Models” and delivery of actionable solutions.
Level 4 (Predictive Analytics) is the ultimate objective. Having reached this level, Deloitte US now has the capability to understand the dynamics of employee attrition in its American offices based on the analyses of time-series-based HR data sets.
"It is crucial for companies to keep a pulse on engagement levels and take swift action when these levels start to experience even a slight decline”
Arhab and her colleagues examined internal and third-party external data sources (e.g. psychographic and lifestyle-behaviors) to identify more than 1,000 predictive variables such as lack of travel, vacation leave and group work.
Age, gender, travel expenses, time sheets and project financials as well as market-related information such as the unemployment rate and GDP were also found to have predictive value, and thus included in the attrition models that were developed.
One of the variables that were found to be a leading indicator for attrition is the level of engagement, as captured by sentiment analyses or survey data. “Therefore, it is crucial for companies to keep a pulse on engagement levels and take swift action when these levels start to experience even a slight decline,” says Arhab.
She encourages CFOs and other executives to “move away from a tribal wisdom-type of thinking and decision-making to a data-driven environment where things are quantifiable, traceable, monitor-able, and where accountability is established across the organization.”
What About ROI?
Given the time required to build up analytical capabilities within the finance team, systems and processes and acquisition of tools, is the adoption of workforce analytics worth it in terms of return on investment?
Working out the ROI is difficult, Arhab concedes. “But we have seen proxies work quite well.” These include basic attrition calculations associated with 1.5–5x salary losses and more complex productivity-loss calculations.
“It is possible to work out scenarios (i.e. from most conservative to most aggressive) to gain a perspective on the amount of realizable gains as well as create stakeholder accountability,” Arhab asserts.
Moreover, Software-as a-Service solutions and cloud computing technology are driving down the cost of data management and analytics. Research by Bersin also shows that companies that have reached the Level 4 stage produced stock returns that are higher by 30% than peers still in the lower levels or have not embarked on workforce analytics.
For CFOs rightly obsessed with the bottom line, these findings could tip the balance towards predictive analytics, despite the time, care, capability building and investment in tools that are required.
About the Author
Jefferson Mendoza is Online Editor at CFO Innovation.
Photo credit: Shutterstock