Your guide to HR analytics
Jessica Twentyman, June 05, 2014
A step-by-step guide to data analytics
Data analytics is not for the faint-hearted, so we have compiled a step-by-step guide to getting it right, from establishing objectives to presenting your findings. By Jessica Twentyman
In the finance department, employees are fed a rich diet of reports, metrics and key performance indicators (KPIs). Without the clear insight these provide into company performance, accounting staff could barely function, much less give a good impression in meetings with senior management.
But for their colleagues in HR, the situation could hardly be more different. In a 2012 Economist Intelligence Unit survey of 418 global executives, an overwhelming 85% said their HR team didn’t excel at providing insightful metrics.
In HR’s defence, the function faces considerable challenges when it comes to assembling and presenting meaningful workforce analytics. Relevant data is often spread across a number of different HR systems, each of which might define the same information differently. And some of that data may be downright inaccurate or woefully out of date.
Above all, HR professionals are often short on the skills required to bring all that data together, run sophisticated reports and queries and then interpret their results, according to Tim Payne, a talent management leader and former head of people at consultancy firm KPMG.
In a report published last year by KPMG, People are the Real Numbers, Payne calls on organisations to provide the resources, time and training required to create a “new breed” of HR analytics specialist. “Historically, statistics has simply not been a requirement of most HR roles,” he says.
Take it step by step
That it’s a challenge for the function isn’t to be doubted but, in a world where everything is based on data, HR needs to take the bull by the horns. But what should HR directors be doing to provide a level of analysis that business leaders value so highly? HR magazine talked to the experts to suggest, on the following pages, four simple starting steps for those new to the world of analytics.
1. Understand the business questions you need to answer
This is vital, and it’s where a lot of data analysis efforts fail, says Ravin Jesuthasan, global practice leader for talent management at advisory company Towers Watson and co-author of Transformative HR: How Great Companies Use Evidence-Based Change for Sustainable Advantage. “What we often see are huge efforts to number-crunch all kinds of data and then look for correlations. That’s too much effort for too little insight – and the kinds of insight it throws up can be pretty spurious at times,” he explains.
Start with the business problem you’re trying to solve, he advises. The issue at hand, for example, may be as simple as measuring workforce diversity, or staff turnover, or training spend or absence levels. However, in order to get a real grasp on where the company currently stands, you need to be able to measure how those metrics change over time: quarter by quarter, for example, or year-on-year.
More complex questions, meanwhile, will seek to discover how HR effort impacts wider business results. For example, how does our investment in learning and development improve our profit margins?
2. Identify the data you will need in order to properly answer those questions
At some organisations, relevant data may reside in a single enterprise resource planning (ERP) system – but this is pretty rare. In most cases, a good starting point is to hunt through a number of HR-related systems for data on recruitment, performance and succession; this, combined with data from exit interviews, reasons for rejection, employee engagement results and employee surveys, can lead to fresh insights.
But to answer more complex questions, the hunt should be extended to other business systems. These could be financial systems, where valuable data on business performance by geography, market and sector resides. Or the answers could be found in sales systems (which sales executives outperformed their annual target?) or those related to customer service (which call-centre staff resolve the most customer queries?).
3. Build and populate your analytical environment
At many companies, including telecoms firm TalkTalk (see box, left), the current analytic environment – a repository where data to be analysed is assembled – is an Excel-based balanced scorecard. But it could be a database or even a data warehouse, depending on the volume of data to be analysed and the complexity of the planned analysis.
Either way, this is where HR will require considerable input and help from colleagues in IT. Preparing data from different systems for analysis involves a process known as ETL: extraction, transformation and loading. There are a range of tools available that can be deployed to perform ETL tasks automatically, according to previously defined rules. In other words, they extract data from relevant source systems, transform it so it’s in a clean and consistent state for analysis, and then load it into the analytical environment.
Running on top of this environment, meanwhile, will likely be a range of business intelligence tools for processing queries and reports and populating scorecards and executive dashboards. In recent years, much has been done by business intelligence tool vendors such as SAP, Oracle, Microsoft and IBM to make these easier to use for non-technical executives, but IT’s advice will continue to be invaluable in determining which are best suited for the task at hand and that fit well with the company’s existing analytical investments.
4. Incorporate findings into business conversations
An investment in workforce analytics seldom comes cheap. In order to get the best return on investment, HR needs to perform analyses regularly and, more importantly, present its insights to the business in ways that the business can actually use, explains Towers Watson’s Jesuthasan.
“Plenty of HR functions have data on staff turnover, for example, but where they fall down is in presenting it to the business in context,” he says.
“Is turnover too high, for example, in areas where we value experience, longevity and in-depth company knowledge, such as customer service, or is it too low in areas where we need creativity, innovation and ‘fresh blood’, such as R&D?”
And the importance of making solid connections between HR metrics and other business measures, such as productivity, should not be neglected, according to KPMG’s report.
“While it’s helpful to track absences by location or versus prior years, if HR could also show that improvements in absenteeism positively correlate with manufacturing costefficiency, then line leaders would be more likely to see the value of HR,” it states.
In other words, by creating a clear ‘line of sight’ between HR interventions and bottom-line profitability, HR analytics can finally provide the proof, with little room for argument, of the definite link between effective people management and superior business results.
Which are the HR metrics that count?
The most insightful analyses typically draw on data sourced from a host of different IT systems and business areas, but HR teams that are just starting to get to grips with analytics often find that a handful of key metrics on their own department’s performance is a good place to start.
These might include:
- Resignation rate: How many employees resigned in a given period, as a percentage of the overall workforce?
- Time to recruit: How long did a vacancy stand empty before a new hire accepted an offer of employment?
- First-year staff turnover rate: How many new hires left in year one, as a percentage of the number of new hires that year?
- Revenue per full-time equivalent (FTE ): What is our annual revenue generated per FTE?
- Profit per FTE : What is our average operating profit generated per FTE?
- Performance appraisal participation rate: What percentage of employees completed their performance appraisal and rating?
By any measure, the HR team at telecoms company TalkTalk seems to be getting it right.
Each month, it produces a balanced scorecard, which presents a wealth of key metrics on internal mobility, time to hire, cost to hire, attrition rates, gross and net profit margins per full-time equivalent (FTE) employees, amount of overtime paid, the ratio of permanent to temporary workers, and more.
This scorecard, based on Microsoft Excel, also allows users to drill down into the information it contains by both group and by business unit and is presented to the company’s executive committee each month, as well as distributed to its board of directors in their board packs.
“Our people and the metrics relating to our people, are taken incredibly seriously by the whole business,” explains Jo Taylor, head of talent management at TalkTalk.
But assembling the information needed to deliver this level of analysis takes a huge effort, a lot of it involving timeconsuming manual processes, she adds, because at present, TalkTalk doesn’t have a single human capital management (HCM) system.
Instead, the data needed comes from seven different systems and must be submitted by relevant HR leaders each month to a single member of the HR operations team responsible for collating it all for data entry – a process that probably takes him a week, Taylor reckons.
Just identifying and submitting the data relating to talent management objectives, for example, takes Taylor’s team around half a day each month, she adds.
The task is set to become simpler, she explains, when the company consolidates its numerous systems into just two, later this year: an HCM system from Workday, and a payroll system from ADP.
“Talk to me in a year’s time and we’ll be able to do this analysis at the click of a button,” says Taylor.
For now, the extra effort is worth making. Taylor adds: “Providing this level of insight is what makes the HR function at TalkTalk a true strategic business partner. It really enhances our image as a department that adds real value.”
The challenges of big data
As HR grapples to come to terms with big data, experts at a recent ADP panel debate discussed big data challenges, what is useful, and important skills for the HRD of the future. ARVIND HICKMAN reports.
Is it little wonder why many HR executives fear big data? Just about everything is measured these days, from personal details, absence rates, productivity levels, the time you take lunch to even, in some cases, the amount of exercise you do.
To make life more difficult, all of this data doesn’t sit neatly in one, easy-to-use piece of kit. It’s often stored in different databases or filling cabinets by different gatekeepers, who are sometimes siloed and have different agendas.
But as technology advances, so too can complexity, and in most companies finding useful data requires battling poor and disconnected data capturing methods.
Even at Nestle, one of the world’s largest food manufacturers, getting the right data is not easy. “Just being able to get the data to ask the questions you want to ask is incredibly difficult,” said Matt Stripe, group HRD of Nestle UK and Ireland.
Stripe explained that at Nestle, HR data is captured by an SAP system, but there are also 360-degree appraisals and leadership development programmes, which sit somewhere else, separate talent management system and financial reporting systems.
The challenge, Stripe said, is to “cleanse” large volumes of data enough to be confident in the analysis to make decisions. To overcome this problem, Nestle has an analytics team of five people – young analysts mostly with backgrounds in financial services.
“We protect them because HR people want to ‘HR them’, they have no idea about appraisals. We just let them create and take all of the data and ask them the really hard questions,” he said.
Confidence in data is important, but so, too, is comparability. This is particularly the case for the c-suite, according to Maria Black, president of ADP TotalSource and former managing director of ADP UK.
“Taking masses of workforce analytics and looking at how it compares, and where to resource and invest, allows a business to stay sharp, lean and execute quickly in a highly competitive globalised market,” Black said.
And the best types of data to use?
Stripe explained that at Nestle there are three phases of data – revealing (how do you understand what is going on better); enhancing (how do we move from lagging measures to leading measures that enhance what we do); and predictive (modelling that is built on ‘what ifs’ and ‘maybes’). “A lot of the value comes from the first two phases, understanding what’s going on in the organisation and making better decisions,” Stripe said. “Big data enables you to put different data sources together at differently levels of the organisation to allow that level of the organisation to make a decision. You get more value from the reveal and enhance.”
Lancaster University Management School senior lecturer Anthony Hesketh argued predictive data was fraught with danger: “The only thing we can predict about prediction is that it is probably wrong.”
Hesketh said the HRD of the future would need to be well networked with other c-level executives; strategicallyminded; able to deliver; and “able to understand and utilise the analytical internal insights”.
Max Blumberg, founder of Blumberg Partnership, added that it was about being “able to see snake oil where it exists”. “You don’t have to be a statistician but you need to understand what a correlation means, the significance of it… [and] not have the wool pulled over your eyes.”
Blumberg strongly believes that data should be a central function and not solely the domain of HR. “Analytics is something that belongs to the organisation, and HR feeds into that. If you start thinking siloed… you move away from an integrated whole,” he said.