September 2014 ~ Strategic Human Capital Management (HCM) Blog

Here are my notes from my keynote presentation from Employe Benefits Live last week:

Introduction

Datafication - a horrible word (not my choice for the session - and I blame Josh Bersin) but an important idea. However I interpret the concept differently to Josh and don't see a need to 'datafy' HR. To me datafication simply recognises the HR has been 'datafied' and seeks to respond appropriately to this new environment.

Key issue is big data - volume, velocity and variety. Despite criticisms and questions on whether big data exists within HR this is a big deal, and it does exist - maybe not in the strict sense of the volume of data (although SuccessFactors calculate their average system contains 5.5 bn data points which is pretty big) but certainly in terms of variety etc.

That's not to say that small data isn't important but big data requires particular attention.

The big shift is from causation to correlation. Big data says we don't need to worry about ensuring something is causal anymore, correlation is enough. So it's not really about the 3 V's at all, it's simply about a different way of treating data (and OK, I know many HR people have never been that worried about distinguishing correlation and causation but let's let that pass for now.)

But big data is only one part of the big picture. The most critical point is that the more important something is within HR generally the harder it is to measure. It's why I'm so against quotes like 'I believe we should be able to calculate the ROI on everything we do and if we can't calculate it, we shouldn't do it' are so dangerous - they send us in completely the wrong direction.

This leads to three very different zones for measurement and analysis:

The main focus is adding value - traditional HR. In this zone our focus needs to continue to be predictive analytics - trying to understand causal links between items of the HR system
At the most important levels in HR, we often can't measure what's important, at least not in quantitative terms and this leads us to qualitative analysis. Analytics need to be qualitative as well as quantitative.
But at the lowest level of value where there's lots of data we can start doing analytics in new ways where we don't care about causation, we simply look for correlation, that's enough. This is the result of the datafication of HR.

What's Been Happening?

What's been happening is the datafication of business. We can look at this to understand what's currently driving change within HR.

The first driver has been social media - every tweet is a data point. The same applies within businesses as more organisations start to use enterprise social networks.

The second driver in business has been the internet of things - so it's not just people who are communicating together, it's our fridges and other devices as well (help, my fridge is full of spam!)

It's why big data is such an important issue in business - and why McKinsey's suggestion that we're so short of data scientists is such a big issue - obviously for HR as well.

What's Happening Now?

HR is now responding to its own datafication - but slowly. The fact that CEOs rate HR's analytical capabilities so low is a problem. It's not all our own fault however - we've had decades of rubbish technology and it's only in the last five years that this area has been sufficiently invested. But until more organisations have good HR systems in place it's actually very hard for us to respond to HR datafication. But we do need to respond.

The same issues which have driven change in the rest of business and transforming HR as well. Social recruiting generates the same level of data points and these are being used to provide new insights by systems like Talentbin, Entelo and Gild, plus in the US, Glassdoor Job Explorer (Mervyn Dinnen had spoken about social recruiting in the morning session of the conference but had actually talked about ieTalent rather than these.)

A bit nearer to the employee benefits space you've got the same thing happening in social recognition - giving thanks, kudos, badges (a Josh Bersin slide) and rewards.

Another key need for HR is to integrate with data outside of HR. Finance, external vendors (particularly for employee benefits) and others eg Glassdoor's new benefits comparison functionality - using the power of big data to help people make decisions about prospective employers based upon the benefits they provide. (It's early days and I couldn't find any information for EBLive case studies up there - eg Maersk, Sheffield University Juice etc - but it will come.)

The internet of things isn't a big driver for change within HR, but it's near relative, human augmentation certainly is. Augmented reality devices such as Google Glass are going to substantially expand the amount of data we're generating and using. And augmented performance - I was wearing my FitBit and it's interesting to see how quickly it and particularly Nike's Fuelband have moved out of the most innovative into the nearly defunct category with new devices, particularly the Apple watch.

But these are all producing and enabling analysis of data.

And all of this is providing benefits. The best case study I'm aware of in employee benefits is Caesar's Casinos. They've tracked thousands of variables about how tehir 65,000 employees use health insurance medical services, such as their choice of cheap or expensive medical facilities and whether they choose a generic or brand-name drug. Eg they found that at Harrah's in Philadelphia only about 11% of emergencies were being treated at cheaper facilities, versus 34% across all of Caesars. Taking action upon this they reduced this to 17% and between 2009 and 2013 made savings of $4.5 million. I like the case study because it's not a complicated analysis of data, but does provide a high financial value (if low strategic value) benefit.

So basically, HR's experiencing the same enablers, the same requirements, the same potential benefits. We've been datafied and we need to learn how to use the opportunity this provides us.

What's Next

What's next is HR getting on top of this new environment. However, we need to be careful not to be led astray by all the hype. Business has been through peak big data hype but HR is following on behind and we're still climbing up to the top.

We're not afraid of big data as a recent CIPD report suggests, we're just cautious about over investing in an area that is still developing and in which our technologies are still not that mature.

But we do need to develop our capabilities in this area so that we know what we don't know. Eg all HR people should be able to use R (and all the other technology that supports data analysis.)

The good news is that there's a MOOC for that! (actually, there's quite a lot of them.)

But this isn't our only data challenge. PwC's report points out the high proportion of decisions that are still made by gut and we need to influence these decisions too. And just because a decision has been made using data doesn't automatically mean it's going to be a good decision.

The same applies in HR as well eg see UCL research in the proportion of selection decisions which are made by gut.

So yes, we need HR science but we need art and a dollop of magic too.

See this article and the Art of HR feature in October's HR magazine.

We may need HR data scientists in our teams but this isn't the future of our profession - what we need more than this is wisdom artists - people who can extract meaning from data and other information - including the sort of fuzzy information provided to us by our intuition, and develop great, innovative, tailored strategies that drive our organisations forward.

people who can extract meaning from data and information – including the sort of fuzzy information provided to us by our intuition, and develop great, innovative, tailored strategies that drive our organisations forward. - See more at: http://www.hrmagazine.co.uk/hro/news/1145749/business-school-launches-survey-hr-artistry#sthash.ZAxKFcwU.dpuf

Actually all data scientists need to be wisdom artists - the value is always in the questions and the story telling, not in the quantitative analysis itself. But this is much more true within Employee Benefits and the rest of HR as big data correlation only provides one piece in what we need to do - we need to be able to combine in gut hunch and qualitative analysis as well. And all of this needs to be integrated together. It's not as if we'll be able to improve one process just by correlation and another just by semantic analysis - or at least it's much likely that both approaches will be necessary for both processes.

Therefore you may want to have a couple of data scientists on your team, depending on its size, or of course you could just borrow these quants from the rest of your business. But if I was you, I'd want your whole team to be wisdom artists.

So maybe wisdomification would be a better (if even more horrible) word to use?