myHRfuture

View Original

How to Use Advanced Techniques in People Analytics Projects

My guest on this week’s Digital HR Leaders podcast is Eden Britt. Eden combines being Group Head of People Analytics with the role of Chief Data Officer for HR at HSBC. He is one of the leading and most respected leaders in the People Analytics space and I always enjoy speaking to him.

In the video clip and transcript below, Eden shares some of the work that he has been doing at HSBC in Organisational Network Analysis (ONA), in using Glassdoor data to understand internal and external trends in employee engagement, and how he is using Natural Language Processing (NLP) at the bank. Eden outlines some great practical examples of how HSBC is experimenting with advanced techniques and using Python in their People Analytics projects.

If you’re interested in learning more about how to get started using these types of techniques in your own organisation, then I urge you to watch the video with Eden below, read the transcript in this blog, or to hear the rest of the interview on the podcast then you can listen or subscribe here.

David Green: Can you give us an idea of some interesting projects that the team have delivered during that time?

Eden Britt: I think the org effectiveness piece is really interesting, so you can quite easily structure your organisation in layers or in different hierarchies to take a look at how the organisation is built and it's quite easy to see complexities within that. So some of the things that we look at are senior grades that are low down in the layer structure within the bank. We look at spans and control obviously to see where we may have layers that are not necessarily working the most efficiently. We've also done things from a network analysis or graphing of network analysis. So we take the organisation and it's quite difficult with an organisation our size to do this, but we put the organisation into a node structure where we using either the position hierarchy or the functional management hierarchy to set that node structure.

Once you've got it in a graph database, what it helps you do is to do all kinds of calculations and questions that you couldn't necessarily do in a relationship database. So when we layer onto a graph structure, if you imagine a node structure of nodes that split out with the most senior CEO at the top and the breakout of those nodes down through the organisation, then as we go down each layer and we layer on top through the use of colour or shape, we can start to see where the regional roles are, where certain activity is done. And so that really helps us to understand the structure of the organisation that helps the business understand particularly tough questions around, why are you built like that? Is it by design? Is it by osmosis? You know, it's just happened over time, that this thing has just grown that way and when you talk to the business about those questions, it's much easier to go with something that you found than a blank sheet of paper and ask an open question of what's your biggest challenge and let me help solve for it.

So I think those are areas that are quite interesting and I think we're getting really good feedback from the organisation on that.

Other areas that we've done recently. We did a great project on looking at Glassdoor information. So there's an API available for Glassdoor, which is essentially a URL where you append certain text to it.

And you send that out to the Glassdoor API and it will send you back some information on the external results for a five-point scale of how your CEO is doing, how your senior leadership's doing, the view on benefits within the organisation. So if you take that snapshot of your company, every three or six months, more or whatever you can start to see if there's an external change and then you can look to, whilst it's not in the most granular format, you can look at similar patterns to your internal surveys and then get a sense of the internal lens or the internal voice against the external voice. And so we did that but then we wrote a python script that helped us to use that API to go out and look at a hundred other organisations within our sector and outside of our sector so that we could pull this data back. You get it back, it's a REST API with a JSON structure that comes back, for anyone who's technical who understands that, but what it means is it's a structured format. We can then read into a data set that then we can look at and we can start to do some analysis, on so that was quite cool.

The other thing that I'm really excited about that we've been doing in the past few months is a lot of natural language processing on unstructured data sets where we've asked specific questions. They could be the engagement questions, but they could also be information around the pay and reward program that we've just been through for end of year or other pulse surveys that we run. So we do it anonymously, but what we look to do is cluster the certain phrases into two, three, and four word groups and then we look for nouns and then we look to try and cluster them into some form of view of what the pulse of what the people are saying and that's actually working out really well. So again a python project, leveraging open source, so it doesn't cost anything, leveraging packages that you can download as part of python that take advantage of some of the things that other people have done, particularly where you want to use packages that already recognise nouns and other words and joiner words and things that it might want to ignore. So that you're really focusing on the value of those, so I think those are some really exciting projects.


ABOUT THE AUTHOR

David Green is a globally respected writer, speaker, conference chair, and executive consultant on people analytics, data-driven HR and the future of work. As an Executive Director at Insight222, he helps global organisations create more cultural and economic value through the wise and ethical use of people data and analytics. Prior to joining Insight222, David was the Global Director of People Analytics Solutions at IBM Watson Talent. As such, David has extensive experience in helping organisations embark upon and accelerate their people analytics journeys. You can follow David on LinkedIn and Twitter and also subscribe to The Digital HR Leader weekly newsletter and podcast.