Lets get analytical

A week or so back I attended an HR analytics conference run by Symposium Events. I'd done this because of my growing interest in both predictive analytics and in making our existing static HR data more meaningful. 

Here's my thoughts on the topic. 

My interest has been growing for the last year or so, promoted partly by the creation of the Torus Group and the need for that organisation to realise some big efficiencies over the next few years in a changing and more challenging environment. I started wondering how much HR could contribute to this and began looking at what we did and what others did. 

I concluded that too much of HR data, both in and outside my organisation, was just sitting there and wasn't being actively sought by the business to drive better decision making and improve performance. It was static, retrospective and didn't go much beyond benchmarking in terms of analysis. 

In short, too many HR departments prided themselves on their ability to manipulate systems to extract mounds of data that was then published to people who didn't look at it, or just looked at it and nodded. 

That wasn't good enough then and as my thinking has developed, I know it's definitely not good enough for the future if HR is to justify its seat at the boardroom table. 

So where do we need to get to?

The conference underlined for me the importance of an MI Analyst role in HR, not just to own the data and its production, but to draw conclusions and offer insights. Particularly for organisations with over 1,000 staff. This role would help us to move towards modelling of data and root cause analysis to identify issues and offer solutions. From there we could move to scenario planning and predictive analysis, and link the data analysis to risk analysis and strategic planning. I don't think this person would have an HR background either, probably better that they don't.

I think, though, having a data analyst in the team is only half the solution. We need to understand what the problem is we are trying to solve with improved analytics. What are the most important things the organisation needs to focus on and which interventions would have the most impact on these things?

In the conference, questions were asked about what aspects of culture impede customer service, and what we can do to improve innovation. You'd also include things like an analysis of complaints or compliments data against specific groups of the workforce, maybe at an individual level, to look at what your best performers do to influence customer satisfaction, or what the worst do. You'd look at recruitment data and where you found your best performers, and whether certain recruitment sources provide better performers than others. You'd also want to explore the link between recruitment and performance data, and absence/performance and behavioural data. There's loads of possibilities. 

Basically, HR needs to find data correlations and use them to start a wider dialogue and build credibility. Don't rush in throwing statistics at the organisation without having some context and understanding the organisational problems first.

I tweeted at the conference that HR needs to focus on what will keep Executives awake in the NEXT few years, not what has been keeping them awake this year or last year. 

Of course there are obstacles, as the conference highlighted. The lack of analytical skills or investment in analytical systems is one, as is lack of support from management or hard to access or hard to correlate data. But that shouldn't stop you from having a go. 

To combat the lack of skills, look to develop them in house. Get HR BPs involved in brainstorming correlations in data, generating hypotheses and going looking for data to support or disprove them. HR BPs should be challenging the data, asking so what, and why, as often as possible.

I think it may be easier to start with a pilot of predictive analytics, in one small area, perhaps an easily benchmarkable one that is present in other organisations, like call centre staff. Do some data correlations there, perhaps using one of the data sets that is easy to identify and measure, for example diversity statistics. Although at the conference I tweeted some of the possible areas of difficulty this could lead one to, but at least it's a start.

Ultimately with predictive analytics we could get to Amazon-style customisation of jobs and work for employees, with work and work environments tailored to produce high performance based on how we know performance and engagement has been generated in the past. I might do a follow up blog on this. 

I'm looking to test predictive analytics soon to see what can be done and learn from it. If anyone has any lessons they can share, I'd be interested to hear them. 

Till next time...

Gary

Ps in other news, I turn 40 this week, a major milestone and a natural opportunity to stop and reflect on things. There's a few bits of celebrating to do first though. Or is that commiserating?