I'm going to break with tradition and use the words of today's guest to frame this week's episode:

To navigate the increasingly complex world of talent, HR needs to grow more quickly in to a strategic advisor. More companies will need CHROs and they will need to have an equal voice alongside CEOs and CFOs in the most critical business decisions. In the coming decades of disruption the management of talent will become the main differentiator of high-performing organisations, this requires HR 3.0.

Those are the words of Keith McNulty our guest today on the Digital HR Leaders podcast. Keith leads the People Analytics and Measurement team at McKinsey and is also a prolific and renowned writer which saw him recognised by LinkedIn as a top voice in December 2017. Keith is a deep thinker and one of the most knowledgeable and visionary leaders in the space. So I know that listeners will enjoy this episode of the podcast.

You can listen below or by visiting the podcast website here.

In our conversation Keith and I discuss:

• The aforementioned HR 3.0 model, an analytically sophisticated and agile function populated by professionals with strong business acumen and problem solving skills

• We talk about the skills and capabilities you need in a people analytics team and we reflect on highlights from Keith's journey of building the people analytics and measurement team at McKinsey, which he's done over the last two and a half years

• We also reflect on the people analytics space and also what drives Keith to share his knowledge and experience with the community

• And like with all our guests we look into the crystal ball and ponder what the role of HR will be in 2025

This episode is a must listen for anyone in the people analytics role, HR and business professionals interested in how people data can drive business outcomes and CHROs looking to build or scale their HR functions.

Support for this podcast is brought to you by Perceptyx to learn more, visit perceptyx.com.

Interview Transcript

David Green: Today, I'm delighted to welcome Keith McNulty who heads People Analytics and Measurement globally for McKinsey to the Digital HR Leaders podcast. As well as being a very prolific writer on the topic and other associated topics around HR. Welcome to the show Keith. It's great to have you here.

Keith McNulty: Thanks for having me.

David Green: Can you give us a quick introduction... Give listeners a quick introduction to yourself and your background?

Keith McNulty: Yes. So my brief life story is that I started my professional life as a mathematician and after realising that I didn't really want to, do that long term. I ended up joining McKinsey and as a McKinsey Consultant, 20 years ago now, quite a while ago, it doesn't seem like that long. And after working on some client work for a period of time I became really fascinated in the question of talent. In particular I felt that the world around us was not measuring and analysing talent in a really systematic or mathematical way, it is still very much done on gut reactions and that sort of thing. So after about four or five years at McKinsey I started to focus much more on the topic of talent. I moved into an internal role. I started working on creating better processes to measure and assess talent and understand talent. And then that developed into in the last five years or so getting much more involved in Talent analytics because what we found through having better processes is that we were collecting a lot more data.

And that data was suddenly becoming available to give us insights. So for the last four or five years, I've been much more involved in the field of people analytics. That led to me taking up a role as Global Leader of People Analytics internally at McKinsey about two years ago.

David Green: Okay, great. Well, we're definitely going to talk a lot about people analytics today. I know that. But I thought firstly that let's look at an article you published last year. One I really enjoyed around HR 3.0 and I know you did a quite a few presentations around the world on that model and obviously analytics is a big part of it.

Can you summarise what you cover in the HR 3.0 article and particularly the three stage evolution of HR as you see it?

Keith McNulty: Yeah. I mean maybe before we go into that, I'll give you a little bit of context around why I wrote that article. At the time I'd been attending a lot of different, HR events and conferences and people were talking about people analytics a lot, but I wasn't really hearing a proper explanation or narrative as to why it was important. People were just saying you should be doing it.

And at the time I was doing a talk at a Santa Fe Institute event and one of the other speakers there was a fairly well-known Economist called Bob Allen from NYU and he showed some fascinating research which showed some statistics on earnings and productivity during the original Industrial Revolution in Britain in the late 1700s early 1800s, and then he compared that to the same statistics today and it really blew me away what he showed which was that we are in the middle of one of the greatest disruptions in history, related to people.

And that together with some research that I saw from the McKinsey Global Institute, which talked about the size of the disruption in terms of Automation that's going to happen, assumed by the year 2030. I started to realise that this is really what's driving the need for change in human resources. That to be able to operate effectively in a massively change environment in the future, HR can't keep operating the way it has been in the past. But I also acknowledge that there has been some movement. So you see around the time that this disruption started which was actually in the late 70s. It's been going on a lot longer than people think. We also see that coincide with changes in the HR function.

So if you go back to the late 70s, HR was really just a back office function there was almost no analytics involved, lots of paperwork, lots of Process Management. If you move to the late 80s and early 90s, we start seeing the beginning of a move towards a much more professionalised HR function and in particular more of a service provider to the business and I labelled that HR 2.0.

It was a distinct change in the role of HR. And it started in the late 80s early 90s although for some companies it didn't really start happening until we got into the noughties. But that was HR 2.0.

But if we look to the future and we see the size of this disruption because it's only just beginning and nobody knows when it's going to end. That disruption means that HR really has to step up even further.

How will we be able to understand and analyse? How will we be able to manage talent in this disrupted world where talent's going to move through businesses much more quickly? Where finding people is going to be more challenging and where retaining them is going to be more challenging and there's going to be some major changes needed to the function to make that happen and certainly if we keep going in the service provider model, that's really the most common model today, companies are really going to struggle in that future talent world. So the idea behind HR 3.0 is this fairly substantial step up. We really need to make to be prepared for that future world of talent.

David Green: I mean it's fascinating stuff and I think all of us that have worked in HR or work in HR, we can see that the expectations from the business are certainly changing.

A constant charge levelled at HR professionals, and I think this is where people do it from a very generous level, is they don't necessarily have the Business Acumen that's required to make that step change to the 3.0 model.

How can HR professionals get that Business Acumen?

Keith McNulty: Yeah. First of all I think it's a valid point.

I think in my 3.0 model there were really three fundamental changes to the business that are required. One is much more data-driven, more analytically driven. The second is that the organisation needs to be more agile. So instead of really swim lane service line type approach to HR there needs to be a combination of that with a much more generalist approach to the function and then third of all to enable that you need the broader Business Acumen. Not from everyone, but from a much more substantial number of HR professionals than we have today.

So in terms of that I'd like to be able to say that education programs have caught up with that and are starting to offer ways of developing people in that way, but I don't think it's actually true and often the education sector tends to take a while to catch up with the Enterprise environment and  work out what's needed.

So there's not a huge amount of opportunities out there for people to take dedicated HR programs or dedicated HR learning that really covers a lot of this new stuff like analytics, like agile, and so it's really I think up to... Immediate changes... Are up to two groups of people.

I think first of all, the HR professional themselves needs to motivate themselves to go out and find that learning. It is there. It's often not on HR programs. It's often in much more general programs. For example, it's very easy to learn about agile and what agile means if you go out there and look for it, but you need to actually do that yourself.

You won't be led into it in the way that education is.. The education is structured currently, I think. But also I think we need HR leaders to be more brave in introducing this stuff to their business. So if you want to have good people analytics bring in experienced psychometricians and other people that are familiar with analysing talent and people. If you want to be more agile bring in some Scrum Masters, listen to them, help them transform your business to a more agile way of working and the third part of that is the talent part, you need to start thinking about bringing in individuals who have a broader range of experiences in the business and not just people who are one hundred percent HR experienced. So in order for HR professionals to operate in this environment, they need to have great familiarity with the broader aims of the business. So bringing in individuals that have that broader exposure, not completely replacing existing HR people, just enhancing them with those individuals would be a fairly transformative piece of work. And that's some of... Among some of the things that I've been doing in my group and in my function, but I think that's one of the, those are some of the things that need to be done. And I think it's dependent on the HR professional themselves and the leaders of the business to really drive that along and make that happen.

David Green: I mean certainly from all the organisations that I talk to, especially the people analytics leaders, it's a real challenge for them to try and create that culture, that data literacy within their broader HR population. And this is a constant thing. How do we excite enable and equip HR business partners to have the right quite conversations with the business so we can work with them on the right projects.

Is that something that you're trying to solve with some of the education you're doing at McKinsey?

Keith McNulty: Yeah, data literacy is known to be a problem in enterprises more broadly and in HR functions, it's probably a little bit more of a problem because historically there's been this understanding that it's not a very data-driven field. Now that is changing but as with all changes, everybody needs to catch up and so, I'm lucky enough to work for a very data-driven organisation. So I don't see the same degree of a challenge where I work, but that said in conversations I have with clients and other externals I do come across a lot of concern around how do we get our people to take data seriously, to understand it, to be confident working in it. I think there's a lot of nervousness and worry among HR professionals because they've not had to do this before and now it's a whole new ballgame.

But, I think the way of the world is that everything is going in that direction. Everything is moving towards data. So I think people need to be brave and they need to step up and they need to engage with data more. Mistakes will be made and that's part of the learning process but by engaging with data related problems that will force you to go out and ask questions and learn and get the information that you need to become more comfortable working with data. So, I think it's important that we stop this concern that people have that they're going to make mistakes or they're going to worry they're not going to be successful working with data. All of that is a learning process and to move along in that learning process you have to be brave and engage with the data in the first place.

David Green: And be comfortable with making mistakes because you learn from them.

Keith McNulty: Absolutely

David Green: And I guess one of the challenges I see as well with analytics is... You mentioned the swim lanes and obviously HR has traditionally operated in swim lanes for a long time. Analytics goes across those swim lanes so you're almost forcing people to come out of their swim lanes a little bit more which actually can be scary. But I guess is an opportunity as well.

Keith McNulty: Yeah, I mean later on I think when we talk a little bit about how my team operates, a big part of that is that HR still has operational pieces inside it, the organisations will often have a different organisation looking after recruiting versus, employee  satisfaction versus the other elements of HR, but an analytic function can't be organised like that because it would be massive if it was so what you need is you need individuals to be able to engage across those lines and I think this group of people who can engage with a more general set of HR questions really in terms of seeing them as business cases rather than swim lanes. That's going to be critical to operating a strong HR function. Those kind of individuals will lie over the people that actually operate the swim lanes themselves.

So you'll have a combination of a well organised operational function that can operate within the swim lanes when necessary but you'll also have a group of people that can take on strategic questions that cross across all of those at the same time.

So that leads on quite nicely to the work you've been doing at McKinsey with the people analytics and measurement team.

David Green: You've been leading it for just over two years now, what have been some of the key milestones along that journey?

Keith McNulty: Yeah. It's moved very very quickly and it is great to work in an environment where you get a lot of support from the business to build a function like this because you can build it very rapidly if you're given that support.

I think if I go back to the beginning of my involvement in this, one of the things that I think a lot of other people working in people analytics will find familiar here is what the team was like at the beginning and it was very much a reporting team. So the role of that team was to provide information to the business on historic and current issues related to talent and a lot of that reporting was very manual.

So there was a lot of human effort being taken up in just transmitting data from one place in the organisation to another. Now as I looked at that and said if we want to build this into a world-leading people analytics team, what would be required here? The first thing that I took on board is that a people analytics team will always have reporting and metrics as a fundamental part of it and it was important that we get that up and running effectively in a way that wasn't highly manual so that we could then start to build to do some of the other things that you'd like a HR 3.0 and people analytics function to do. So about 50% of the team focus on reporting and metrics and part of that is being involved in defining what metrics are and owning them for the business.

But also getting them out there to the people that need them in the time that they need them and the elements of that are what I would call ad hoc, which is where people just need certain data and it's not data that people will usually ask for and so we will have members of the team that just deal with that sort of thing. But then more importantly we've set up people that work in a kind of a devops model where we actually develop our own internal software products to help people serve themselves with the data and that's been a big move. I think in the last couple of years which is to enable the business to access data for themselves, but therefore vastly reducing the amount of time that people on our team need to spend actually getting that data and procuring it.

So there's a big technical element of the team, which is developing products for people to access data and metrics, that frees up capacity when you do that. So that then allowed me to think about what are the other parts of a group like this if they want to really, you know, hit the HR 3.0 Benchmark and this is where we move on to what we were speaking about earlier, which is addressing strategic questions of the business that cross many functions using data. Now to do that you need two very fundamental groups of people.

The first is you need data scientists. Because you're going to be playing with that data in all sorts of different ways. You're going to be using some fairly advanced modelling. So I have a cross functional data science group that have expertise across a broad range of areas, like statistical modeling natural language processing.

We do a lot of work in graphs and that group is a fundamental part of solving some of these business problems, but they can't work just on their own because data scientists have a lot of talent but one of the talents they're not known for is their interaction with the business. And data scientists tend to like working on the actual details of the problem.

They don't tend to enjoy the parts that are involved in the business communication and those sorts of things. I'm generalising a bit but I think it's relatively fair generalisation. So you need to enhance the data scientist with a group which we call translators and translators do that communication with the business.

So they help define the problem. They help turn it from a kind of fairly ambiguous type of concept into an actual set of steps that you can go through to solve the problem. And then that problem is communicated to the data scientist. They work in partnership with the data scientist to take the analytical steps needed to solve it and then they think about how do we communicate that back to a population which doesn't understand data science.

So how do we make it intuitive to the population that we deal with? So those two groups work very much hand-in-hand on a wide range of initiatives and that model I think is, starting to look like more broadly within the analytics space, not just HR, the successful model of operating in an Enterprise and then behind all that if you think about the reporting and metrics, you think of the data science, if you think about all of those things that all relies on having really good quality data, you can't do good work if you don't have the data. It's rubbish in rubbish out principle. So one of the groups that we started to develop very quickly is the data Engineering Group and the purpose of the data engineering group is to make sure that our data is a very very high quality that we manage it well that when changes happen to it, those are well  communicated so people know about it and having that as a foundation is very very important because it means we... I'm sure that you hear a lot that one of the biggest challenges that people have not just in the HR space in many areas is data quality, right? That some people say 80% of data science is data cleaning.

So I wanted to get away from that. And so if we enable a group of data engineers to ensure our data is clean, well structured, understandable that takes that away from the data scientist. So the data scientist can really focus on the actual analytics itself and drive to high quality answers. So those are really, you know, that's how the team has been built around those core principles

David Green: And in terms of growth obviously you don't necessarily need to go to how many or how big the team but it's grown significantly presumably?

Keith McNulty: Yeah. I mean it has I mean not as much as it would have had to if we hadn't addressed the data self-service and the automation aspects. Being able to automate the delivery of data has taken a huge amount of capacity, manual capacity away, which we can then repurpose for many of the other things that we would like to do.

But another reason the team has grown is because we want to get really specific expertise into it. So, to work in this space now, you need to know a lot about natural language processing because a lot of HR data is text data. You need to know about ways of understanding connections.

So we do a lot of work around graphs and networks and those sorts of things. Psychometrics is a big thing. In the end we're analysing people and talent. We need psychometric expertise. And then you just need to be able to do all the statistical modeling that data scientists do. You need to know about models, which one's work for which problems.

So a lot of the reasons why the group is bigger is because we needed that specific talent to come in so that we had that expertise.

David Green: As you said, there's a lot of different skills that are required to really be effective in people analytics.

So some things I need to touch on from there. So firstly by automating as much of the reporting as you can you're almost generating an appetite for data, I guess in the business they can start to then understand actually we can get this stuff out of HR. This is quite exciting and that probably generates questions.

Keith McNulty: Yeah. I mean that's never been a problem at McKinsey. McKinsey is very data hungry. They have a highly data-driven culture. And that is a wonderful thing because it means that we operate in an environment where we're highly valued but it is challenging as well because the pace at which you have to deliver and the complexity of some of the requests are very very high.

So for us it's been less encouraging people to ask for data, that's always been there. But it has been an interesting journey to try and say how do we enable these products to be able to answer all the different varieties of questions that will come in from various parts of the business about our talent. But there is definitely a shift in culture that you can achieve and speaking more broadly here, which is this idea and I think a lot of companies still operate this way where you know, there's a few people sitting at some desks and then somebody else will just say to them can I have the data please and they're the only people that can go and get that data.

That's a massive lock on the resources of the business. It means that the people that want the data can't get it at the speed they need it and it also means that often that group are overworked because all of that stuff is coming their way and they only have manual ways of solving it. So this idea of being able to expand your technology and create an environment where you can automate that just opens data up. You obviously have to do it in the safe way and think about all the guard rails around the data. But it opens the data up and it allows the business to be able to access it more freely.

Which aids decision-making, makes things more rapid, more agile. And so that's really a huge thing. And that's why I always say... and there's a few people out there that have said, that reporting and metrics is not really part of people analytics. It is one of the most fundamental parts of people analytics and you have to get it right and you have to know how to deliver in an agile and rapid way and help the business make decisions because talent is moving so quickly now as I mentioned earlier, so the data needs to be able to move quickly so that you can make decisions about talent.

David Green: And I think what's interesting is a lot of companies that have separated people analytics from reporting tend to bring it back after a period of time and maybe it's because they come to the same realisation that you've just explained.

So one of the other roads I think is really exciting is around the translator role. We hear quite a lot about... and I know there's some excellent stuff published by McKinsey around the importance of translators in any type of analytics, whether it is people or elsewhere in the business. Where are you finding the translators from? What sort of backgrounds have they got because it's quite a specialist skill in many respects.

Keith McNulty: Yeah, I mean my personal belief is that translators need to come from the business because each business operates in a different context and you need to understand the culture, the way it works, you need to understand why are these questions being asked? What is it that's driving these questions? The way I describe it is if you take a scientist out of a scientific institution or university and you ask them a particular question about any field really they will have a particular approach and it's often quite generic and it will consult, a lot of the general research on the topic and what they'll deliver back might be useful but not necessarily something the Business Leaders feel they could implement.

Because it doesn't have that, specific understanding of the context that the organisations operating in. Now, if you've got a translator who you've brought from the business when that question is asked the translator is much more likely to say to themselves: I know why this question is being asked. I understand what the challenges of the business are and why we're asking this question. Which means that they can then communicate it in a more effective way to the data scientists that the data scientist is really working on the core aspects of your solution, but then when they bring it back to the business, they cannot just say this is what we've seen, this is what we've discovered but they can also say this is what you can do about it. These are some of the solutions that you can put in that might work in our context because they understand the business. So translators in our group have been sourced from the business and I think in most good analytics functions, you need to source your translators from the business. So people who have worked in broader areas of your business, but have a real interest and love of analytics and data, those are ideal profiles for translators.

David Green: And I think perhaps one of the challenges that other organisations are facing is that sometimes they are trying to get people from HR to do this role. As you said, they haven't necessarily got the business experience. And I think there you have a people analytics team on its own with quite technical people in it, and they're relying on their HR business partners who maybe don't have the understanding of the business and certainly not the understanding of analytics.

Keith McNulty: Yep. And I do hear from a number of people analytics professionals this feeling of being a bit stranded in the organisation and a bit lonely not understanding where they're having the impact and I think a lot of that is driven by the fact that that communication is just not there.

So the flow from the business decision maker through to the people analytics group and back again is not well designed and I think that's a relatively common phenomenon, but the more that people can think about bridging that gap and thinking about bringing translators in the easier they can address that.

David Green: So I've seen you speak Keith at a number of conferences and other members of your team actually and I know you're doing some really interesting work at McKinsey. Can you give listeners a couple of examples of some of the projects that you've done and the outcomes that that's delivered as well.

Keith McNulty: Yeah. So one thing which we've been very involved in, one of the things I'll mention is the name of our group is called people analytics and measurement for a reason. So when I thought about what people analytics is, I realised that you can't do good analysis without good measures and one of the big challenges in HR is the measures aren't there.

So when we look across the work that we do in the business one of the questions we asked is where do we need to get better measures to better enable our analytics in the future and there are a few areas where we're doing some interesting and fairly cutting-edge work. So one area is this idea of how do we understand people's thought processes?

This is really interesting because McKinsey, has a really strong reputation for how it does it's recruiting and it's interviewing and one of the reasons why it has that reputation is because in our interviews, we focus very heavily on understanding how people think about problem not was it to get the right answer or not, but how they think about the problem and one of the challenges we've been dealing with for a number of years is you can only interview so many people. So how do you open that up? Are there ways that technology can allow us to understand among a greater number of people how people think and in order to be able to tell what type of a career would best suit them in, would they make a great McKinsey Consultant, would they make a great digital designer or a data scientist.

So we've been working on a digitised assessment that really involves tasks not questions. So instead of a traditional test where you'd be asked the question you had like A B C or D and you had to pick one of them. In this type of assessment you're actually put in front of the computer screen and you have to do stuff and the things you do help us understand how do you go about solving a problem. And that helps us get measures which we're hoping in the future will help us to be able to better understand types of problem solving or problem solvers people are and I'm absolutely convinced that we can get much more intelligent at fitting people with careers that best suit them.

So a lot of people are unhappy or they perform poorly in their careers because they've just been poorly matched with what they do, what their skills are. So getting a better understanding of people's skills will help us better match, you solve problems like this. So therefore you would be best fitted in an environment that does this type of work. So right now we're really focused on problem-solving for that and we've already started implementing this assessment in a number of scenarios. We've got incredibly great reaction from the actual candidates themselves who think it's much more interesting than doing a paper and pencil test probably unsurprisingly, but certainly excellent response from the candidates. And we're now starting to think about how we expand it and how we open it up to more people. And hopefully in the future we'll be able to move outside the problem-solving realm and start to think about is this a way that we can understand, what people.. How people behave and what types of job would suit their personality traits and those sorts of things. The whole personality space is much more difficult to measure much more challenging psychometrically. So that will be a little bit further down the line I think. Certainly, what we're developing here is probably, the most cutting-edge psychometric tool that's out there at the moment. Something I'm very proud to be working on but I think, I also feel very fortunate to work for an organisation that allows me to work on something so amazing.

So that's one of the things. I think another area where we're trying to understand better measurement is this concept of connection and interaction. There's a very good expert on organisational design called Naomi Stanford and I watched her speak at a conference I was at last year and one of the things she said which really fascinated me was what is organizational design she asked?

And I think a lot of us understand organisational design to mean tree diagrams and who's who's boss and those sorts of things. But what she said is organisational design is the study of how work flows through organisations. And that made me realise that to understand how people interact is going to be a critical element of people analytics.

It's one of those things where we can really raise the discipline above what we've studied in the past and bring it to a whole new level of understanding so we've been doing a lot of work around networks. We've been moving our data into graph format so we can understand interactions and links between individuals and what do they know? Who do they know? What have they done? When have they done it? And allowing us to analyse through connections rather than the standard way of looking at data, which is in tables. So we've been doing a lot of work around that and that is starting to feed into a lot of the general data science that we do. So when people ask us to do modelling and things we can start to bring in measures of connection and things like that into the models, which we weren't able to do before.

I think the other area which I've mentioned which we're doing a lot of work in is text and language. So much of what comes in through any talent process is text and language. When you evaluate people, often you're evaluating them on the basis of comments that people have made about them those sorts of things when you survey people or get their opinions on things. It's often in the form of text. And I think a big challenge for organisations in the past, especially if they're large organisations has been wading through all of that, a human being can only take in so much they can only understand and process so much.

So if you're an organisation, you've got like 10,000 comments from an annual survey. How is a human being supposed to group that into a set of topics? That's very very challenging. But within the natural language processing field, we have algorithms that can do that. So we've been working a lot on being able to automate the process of breaking up large amounts of text into distinct topics and being able to say, on average when our people are answering a question like this. These are the three or four topics that they tend to focus on. And that's fantastic for our leadership to be able to have that level of synthesis.

So, those are some of the things that we've been really pushing the boat on out on recently.

David Green: All fascinating areas. And I think that we're only at the early stage of both the analysis we can do on text as it relates to people within an organisation and then networks as well. I'm seeing some really interesting projects going on around the world around understanding networks, understanding the importance of sales performance, customer loyalty, all those sorts of things.

Keith McNulty: Yep.

David Green: And as you said how the work flows through the organisation as well rather than just looking at the org chart which is how we have traditionally done it. So I suspect that you'll be doing more of the same for the next question. But what's next, what's on the horizon for the next 12 to 18 months?

Keith McNulty: Yeah. So I think all of those are still as you say in relatively early stages, we've seen some interesting conclusions from some of our work. For example on the network stuff. We've done some analysis around what does it mean to be connected in an organisation, is there a certain minimum level of people you should know and we have started to learn that there are lines where you can say if people are below that level of connection there's a much greater likelihood that you might, that they might leave the organisation in the future. That sort of analysis. So we're starting to see interesting results that are coming out from the early stages of work, but there's a long way to go in terms of enabling your data model to be able to do the most advanced analytics in this space. So a lot of our focus will be just on pushing those areas I already spoke about. I would say in addition to what I've said, there's two areas that we will continue to push on. One is continuing to develop this self-service model. So we've come up with some pretty good frameworks where we think about how do people access data in an organisation? And what levels do different people need data? And what journeys do they go on when they look for data?

So I think that's one thing that we'll continue to develop our self-service model against that framework. So that not only are we providing data, but we're providing the right data to the right people at the right time. So pushing that further is very important. Another area is cloud computing or distributed computing. So a lot of, especially in large organisations, a lot of the people analytics topics can get into the realm of huge amounts of data, particularly if you're dealing with text data and you can be very limited with what you can do unless you can have access to some sort of distributed or cloud computing resource.

So we're starting to move towards doing more of our analysis in large kind of computational arenas so that's a kind of technical area that we're pushing on. The last thing which gives us a creative edge to what we do, which I think is really really important, is we're experimenting with data art. So it's this idea of how do you get people to engage with data in your organisation? And one of the things that I thought is, we've been thinking a lot about making data relevant and making data accessible. But how do you build an emotional connection with data? How do you make people go wow, I really want to see that. I really want to look at it. And so we've been experimenting a lot lately with creating art that tells the story of our firm which is really interesting. I'm cautious about the term storytelling in our field, but where it does work very well is historical data because there is a story to be told. So we've been working with a few of our locations to use data to tell the story of their history and that's been really amazing work and some of our graphic design experts have been really involved in that.

And it's really just a passion project for us, but it's really developing into something that I think will have a bigger application which is this idea of how do you build that emotional connection with data and get people to love data more in your in your organisation. That's something we'll be building more of I expect.

David Green: So I've read the article you wrote about storytelling. So I've got the context as well, but just for the listeners that haven't read that article yet, why the concern about storytelling?

Because a lot of people talk about it.

Keith McNulty: Absolutely. So one of the problems when you don't have a data literate or research literate organisation is that people can very quickly jump to conclusions and I think one of the problems in most enterprises is that people say I'm going to do an analytics project now. And I'm going to do it so that I can prove something. And that is the wrong approach because that puts you under pressure to prove it. What you need to be saying is I'm going to do an analytics project to get the answer to this specific question and the answer might be you were wrong. Okay, but it's really important I think to not be taken in by that type of desire for storytelling and that is a big pressure in most organisations to say I have an objective here and I want the story to come out that looks like this. But from analytics point of view, you have to go from the other direction which is to say I'm going to analyse this and see what it tells me and then when I know what it tells me I'll then think about how to present that back, but it might tell you nothing so you shouldn't be under pressure to tell stories in that way through your data. But of course as I said, if you're talking about historic data, then that's a story that's already happened. So there's a lot of potential for stories...

David Green: And you can make that story more captivating by using data and art as you said...

Keith McNulty: Absolutely.

David Green: We could probably talk all day but we'd better move on.

What excites you most about people analytics and where it could develop and then the flip side to that. What's your biggest concern currently about the space?

Keith McNulty: It's interesting. So if we go back four or five years when people analytics really burst into it's boom time. It did that on the basis of something which I believe never really materialised which is this idea of highly advanced prescriptive algorithmic modelling and predictive modelling and this idea that you could predict when someone's going to leave your organisation and those sorts of things. And it doesn't surprise me that that hasn't panned out because there's many reasons why it's very difficult in a talent context to be able to do that type of work.

I mean if you want to know what Netflix movie to watch that's easy because it's billions of data points and they can analyze it all and they can recommend to you what movie to watch but in a people context we don't have that level of data. We don't have the types of algorithms that can really accurately pinpoint those sorts of things.

But it was that, I don't know if you agree with me or not, but I think it was that that really was the forcing device to push people analytics. This idea that people thought you could do this. So it's amazing. I thought in many ways people analytics came on the scene out of something that never really materialised.

But it was a good thing that it came on the scene because what excites me about people analytics is the fact that we are now opening up this discipline to bring in expertise that organisations never had before. So psychometrics, network analysis, natural language processing all those things I've mentioned are things that are now being allowed to move into this space because it's opened up and because people are realising there's something we need to do here. So even though the initial foundations may not have been solid and may not have been the thing that really brought it into its own. It has created an environment that's allowed it to develop in a way that has brought a lot of really specific expertise into it.

So what personally excites me about working in this field is the sheer amount of different types of models and methodologies that I'm using, the different types of expertise that I'm bringing into these problems. It's so varied and it's so cross-disciplinary. That's really what excites me about it.

In terms of your second question, which is what worries me the most in people analytics. I think two things one is I think it's a lot of people who say they're practicing people analytics, but they're not really practicing people analytics and the danger is that if you get a lot of poor quality examples of what's happening in this space. It can make decision makers say well, there's not really a lot there and maybe I shouldn't be investing in this after all. So I think it is very important that when we communicate work to the broader people analytics community that we're careful about quality of that work and that it is genuine insight. And we're not just talking around the topic a lot.

I think that's one thing and then I think the second thing is that I've not seen the kind of piece that I've development but I would have ideally like so I think the Bersin report showed that there's a lot of organisations still feel they're fairly basic in their people analytics operations.

So I think we need to get to the bottom of why we're not moving rapidly on that or what's slowing us down because I would like to keep up the pace of excitement and the pace of development that's been going on over the last few years. And I worry that it'll eventually peter out if we don't keep that that pace up so..

David Green: Any concerns on the ethic side of things and what that could do to the field.

Keith McNulty: I think GDPR has been a great development for...

David Green: We agree with that...

Keith McNulty: ...for ethics. Because it's forced us to address the question. There was no forcing device there before and so, some organisations naturally have strong ethics and therefore, when they look at GDPR, they're probably looking and saying well, it's not telling us anything new we were doing that already, but there's probably some organisations out there that thinking whoa, okay, there's some things we have to do differently now, so that's been great as a forcing device. And what I really like about GDPR is the forcing device orientates around fairness. How do we make sure with all this technology that we are actually being fair to employees? Because it's very easy using algorithms to start to drive probabilistic judgments on things that can result in a decision about someone's future career based on a probability. And sometimes that probability is not even very high.

And it's important that we don't go down that route. It's important that our decisions are always made fairly with all of the employees skills in mind, which is why by the way, they'll always be a big role for people in the HR space because people need to make that judgment, machines won't be doing that anytime soon.

So ethics is a very important part of the business, but I think we're much further along because of GDPR and I think that that's great. And if there's HR professionals that don't yet understand GDPR I would have encourage them to get out there and understand it because it's a very good basis to operate your business around.

David Green: I agree I think GDPR is actually is a good thing and it's easy for me to say because I don't have to implement it within an organisation, but I think in long term it really forces companies to act as you said really think about the employee. And actually think if we can't really communicate what the benefits of this are to employees then maybe we shouldn't do it.

I think that's a good thing and that hopefully it helps protect the field.

Keith McNulty: Absolutely.

David Green: So we talked a bit about obviously the wider context bit about what you're doing in McKinsey and the people analytics and measurement role. Somehow you manage to be a prolific writer as well. And obviously a series of different articles there, some that are talking about the work that you're doing others talking about some of the wider context.

What motivates you to do that to keep going and what new experiences have you had from doing it as well? 

Keith McNulty: Yeah, I think I'm part of it is selfish. I realised that in a chaotic space like this where there's a huge amount going on and it doesn't really have a great deal of organisation to it.

It's important to get clarity in your mind about what's important and what's not important and writing is an incredible way of doing that. So the actual process of sitting down and creating a linear narrative around something is an amazing way of clarifying something in your own mind. So a lot of the very clear thoughts and opinions I have on things has come out of the process of writing.

So in many ways I do it for myself. But of course, there are a lot of people out there in the HR community that are really trying to grapple with stuff at the moment. They're seeing all this stuff come their way. And they don't know what they should be focusing on, what's important, what's not important?

So in many ways by my own process of clarifying that in my mind and writing about it. You can share those perspectives with others and hopefully that will help them navigate their way through this so that's been a big motivation for me to do that. And I think both of those reasons are important.

If any one of them weren't there, I probably wouldn't be writing as much as I do. But the writing aspect is very important. I think one other reason I would say is because I do two types of writing. Some of my writing is more general about the future of HR and those sorts of things.

But I do quite a bit of technical writing and the reason I do technical writing is because I would ideally like to model to the HR Community a journey that I've taken which is that if you talked to me three or four years ago, I didn't know how to code I would have no idea how to do this stuff but I put that effort in and I learned because I realised that to work effectively in the world of the future and to be credible in the analytics space you have to really know that stuff. So I actually took that journey and I want to help and encourage other people to do that because I think that's very important. You can't just sit there and talk about people analytics. You just have to go and do stuff to make yourself better at it. And so a lot of my writing both encourages that but also gives people tips and shows them the ways of doing things, shares code with them and you know, I hope more and more people will do that because in the end we're moving towards environment where a lot of our development is in open source. So people can collaborate a lot more on topics. And you can enable faster movement towards that by sharing your own experiences and your own expertise.

David Green: Please keep it up because I know there's lots of people enjoy it, I enjoy it and I think it's good when people are actually doing the work like you, like Dawn Klinghoffer, Patrick Coolen... That actually share, some of their experiences. I think it's inspirational for those of us out there trying to learn so keep that going.

So that leads us on to the final question. This is a question we ask everyone who kindly comes as a guest on the show. Where do you see the future of HR going? So if we look to 2025 and please feel free to go to go beyond that time span if you'd like to.

Keith McNulty: Yeah, and I think I have to go beyond it because 2025 is not very far away.

And one of the things we mentioned earlier is that I don't see us moving at pace. So my honest answer is hopefully there will be a few more organisations that are starting to touch on HR 3.0. There's a few that are in that space now. A handful i would say. Hopefully a little bit more than a handful by 2025 but I don't anticipate rapid progress just on the pace of what I'm seeing and economic conditions will affect that as well.

So, things go up and down with the economy and we can't really predict to what extent it will continue to be an area of growth for companies. So 2025 seems a bit too short for me. As I look out past that I would hope that a few things start to happen. I would hope that we start to see a reshaping of what HR is within the Education and Training sector so that professional organisations and universities and other training institutions start to reshape and rethink what is the role of an individual coming out of a qualification moving into an HR function and what types of roles could they possibly play? And move it beyond the kind of you know service line idea that we've been operating around for some time.

So I hope that that starts to develop and then relating to what I said earlier. I hope that we start to see some greater productivity from HR professionals themselves to skill themselves up. Particularly around analytics, but also around agile and what it means to operate agile within organisations and also a little bit of positive movement from leaders like CHROs in terms of pushing these disciplines more within their functions so that we gradually move away from that...Somewhat away from that swim lane idea and have much more generalist capabilities and competencies within the HR function. But as I said earlier in the interview, I think it is really in the short-term reliant on the professionals and the leaders to push that and if they do that, I think we'll start to see the educational institutions and the professional organisations follow suit. And move that that ball along more.

David Green: And in terms of automation of the HR function, where do you see that progressing over the next five to ten years?

Keith McNulty: So I think there is considerable opportunity for automation to play a role in HR. It really depends  on scale.

One of the things is when we have these discussions we talk about the HR function, but in some organisations the HR function is like 10 people and in some organisations the HR function is thousands of people. So scale is a big factor in automation and in order to effectively automate a process you can do that in small and large scale businesses if the process is easily predictable. So you might be seeing for example that as we move towards the future, simple interactions with employees, which we already see in some organisations. Like can I see my payslip or can I find out more about my benefits or those sorts of things become more automated. But anything beyond that where you're asking for a recommendation to be made or an action to be taken and for the machine to make that decision and recommend that... I don't see that happening in the near future simply because no organisations are really operating at a true scale that would allow that to develop. To create accurate recommendations on anything more than basic and decision-making. So, I think where automation has a role to play is in delivering more rapid, general processes in the organisation. And it's often about employees or professionals in the organisation being able to access stuff for themselves. And that's where we've gone down with our analytics  route, which is moving more towards a self-serve type of analytics model where people get all the key metrics for themselves and the organisation.

David Green: Great. Well Keith, thank you very much for being on the show. It has been great to have you. How can people stay in touch with you via social media and access the stuff that you're publishing?

Keith McNulty: Yeah, the two most common sources that I use. LinkedIn. So I do a lot of writing on LinkedIn mainly around more general HR and people topics. Some articles are in psychometrics there as well. And then for more technical writing, so if you're kind of more of a data science type and you want to get into the nitty-gritty of stuff, I do some writing on medium and that's more, examples of how you can approach a particular problem or ways that you can code things.

So those are the two major sources of information that I'd recommend people if they want to follow some of my writing

David Green: Perfect. Well Keith, as I said, it's been a pleasure to have you on the Digital HR Leader show. I'll see you soon. Thank you..

Keith McNulty: Great to be here. Thanks.