Analytics, modelling and forecasting: using people data for insight
8 minute read
HR software has the potential to capture more HR data than ever. Is your HR team making the most of this opportunity to gain real insight into your people?
When the pandemic turned business as usual upside down in early 2020, HR teams were faced with more questions than they had answers for. Did their organisation have the capacity to support the shift to working from home? How would furlough impact pay and staffing levels? How could essential departments deal with potentially high levels of sickness absence?
Research carried out by the HR Analytics ThinkTank, a partnership between universities and private companies that compiles intelligence on the people analytics field, shows that data played a crucial role in decision making during the Covid-19 crisis, helping HR teams address immediate issues and answer unprecedented questions. Nigel Dias, the think tank’s founder and managing director of consultancy 3n Strategy, explains: “The pandemic saw businesses making decisions they’ve never made before, and there were questions no one had ever answered before. People were either guessing, or they had data which enabled them to make better choices.” People analytics – the discipline of using HR and related data to make organisational decisions and even predict outcomes – will continue to play a key role in decision-making as organisations recover from the pandemic, the think tank concluded. “The next stage will see people analytics teams answer new questions, including questions about how employees are coping with their new working arrangements, including the challenges of working from home, risking their lives as essential employees, or leaving their employment – and the overall mental health of employees as well as dealing with the uncertainty and stress of the situation.”
For many HR departments, however, the pandemic represented a shock to the system in terms of the data they had available and their capability to deal with it. Consulting company McKinsey identified four key uses for people analytics during the crisis: workforce sentiment analytics, looking at how people felt; workforce protection data, such as social distancing or how many employees are in the office; visibility of who is working on what remotely; and workforce availability. To make robust decisions in these areas, HR teams often needed to access reliable and real-time data feeds from multiple systems outside their central HR system, such as recruitment software and rostering systems, and surveying or employee engagement tools.
The pandemic also brought into sharper focus the debate around HR professionals’ skills in people analytics. “In the past 10 years the companies that do this really well, such as Facebook and Google, haven’t really changed,” says Professor Andy Charlwood, chair in human resources management at the University of Leeds. “There’s a feeling among HR that they’re good at managing people so they question the value of data. Either that or they believe their HR system is there to make processes more efficient rather than create insights.” There are also mundane practical issues such as staff setting up fields in different formats so data cannot be compared, or failing to input data at all. “Sorting out the data is a big job and not a nice job, and so it seems to become more arduous than it actually is,” he adds.
A survey by the CIPD carried out in 2018 found that only just over half (54%) of organisations had access to people data and analytics, and only four in ten felt their HR team would be able to tackle business issues using data analysis. Three-quarters who had access to people data were using it to tackle workforce performance and productivity challenges, and in organisations where respondents felt their analytics culture was strong, 65% reported better business performance compared with competitors. The CIPD advised a cross-functional approach to improving HR’s skills and confidence in this area. “There are clear differences in the perspectives of HR and finance professionals, and other professionals using people data,” said its report. “Non-HR functions need encouragement to increase their use of people data in their decision-making and HR has a role to play in generating trusted, relevant people data to serve wider business needs.”
Sources of people data
Organisations can draw on a range of data sources to model how their workforce looks against different parameters now and in the future. These include but are by no means limited to:
- Central HR system data (basic information about individuals, such as name, job title, department)
- Demographic data (from the central HR system and through disclosure of, for example, sexual orientation, socioeconomic background and ethnicity)
- Recruitment systems (for example, how many hires, time to hire, and diversity of new hires’ backgrounds)
- Onboarding system data (eg drop outs, time to productivity, targets for training)
- Learning management systems (eg course completion, skills needs, popularity of courses, compliance)
- Employee surveys (eg in which departments are people most satisfied? Who needs support?)
- Absence records (if not held in the central HR system)
- Performance management tools (such as feedback scores or comments)
- External sources (such as job boards, pay benchmarking, and social media)
Professor Charlwood recommends organisations that are starting their analytics journey begin with the business problem, rather than expecting the data to present solutions. “Take sickness absence, for example. It’s easy to see who is off sick, but not necessarily know why. Perhaps the data in the HR system shows that it is women in their late-30s who could be taking time off to look after children, but is that actionable insight?” he says. “To get to the insight you need to collect data in the right way, so for absence that might be through return-to-work interviews. Once you know the answer, that can inform how you change how employees are managed or their work is designed.”
Spending time at the start of any data analysis project identifying the right questions, then defining the types of insights the organisation would like to see from the analysis is crucial. Dias adds: “Anything HR achieves is a result of a decision it has made, and at any time HR professionals could be making hundreds of people-related decisions. If your gender pay gap is X, then what decisions were made to get there? If you want to reduce it, what is the factor that will make the change?” One of the barriers can be that seasoned HR professionals feel that they have been making these decisions for years and “don’t want to feel that a piece of analytics is more powerful than their expertise.” The key, says Dias, is in understanding that the data can help teams make better choices rather than replacing their years of experience.
One area where the role of analytics has come to the fore during the pandemic is employee experience. Suddenly, organisations craved more insight than ever before into how employees were feeling, their stress levels, and their physical wellbeing. Many opted for more regular pulse surveys on topics such as mental health, producing a wealth of data points for analysis. Rob Robson, director of people science at the People Experience Hub (PX Hub), says the same principle applies as with any strand of data. “Why do you want to collect data from employees in the first place? What problem are you trying to solve? What outcomes do you want? Going straight into investing in a platform could close off options. Your research question should be your mantra; you need to determine what you challenge is, before you think about how you’re going to solve it.”
Collecting data from employees about their experience, or asking them to disclose additional personal information, can feed into useful insights on improving the business culture and diversity and inclusion – but it also requires careful handling. Data protection regulations (the GDPR) require organisations to be transparent about how and why they are collecting personal data, and to minimise the number of people who need to handle it. Informing employees how their data will be used, presenting the insights to them, and showing that they have been acted upon can help build trust. Aggregating and anonymising any datasets from particular groups ensures individuals cannot be identified, says Dias, but often the organisation’s culture will drive how receptive it is about using people data to drive decisions. He adds: “Engagement data can be hard to marry up with other strands because it’s anonymised, but the biggest factor in building a data-led culture is leadership. If you don’t have sponsorship at the highest level, every other measure will be undermined.”
When it comes to modelling people data, how systems are integrated beneath the surface will influence the efficiency and reliability of insights for HR. There are a number of tools on the market, such as Visier and Capterra, that will draw data out of different systems and automate the processing. Microsoft’s Power BI tool, for instance, will generate data visualisations and insights from CSV files. “If you’ve integrated systems so that data is up to date in real time then you can make evidence-based decisions based on the data you have. You’re running a board report that’s up to date, rather than working from an old spreadsheet,” explains Phil King, director of sales at Ciphr.
The key to unearthing deeper insights can often lie in overlaying data from different sources. At a simple level, this could be marrying up data on attrition in a certain demographic group with the same group’s engagement scores. Building it up step by step can build confidence, adds King. “Start with a question such as ‘How do candidates rate the recruitment process?’ and once you’ve got that, you could look within that at age, gender, hiring manager, and interview stage. HR systems collect a huge amount of data and it’s a case of getting it all in one place.”
Robson adds: “It’s about looking for patterns in the data, and working across different indices to see where something might have an impact on, for example, your change strategy.” A culture that is supportive of data-led decisions is crucial, however. “If you identify a problem in the recruitment process using that data, for example, will it be taken on board and addressed?” asks King. “What will be the action based on what you have found?”
The next step is using data from your systems to predict future outcomes – a practice that is already established in areas such as marketing and engineering, but can be more problematic in HR. “Statistically, it’s more complicated,” says Professor Charlwood. “A marketing team, for example, might want to know which version of a website makes people buy more stuff, or in a factory you might want to predict when a machine will need maintenance. In HR, predicting whether a candidate will perform well and making a decision based on that could expose you to legal risk. Or the data could suggest that 20 people are going to leave, but which 20? It’s much more complicated with people than machines.”
Instead, he advises using the suggested insights from the data to augment future decisions rather than make them for you; improving access to wellbeing support if a data model predicts attrition due to burnout will benefit everyone, for example. Often it’s a case of spotting a potential anomaly that provokes further investigation. “The data might not give you an answer but will give you the question to ask, a red flag or conversation starter,” he adds.
In a time where making reliable decisions is a challenge and the workplace context is constantly shifting, HR teams will need to rely on data more than ever before. In the months and years to come, the post-Covid workplace will throw up questions about safety, productivity and wellbeing, many of which can be answered by data in HR systems. But it’s also an opportunity for HR professionals to improve the decisions they may previously have made based on intuition or experience. As Robson concludes: “Interpretation comes from the tools, but at a more strategic level you can’t replace human judgement and the application of experience.”
Five key takeaways
- The effect of the global pandemic has accelerated the need for data-led decision making and this will continue to evolve
- Effective integration of HR and associated systems can improve data quality and the level of insight produced
- A data-driven culture can boost HR’s confidence to use evidence-based insights, particularly when these insights are acted upon
- HR professionals should talk to colleagues and peers in other functions about how they use data, or the benefits they could get from using people data
- Predictive analytics can suggest potential scenarios, but should inform decisions rather than make them
This is an extract from Ciphr’s book Good Work, Great Technology: Enabling strategic success through digital tools. For more insight into how technology can change work for the better, download the complete book for free, now.