We’ve read a lot of blogs on People Analytics. There’s a lot of good content out there. We’ve read about different experiences, methodologies and approaches. Each blogger typically has an area of focus – whether that’s a specific dataset or new analytical technique. But what if you tried to take a completely holistic approach? What if you tried to build end-to-end People Analytics?
What might be discovered if you incorporated ‘any and all’ data from across the myriad of systems collecting information about your people? What if you combined the best in predictive modelling with the latest techniques in language processing? I think you get the idea. We’ve been working to create a complete end-to-end people analytical process and the results have been surprising.
If travel broadens the mind, what happens when you take a journey to the far end of People Analytics?
Preparing for the journey.
Before going any further, we should really define ‘end-to-end analytics’. Broadly speaking we mean integrating three analytical approaches:
- Descriptive analytics: a complete view of historical data.
- Predictive analytics: a statistical view of the likely near future.
- Text analytics: a qualitative view of people’s perceptions and opinions.
These three elements quite often appear as steps on a curve of analytical maturity. Bersin’s maturity model is probably the most popular of those widely shared on social media and blogs. It’s a great model but we might suggest a slightly more nuanced view of maturity. One that when fully realised, is part of our definition of an ‘end-to-end’ People Analytics approach.
Firstly, there are levels of maturity within each of the analytical approaches. Descriptive Analytics can be a simple dashboard displaying just a few metrics and delivering limited insight. It can also be an interactive insight visualisation platform that allows users to explore different analytical paths in real time. It can also involve combining and cleaning data from a myriad of systems, including operational systems containing performance and productivity data not normally handled by the HR team. Clearly there are degrees of maturity within descriptive analytics.
Secondly, maturity should also be measured in terms of accessibility and scalability. Fority is a question of integration and business impact. Mature analytical approaches are coordinated with each other and incorporated into strategic tools (e.g., workforce planning) as well as core business processes.
We’ve been working with some leading businesses to deliver all three analytical approaches at the highest level of maturity. It’s not easy but the insights and impact make the effort worthwhile.
How we got there.
The detailed steps of building a People Analytics capability are too numerous to list here. Especially when deploying all three analytical techniques at the highest level of maturity. However, we believe the key to doing this quickly and effectively is a new type of working model.
When working with any new client, we use a ‘tech-enabled consulting’ model. This means (a) using technology to short-cut laborious analytical processes and (b) deploying expert resources – in this case a team of data scientists, organisation experts and technologists. We work closely with our clients HR teams through-out the process, listening carefully to the specifics of their businesses and gathering the all-important context required to interpret the data.
Quite quickly, the joint team can build analytical tools together that create a sustainable set of advanced capabilities. Collaboration, access to experts and intelligent tech enables us to get to ‘end-to-end’ analytics must faster than was ever previously possible.
The Benefits of Travel.
Seeing more of the world is a good thing according to Twain. Travel is ‘fatal to prejudice, bigotry, and narrow-mindedness’. But what are the rewards of making the effort to complete an end-to-end People Analytics journey?
We’ve uncovered many valuable, hidden insights in our work. Here are some those findings that we think demonstrate the value of a mature end-to-end process that combines description, prediction and text analytics:
- FUN. It’s a small word but it’s very powerful. When people write about their workplace, there is a huge difference between ‘great’ and ‘fun’. By combining predictive and text analytics techniques, we can see that FUN matters, especially when it comes to future absence, attrition and productivity rates.
- RELATIONSHIPS. When we create metrics, we don’t just look at individuals but also the people around them. Great descriptive analytics plugged into a strong predictive model shows that teams copy each other’s behaviour over time and impact of a good manager can be significant – future performance is created by good leadership.
- CHANGE. Few people like change. Disruption to teams and organisation structures can be very predictive of lower future performance and absence. Quantifying the impact of this and then linking it to workforce planning is very powerful – it allows HR teams to model the impact of a new organisation design on future key performance indicators.
- PROCESS. Analytics can also help measure the effectiveness of HR interventions and tools. By integrating predictive tools with recruitment assessment data, including free text answers, we can see which questions are most effective at talent spotting – if a question isn’t predictive of future performance, replace it to improve overall assessment accuracy.
- FEEDBACK. Many deride analytics as being ‘theoretical’, providing statistics without testing to see if remedial actions are viable. Surveying capabilities that are linked to ‘intelligent language processing’ allow HR teams to engage employees in conversations about change – testing possible actions, hearing ideas of how to make organisations more effective.
The detailed insights we find, together with the great HR and People Insight teams we work with, are far too numerous to list here. There are predictive relationships that appear common across many different businesses and some that are surprising and unique to each individual organisation.
This wasn’t meant to be a list of insights that applies to all. We just want to highlight the possibilities of mature analytics, and share news of our own progress as we develop more sophisticated HR analytical techniques and technologies. We’re excited by what we find when we work with new clients and even more excited about what is still to be discover.
End-to-end HR analytics isn’t necessarily easy to achieve. It requires some work and some very clever technology but it is far more affordable and far faster to implement than ever before. What seems important, from our perspective, is that the rewards of making the investment are clear.