Blog: Analytics Can Tell A Story: An Experian Case Study

How can using key data help to reveal new solutions to longstanding issues?

What was the problem?

In 2019, staff turnover was an issue that was attracting the attention of the executive team at Experian, the credit reporting company. “Attrition was 3-4% higher than the executive team wanted it to be, and there had been regular discussions about what actions we should take. We knew reducing turnover could add significant value to the bottom line,” said Olly Britnell, Head of Global Workforce Analytics. Britnell’s team built a model that predicts turnover by analysing around 200 attributes that might drive ‘flight risk’, such as team size and structure, supervisor performance and distance/ length of commute.

How did analytics help them solve it?

Given Experian’s core business capability in analytics, Britnell’s team was able to work with the internal experts who were responsible for the company’s credit scoring model, and they built the resulting employee turnover model using the same methodology. The model provides a risk score for every employee in the organisation, based on data from a wide range of sources – including the HR system and consumer profiling data.

Managers can also use the model to test options for reducing the turnover risk for key individuals or groups, and to build retention plans accordingly – and they can do this at individual, team or unit level. For example, the model showed that being part of a team that comprises more than 10 or 12 people increased flight risk, and the business now uses this insight for organisation design.

Some of the insights that emerged weren’t necessarily surprising, but they provided evidence to back up hunches. For example, if an employee had recently moved out of London, thereby significantly increasing their commute time, they were much more likely to leave. “The model has given us tangible and evidence-based information to prompt action in the organisation,” said Britnell. “What’s powerful is being able to combine data to build an overall perspective on what’s happening to any employee at any given time. This is what allows us to predict the outcomes we should expect if nothing changes.”

What next?

The model is being rolled out globally by region, starting with the areas where turnover poses greatest business risk. The analysis showed that there were significant regional differences in the factors affecting turnover, so the model differs slightly between the UK, Asia Pacific and North America. “We learnt a one-size-fits-all approach doesn’t work,” said Britnell. “Each region is different and the types of actions you might take are different, depending on where you are.” The model is refreshed every six months to check the predictions are still valid. Over the past 18 months attrition has fallen by 2-3%, with a net saving to the business of $8-10m. It is difficult to attribute all of this directly to the predictive model, but it has undoubtedly had an impact.

Why does it work?

Britnell believes the success of the rollout so far comes down to two factors. One is strong business buy-in from senior leaders who value the potential contribution of analytics, combined with strong HR leadership to support the implementation. The other is that the analytics team see their role as being not just to produce a model, but also to make sure the benefits are realised in the business. “I see it as part of my team’s role to keep the HR team’s toes to the fire. We follow up with business partners to check what they are doing about those people who have been identified as high risk. We also find that the more accurate our predictions, the more that gives us permission to push a bit harder to make sure actions are being followed up,” said Britnell.

Want to learn how you could use workforce analytics to support your HR team? Check out CRF Learning’s On Demand course, Workforce Analytics and Storytelling.