Getting insights from People Analytics is the easy bit. ROI? That’s far, far harder.

People Analytics

Netflix’s recent documentary, The Great Hack, exposed the dark side of data exploitation using its power to influence millions of peoples’ voting intentions.

If the claim that “Data has surpassed oil as the world’s most valuable asset” is true then why is it that the benevolent intention of people analytics to improve workforce productivity rarely leads to the kind of impact seen by Trump’s (first) presedential campaign?

It’s an all-too common story. A forward-thinking HR director invests in a people analytics platform to mine the company’s people data. They’re hoping for insights that might boost productivity, reduce sickness absence, and cut attrition, among many other benefits.

The new system springs into life, using complex algorithm after complex algorithm, trawling the shifting sea of employee data for hidden connections. Finally, it brings forward a tantalising list of “actionable insights”, and then… well… nothing much happens.

There’s a good reason for that. Yes, selecting the right technology is important, but it’s only the start. Convincing a sceptical organisation to make workforce decisions based on data is just as important – and it’s a whole lot harder to achieve.

If people are your great asset, then people analytics is your greatest ally… potentially

People analytics is now established as a source of value for HR. Used well, it can help uncover hidden workforce issues, improve the quality of recruitment, reduce absenteeism, predict which staff are most likely to leave and improve competitive advantage.

And as a result, a growing number of employers are investing in HR analytics tech; 71% of companies see people analytics as a high priority.

Going back to our recently-procured people analytics system as it spits out an endless stream of so-called actionable insights – when does information become insight, and when does insight become actionable?

You might have seen a dashboard which is supposed to provide powerful insight. You ponder for a few moments, hoping that the thing you’re missing will hit you. You scratch your chin; maybe you mutter “that’s interesting” before shrugging your shoulders and moving on to the next equally actionable insight to ignore.

To get anything to be of actual benefit from your people analytics, it needs – as a bare minimum:

  • To be trusted
  • To be understood clearly
  • To be important to the decision maker

You also need to overcome any tissue rejection from your organisation’s culture.

It’s no wonder so many supposedly actionable insights end up on the “too difficult” shelf, slowly gathering dust.

There are lots of reasons for that. You – or your colleagues – might not trust the findings. Your organisation might be resistant to change. Sometimes the promised benefit isn’t worth the disruption, and sometimes there’s not really any clear action to take: it’s “just the way it is…”

None of this is the fault of the data, or the technology. But it stops you achieving the return on investment you expected – and every suggestion or initiative that falls flat undermines HR’s hard-won credibility.

Inertia is cumulative. Every change that isn’t implemented makes future change less likely too.

The problem with people analytics? People.

It’s tempting to think that just having insights into your workforce’s behaviour makes improvement much easier. Here’s the thing some HR tech companies seem reluctant to say: dashboards are not a magic bullet. The truth is, spotting an “actionable insight” is only the beginning of the job.

You still need to work out which insights are worth acting upon, and what you should do about them in the real world. You need to sell the idea – to leadership and, often, to managers throughout your organisation. You need to provide evidence and track the results. You need to deal with real characters, and their real-world concerns.

Getting ROI from your Human Resources technology isn’t really a technology challenge at all. It’s a human one.

Choose your recommendations wisely

When you implement HR analytics, there are a number of things you can do to stack the odds in your favour. The first is to be careful where you start.

If you spot promising-looking patterns in the data, it’s tempting to start highlighting them straight away, to show the value of the analytics as quickly as possible. But some insights are more actionable than others – if the one you’ve identified isn’t sufficiently important, or if there’s no obvious action to take, it could have the opposite effect.

Data analytics can’t usually tell you real-world impact of what you’ve found. For example, imagine you see there’s a 30% chance a given employee will leave; how you respond might depend on other factors about that employee (such as their performance level).

Analytics can tell you the chances, but without context that’s not insight – it’s just information.

Communicate in a way that suits your culture

I remember implementing a recruitment analytics programme at a major financial services company, in the days before anyone in HR had really thought about big data or AI.

For six months, we piloted an AI project to predict who to recruit, based on their application and a series of aptitude tests. 200 hires later, we had doubled the number of successes; absence was down, employees were staying longer, and performance had increased.

The statistics were impressive – and we were excited to report back. But when we showed senior management our findings, they were unmoved. This was a billion-dollar company that had been trading for over 150 years, had more assets under management than Greece, and continually performed well; why fix what isn’t broken?

That mood music dramatically changed when an MD from one of the divisions talked about a customer service agent that had been hired as a direct result of an AI prediction. At first glance, he explained, this was as far from sort of the person they normally hired – they had the wrong background, wrong past experience, lacked the minimum qualifications, and even wore the “wrong” sort of clothes (jeans and a black “The Dude” T-shirt).

But then he heard this maverick employee talking to customers… and witnessed what he described as “the best customer service call I’ve ever heard”.

That recruit became our Customer Service Agent of the Year, and the AI insights became an integral part of the company’s future hiring decisions.

Our statistics were great, but that leader needed to see the real-world impact of the work. And the culture meant that once he was convinced, the organisation followed.

When you’ve found an insight that’s important enough to be actionable, you need to communicate that to the right audience, in a way they’ll understand. Remember, your insight is only important if it’s important to them.

This is often about how you frame the information.

A sales director is more likely to respond to a proposal showing an improvement in Sales Order Values than a reduction in the attrition rate of high performers.

A Medical Director cares more about having six more nurses on the ward than a 0.5% reduction in absenteeism.

A CEO is more likely to notice a 0.5% bottom line improvement than a £250k saving on the training budget.

Collect your data with the outcome in mind

Even with worthwhile insight that’s easy to communicate, many actions fail to happen because the evidence doesn’t seem robust. If a finding seems counter-intuitive, it’s easy to question whether the data’s accurate. That can mean a recommendation gets rejected – or that a good idea is dismissed out of hand.

One way to avoid this is to decide – in advance – what impact is needed to achieve the goals set out in the business strategy and HR strategy, and how to measure that.

All too often, key measures are based on the data that’s currently available, which severely constrains the impact that KPI can have. Switch it around: choose the impact first, then determine which measures are needed – which in turn defines what data you need. If the impact is seen as important enough, data collection and quality can always be overcome.

I’ve hosted many Analytics Strategy workshops with HR and business leaders. My opening line invariably includes something along the lines of: “Today we’re not going to worry about what data we have or don’t have – if it’s important enough to acheive the business goals, we can find a way to collect it.”

This is one of the most cathartic sentiments – it releases the inner business mojo of the leadership. As a result, organisations get the perfect measures to impact outcomes, a series of tight improvement hypotheses, everybody on the same page and most importantly everyone engaged to act.

It’s also important to realise that unlike management information, people analytics doesn’t need 100% accuracy of data. Analytics is about directional trends, and statistically valid exceptions – so a data set missing a few absences here which make 0.1% difference is not important. Generally in-built inaccuracies remain relatively constant. You can still get a good sense of the scale of an issue, and whether it’s getting better or worse, for how long, where and why.

And often, that’s all you need to make a good decision.

Data, technology and people skills, together

I’ve seen many clients transform their businesses by applying people analytics to spot and solve problems, and get a stronger, more engaged workforce. But technology and data can’t do that alone – ultimately, it’s just a smart tool.

Using it correctly also requires the collective expertise of HR and the business – along with a huge dollop of empathy, persistence and patience.

Leave a Reply

Your email address will not be published. Required fields are marked *