Developing empirically-driven HR leaders

Some HR leaders are trying to build HR analytic capability by hiring a small team of data scientists or using a consultancy to do pilot projects. The hope is that by doing good, valuable work they will encourage a change to a more empirically-based HR function.

The State of Sarawak in Malaysia is taking another approach and I have high hopes that it might prove to be more effective. Instead of building a small team of highly capable analysts to provide a ‘supply’ of People Analytics they’ve focussed on building demand by making analytics-understanding a core skill in HR.

Re-skilling HR

Earlier this year I started talking to Dr. Sabariah Putit, now Deputy Secretary of State responsible for transforming the civil service in the state. She hatched an ambitious plan - she wanted all her senior HR team to be trained on People Analytics.

We’ve done this work before, working with senior HR leaders to help them understand how analytics can help them in their organizations. However as our conversations progressed it was clear that their ambition far exceeded anything we had previously seen.

Analytics should be a core skill of all HR

In September we completed the first wave of training in Sarawak. During this wave around 110 of the most senior HR people in the state have been through our 2 day in-house HR analytics course. This includes not only people in the central HR functions but also the most senior HR people in various departments and regions of the State.

It’s worth stating that I don’t believe that the majority of these people should or will be doing analytics, at least not at the level which we’re seeing firms build HR analytics teams. Hence the course we’ve developed over the last 3 years focusses on giving HR leaders an understanding of how to apply analytics in their organizations.

Being good consumers of Analytics

My take on the need for most in HR is they need confidence to use data (or information as we call it) to make empirically based decisions.

I want HR managers be comfortable to turn to data or analytics to inform decision-making. It’s important that they don’t get intimidated when more analytically-driven functions are using data to make decisions or argue their case.

There is a mythology or hype about analytics at the moment that you need a team of PhDs to do this work. I want to show that many of the core techniques and approaches are at their heart simple. When a software company or consultancy tries to sell a shiny new product / project based on ‘predictive analytics’ I want the HR managers to understand what it is, how it could be useful and how to ask sensible questions to evaluate whether it’s likely to be effective.

Finally, as organizations do build-out analytics functions in HR I want to ensure that there is a common language between the HR specialists or managers and the analysts (who often will have a non-HR background). It’s vitally important that these groups can come to a common understanding of the real need as quickly as possible and that HR managers feel confident managing analytics projects or shaping them to the needs of their clients.

Ultimately our approach is to provide a start to building a sustainable culture of empirically-driven decision making in HR. Much of what we hear both written and at conferences is about building capability for the supply of analytics within HR. What we’re attempting to do is build the demand.

Focussing on the technical aspects of analytics can be a red flag

Whilst the course that we delivered in Sarawak was called ‘Talent Analytics’ it would probably be better described as ‘How to use information to deliver empirically-based HR’. Needless to say the latter isn’t so snappy.

Too many analysts, especially those without deep HR domain expertise, seem to focus on the predictive capability of their models. They discuss which algorithm they use, or how well it can score on various definitions of accuracy.

Our take is different. We think you need to be focussing on business objectives and achieving those objectives requires not only a prediction but also action. I’m absolutely convinced that often to get the model to perform well you need to focus on variables or features that realistically you can’t do much about. As I’ve previously noted, far too few analysts use a loss function which is absolutely critical when understanding the predicted benefits of taking action.

I also frequently see a desire to overcomplicate the situation. Without a good understanding of the potential ROI of analytics (something that I spend a lot of time discussing on the course using an information economics approach) it’s too easy to over-invest in doing analysis where bringing some much simpler information to the table could be almost as effective. Analytics becomes an expensive R&D function. Our goal isn’t better analysis, it’s better decisions. If you can realise 80% of the opportunity with 20% of the effort you should almost certainly do that first.

The course we’ve developed really has two broad components. The first is number-sense and what’s probably best described as decision-making under uncertainty. The second is an explanation and demonstration of some of the techniques and approaches that we think are the most exciting and useful. We’re not really trying to teach people how to do predictive modelling or social network analysis (though we show it’s not prohibitively hard to do and in Sarawak we conducted a social network analysis to look at collaboration in the HR team to bring it to life) but really to show that these techniques exist and what they can do. All analytic techniques require you to think about the workforce and HR in possibly different structured ways. Understanding them can even challenge your existing beliefs about how our organizations work.

HR analytics is not just for HR

Building a culture of using data to inform people decisions can’t be something left totally to HR. As part of my time in Malaysia I had the honour to present to about 50 of the State’s top civil servants. In some ways I enjoy talking to ‘the business’ about the work we do more than HR. The feedback and engagement we get is both challenging and enlightening.

One of my personal highlights was discussing with the head of the largest department, an engineer, about applying systems dynamics to understand how the workforce develops over time. The realisation that he could apply approaches that are familiar to him as an engineer to his organisation was wonderful. I really hope that he is able to apply it into his organisation, even if only at the level of a mental model.

In my view, a disproportionate amount of effort has been spent creating analytic capability in some HR departments without focussing on the cultural change needed to embed these new ways of working. We could argue that there has often been a ‘build it and they will come’ approach. I think we need to think of this as introducing a new ‘product’ to the HR offering, and it’s absolutely critical that developing the market is at least as important to ensure that results and efforts are sustainable.

ABOUT THE AUTHOR

Andrew is one of the pioneers of the European People Analytics scene. He is the founder of OrganizationView, creator of the open-question employee feedback tool Workometry and the co-founder of the People Analytics Switzerland community.

Andrew chaired the first European People Analytics conference - HR Tech World’s 2013 ‘Big Data’ event and has been co-chair of Tucana’s ‘People Analytics’ conference in 2014, 2015 & 2016. He teaches HR Analytics and Data-Driven HR in Europe and Asia and is a member CIPD’s Human Capital Analytics Advisory Group, setting standards and content strategy for HR Analytics content.

To keep informed of all Andrew’s writing, here and elsewhere please subscribe to OrganizationView’s Newsletter or follow him on Twitter.