Episode 289: Data Inspired: Turning People Analytics Into Better Business Decisions

When did the data last change your mind?

Organisations have invested heavily in data, dashboards and analytics technology. Yet better information doesn’t automatically lead to better decisions, particularly when it challenges established beliefs or the way things have always been done. In this episode of the Digital HR Leaders podcast, David Green is joined by Sebastian Wernicke, author of Data Inspired, to explore how organisations can use data to challenge assumptions, identify new possibilities and turn evidence into action.

Sebastian has spent almost 20 years helping organisations use data to grow revenue, reduce costs and understand risk. He is a partner at the economics consultancy Oxera, where he leads the data science and AI team. Drawing on his book and consulting experience, Sebastian explains why improving decision-making requires attention to culture, leadership and human behaviour alongside the quality of the data itself.

Join them as they discuss:

• The difference between being data-driven and data-inspired, and why organisations need both
• Why more data rarely changes minds on its own, and how psychological safety enables people to challenge assumptions
• What a consistently green dashboard might conceal, and how targets can distort workforce measures
• How to start with the business challenge, judge whether evidence is sufficient and use targeted experiments to fill gaps
• Sebastian’s data triangle, and why reducing engagement or performance to a single score can undermine understanding
• How better decision-making brings HR and technology teams together, and why relationships and critical thinking matter for future work

‍

This episode is sponsored by 365Talents, a Docebo company.

Every organisation approaches skills in its own way, so 365Talents designs its talent intelligence solutions to adapt to how each business works. HR teams use them to put skills-based strategies into action more quickly and to see the impact across the whole organisation.

Find out more about 365Talents

Building or refining your skills strategy? The 2026 Skills Toolkit brings together 14 free HR templates covering every stage of the journey. It includes a business case built for your CFO, a structured framework for comparing vendors, a strategy you can take to the board, and practical tools for day-to-day HR operations and rollout.

Get free access to the 2026 Skills Toolkit

Resources:
Data Inspired
Oxera

‍

This episode of the Digital HR Leaders Podcast is brought to you by 365Talents, a Docebo company.

[00:11] David Green: When did the data last change a decision in your organisation?  Most of us have more data, more dashboards, and more sophisticated tools than ever.  Yet many HR and people analytics leaders still find themselves asking why that investment isn't having more impact on the decisions that matter.  I first came across today's guest, Sebastian Wernicke, at the Wharton People Analytics Conference in 2019.  He delivered a fascinating and immensely humorous talk analysing what makes a successful TED Talk.  Sebastian has spent almost 20 years helping organisations turn data into value.  He's a Partner at the economics consultancy, Oxera, and author of the recently published book, Data Inspired.  

In our conversation, Sebastian and I explore why more data rarely changes minds on its own, how organisations can become very good at reinforcing the status quo, and why a dashboard that stays green should prompt curiosity rather than end a discussion.  We then bring this into HR and people analytics.  How do we start with the business challenge rather than the data available?  What should we do when the evidence isn't good enough?  And when might a targeted experiment help us answer a question that another report can't?  We also discuss how metrics shape behaviour, the trade-offs involved in simplifying complex workforce data, and why linking engagement scores directly to manager bonuses can undermine the understanding we're trying to gain.  Running through the conversation is the importance of psychological safety, helping people challenge assumptions and change their minds without putting their credibility at risk.  There's plenty here for anyone who wants the numbers in a meeting to lead to better decisions.  So, without further ado, let's get the conversation started.  

Sebastian, welcome to the Digital HR Leaders podcast.  Looking forward to our conversation.  I do remember seeing you on stage at the Wharton People Analytics Conference.  It was probably 2018 or 2019.  You were doing a great talk about TED Talks, actually, and the data behind it.  Before we get into the conversation, can you give listeners a quick sense of who you are and the work you do?

[02:27] Sebastian Wernicke: Yeah, of course.  So, great to be here.  If I say who I am, I would probably describe me as a very, very curious person.  So, I think that also then leads to what I do.  I've been working in data for almost 20 years now; so, in data, with data, trying to help organisations turn that into value; so, whether it's using data to increase revenue, to save costs, to understand risks better.  And currently, I'm a partner at Oxera, which is an economics consultancy, where I lead the data science and AI team.  And as you mentioned, I've also gone among the authors, bringing some of these experiences on physical pages, which is still a thing these days.

[03:09] David Green: Sebastian, can you provide listeners with a kind of helicopter view of what you cover in the book, and what led you to conclude that becoming data-driven is necessary but no longer ambitious enough?  

[03:20] Sebastian Wernicke: Yeah, I mean, one thing you will notice among these names that you just mentioned is that you would classify probably none of them as data people.  And, I think that is also what the book is about.  So, it says 'data' in bold letters on the title, but it's mostly a book about the culture that a company needs in order to be successful with data.  So, I make that distinction between data-driven and data-inspired.  And data-driven is, I think, the way that we all tend to think about data these days.  So, it's about, let's look at the data to optimise what we're doing, to systematically look at the operations, to optimise performance and to get a little better every day at what we do.  And this is, I think, something that is here to stay, and it's, of course, something that you need to do, but it has moved from being a competitive advantage to just being the baseline.  Essentially, everybody's doing it.  

I think there's another thing that you can do with data and another way that you can think about data, which is the data-inspired mindset, which says, "How can we use data to not just do better what we do today, but actually to see how we can transform and reinvent the company?  So, what are our next big steps?  How can we uncover that in the data?  And how can we use data not just to do better what we do today, but actually to find out how can we change, and how can we change on an ongoing basis?"  And that doesn't depend so much on the data itself, even though you need it.  But if you want data to spark curiosity, to challenge the fundamental assumptions that you're making about your business and your company, and to drive that transformative innovation, then you need to think about culture, you need to think about mindset, and you need to think about things like decision-making, how does it happen, and how do we ensure that data can actually play that role.  

[05:21] David Green: In the book, one of the things that struck me is you argue that many data initiatives succeed at making organisations better at staying the same, which I quite like.  What does that look like in practice, and why is incremental optimisation so often mistaken for transformation?  

[05:39] Sebastian Wernicke: Yeah, I mean we live in a world that is moving quite fast and is quite complex, and I don't have the impression that it's becoming any easier.  And so, it's oftentimes, I think, very tempting to turn to data for comfort.  So, we use data for assurance, essentially.  We use it to build dashboards with performance indicators where we try to see, is everything green?  And as long as everything is green, we can be satisfied.  Or we try to add data to decision-making, not really to influence our decision-making, but rather to have something else except ourselves to point at if that decision turns out to be wrong.  And we can say, "Well, we actually had lots of data going around the data set, so it wasn't just me making the decision".  So, there's this great temptation of data just reinforcing the status quo.  So, if it's on the dashboard, you are monitoring what's going on today, but you're not asking, "Am I even looking at the right things?  Are there other things I should be looking at?"  As long as it's green, you're satisfied.  And I think with the decision-making, it's the same, just because you have data around, well, when you have data around, it's very easy to tell yourself, "Well, of course we are making data-driven decisions here because we have all this data surrounding us, more dashboards than Netflix channels, essentially".  

But at the same time, if you're asking yourself, "Am I actually changing my mind based on something I'm seeing in the data?  Is my manager changing their mind based on something that they see in the data?  Is the board using data to change their minds?"  Usually the answer is, "Well, not so much".  And I think that is a stage that we need to come to, because I think you already mentioned the big point, right, what is the purpose of data?  The purpose of data is to act differently and to decide differently.  Because if you make the same actions and the same decisions with or without data, it becomes a very, very expensive hobby.  And then, you're not data-driven, you're data-decorated, in a way.  

[07:46] David Green: Yeah, I like that analogy there.  So, people analytics is a lot of the work that we do, and a lot of the people listening to this episode will be working in people analytics, and certainly in HR.  And I think we're going to learn a lot from the work that you do outside HR as well, because I'd argue that HR is, in most cases, a little bit behind the rest of the organisation when it comes to using data and driving decisions and outcomes with data.  But one of the frustrations we hear from organisations in people analytics is that they have data, lots of data, more dashboards, as you said, lots of dashboards, and more sophisticated tools than ever before to actually create these dashboards, to create and store the data as well.  Yet many people analytics leaders, but also many CHROs, many business leaders remain frustrated by the impact.  Indeed, in our research that we've been doing every year since 2020, consistently around 60% of people analytics teams are what we classify as D teams; they're essentially reporting, and that functions only.  That frustration, where do you think that frustration comes from?  

[09:00] Sebastian Wernicke: Yeah.  I mean, first of all, we should say if you're data frustrated, you are not alone.  I think data frustration really is a broad phenomenon, and I think it's a symptom.  And one of the causes of that data frustration is what I like to refer to as the data deficit theory.  So, I think it's very intuitive to operate on the assumption that all we lack is data.  And as soon as we add more data to the organisation, as we add more data to the decision-making process, or as a vendor might tell you, "Just get the right data to the right people at the right time", as soon as we do that, the rest will just happen on its own; that data is going to find its way into the thinking of people, it's going to overturn convictions that people have; it's going to make us more rational decision makers.  The unfortunate thing is, if we assume that, we have decades of psychological research that we are working against.  So, it's very, very well-known that our brains are fundamentally resistant to letting data guide the way, especially if we believe something strongly, if it challenges a pre-existing belief, or if it challenges our instincts.  And I think that a lot of the frustration is coming from expecting data to do its job on its own, and then it doesn't happen, and then you made all of these investments, but you're not seeing the change.  

So, one crucial thing to do if you want to avoid data frustration is, I think, to not just look at the data, but also look at the, what I would call the decision framework of your organisation and the decision-making culture and say, once data arrives, in that room, it arrives at the door, do you even have a receptive counterpart for that?  In meetings, are you able to discuss data?  Are you able to change strongly held beliefs?  Are you able to walk back something, even if you're the senior leader, once the data agrees with you?  And we need to look at these components as well.  And I think also, that is where HR, in my mind, can play a very, very big role, because those two discussions, the discussion about the data, which is essentially seen as a technical topic, and the decision-making culture, which I think is something that HR can take and drive, they need to happen in the same room, because if they don't go together, you're just not connecting.  You have a pipeline delivering this awesome data, but nobody actually being able to connect to that.  The plug doesn't really fit the socket, in a way.  

[11:50] David Green: And is one of the problems, it's certainly something we see in HR and in people analytics, is one of the problems that companies, well, analytics teams maybe, are too ready to provide the data or start with the data, and not start with the problem that is trying to be solved?  That could be, I don't know, that could be attrition in a key business area, it could be driving revenue, for example, if we look outside.  Do we too often start with the data and not with the problem we're trying to solve?  

[12:16] Sebastian Wernicke: Oh, yeah, absolutely.  I even made that a chapter title where I said, "Purpose first, data second", because it's so important to realise that in the end, data is a tool and we should not mistake it for the solution.  And I think the same thing goes, by the way, for everything that is built on data.  So, of course, nowadays everybody's talking about AI, but AI fundamentally is also just a data tool, even though it's amazing and it's very helpful and it's going to revolutionise a lot of what we do, but it's a tool that takes data, processes data and it outputs data.  And so, I think that it's quite important that when we are adding the data to what we're doing, we realise that the data itself is not going to bring the certainty, it's not going to do the complexity reduction that we might be hoping for, but rather, it's going to present us this very rich picture of reality if we let it, and I think we need to be prepared for that.  If we go in with the expectation that we're just going to get this yes or no answer from the data in the end, then we're bound to be disappointed by it.  

[13:38] David Green: This episode is sponsored by 365Talents, a Docebo company.  365Talents takes a flexible, tailored approach to skills and talent management because no two businesses are the same.  Their adaptive talent intelligence solutions empower HR teams to move faster, build skills-based strategies, and deliver real impact at scale.  Want to learn more?  Visit 365Talents.com.  And don't forget to explore the complete 2026 Skills Toolkit, 14 HR templates spanning the full skills journey, from the CFO-ready business case, through structured vendor evaluation, board-ready strategy, daily HR ops, and rollout.  Head over to the show notes and get your free access today.  

And you talked about HR's role in the decision-making culture and maybe a culture of curiosity as well.  In your kind of experience, Sebastian, and maybe in the book as well, what distinguishes a genuine culture of enquiry from an organisation that simply produces and churns out and circulates a lot of data?  

[15:02] Sebastian Wernicke: Yeah, there's two things I would look at if you're sort of trying to diagnose where you are on that spectrum. One thing I already mentioned, which is look at your meetings where decisions are made, even though that is a simplification, decisions are never just made in one meeting, as we like to think.  But look, you can imagine it and then basically just ask the hypothetical question.  If the junior analyst in that meeting comes in with some data contradicting a strongly held belief, especially if it's a belief by one of the more senior people in the room, would data be able to change the direction of the discussion?  I think that is something that would distinguish a data-inspired culture from what you would see in most organisations today.  The other one is the curiosity and enquiry.  I think it's sort of what is the default reaction once that contradicting data comes in, or once the dashboard turns red, which of course it tends to do from time to time.  Is the reaction to say, "Okay, we need to immediately fix this and get the number back to green?"  Or is it more a reaction of curiosity and enquiry?  So, are you just trying to get the number back into the green, or are you saying, "Okay, let's get curious about this.  Let's really understand what's going on here and uncover the root cause"?  

So, there are techniques like, I think Amazon popularised it, the famous Five Whys technique.  So, you ask why five times, once you see something going wrong, to make sure that you really uncover what is going on under the surface, rather than just applying that band aid to whatever problem you're solving.  So, it's allowing yourself to go deeper into the data, to fall in love with a problem, essentially.  At some point, of course, you have to then dive back up and make a decision to move on.  But I think that's what distinguishes these two cultures.  Are we just having knee-jerk reactions to what we see as anomalies and outliers in the data?  Or do anomalies and outliers make us genuinely curious and put us into an exploratory mindset?  

[17:23] David Green: And of course, one of the ways that we move from insight to decision to outcome to value is the power of the story that we put behind the insight.  Now, obviously, if you've got a counterintuitive finding that the senior leader maybe doesn't expect, goes against their instinct, then that story needs to be even more powerful, doesn't it, and compelling to actually drive that?  I mean, the storytelling element of this, how important is that?  

[17:49] Sebastian Wernicke: Yeah, storytelling is powerful and always goes back to the Spider-Man saying, right, "With great power comes great responsibility".  So, stories are an extremely powerful way to package data and to give it a direction.  But the responsibility, of course, comes in that these stories can also be manipulative.  I mean, you can take in the same data set; usually, the same data set will not just tell one story, there are several stories that you can tell.  And so, I think if you're trying to explore alternative scenarios, for example, stories in the plural can be quite powerful to actually make these come to life.  At the same time, if you're looking at data and saying, "Okay, here's the one story, this is where we're going", you are, of course, running a danger of losing the nuance that is quite critical if you actually want to be curious about it.  Because, in a way, a story oftentimes I think also tells us already what the solution is and what is really going on.  So, let's say, as a selling device and as a communication device, I think it's quite powerful.  But I would also be careful about introducing it too early.  And I think that also requires a bit of a mind change to say, it's not just that we need to make the data attractive and easy to process, I think we also need to learn how to be comfortable with the discomfort that data may cause us when it's all very complex there is not yet a clear story.  And there may never be a clear story.  There may be alternative explanations.  

If we're really willing to listen to data, I think we also have to listen to these nuances rather than saying, "Okay, so, what's the one story, what's the one bottom line here?"  We do that in the end, of course to again, we need to make a decision at some point, we need some action at some point.  But we do it too early, I think we can also set ourselves on a path and not look left and right.  And that would be a bit of a waste of the beautiful things that data might be able to tell us.  

[19:55] David Green: And I guess that that stage, that early stage, you probably want to be formulating hypotheses as to why something could be happening or not happening, and then doing further enquiry and further testing, I guess, to try and find out, "Is that true; is it not true?"

[20:09] Sebastian Wernicke: Yeah, hypotheses and also just understanding what is important to you.  Again, we come back to the purpose.  So, one example I have in the book is where I show you three TV shows with the user ratings that these shows get.  And there's a very simple question posed, which is, based on these user ratings, you are to decide which of the three shows is rated as the best one.  And it turns out that you can make a perfectly rational argument for each of the three shows purely based on the data.  And so, one would be where you say, "Well, this has the best average rating.  The other show has the most five-star ratings.  Another show has the most ratings and it's rated very high, whereas the one that has the best average doesn't have as many ratings.  So, it's a bit uncertain what the result is".  So, the important thing before you look at the data and ask yourself, "What is the best show?" I think, is to have a discussion in this case to say, "Well, what does best actually mean for us?"  Is best when we're thinking about our viewers, is best that we say, "Well, we want to produce shows with a high average rating?"  Is it that we want to delight our users, even at the risk of maybe not delighting or maybe even offending others?  And you come to different conclusions based on that.  

So, this is also, again, I think it shows you that data alone cannot lead the way.  So, it always needs to be accompanied by a search for purpose and saying, what am I trying to find here?  And then, you're absolutely right of course, then I can start formulating hypotheses and saying, "Okay, so given that this is our purpose, our data might be trying to tell us A, B, or C, and then you start exploring that.  But at the end, I think this is also one of the annoying facts of data.  Even then, you will rarely arrive at a clear yes or no answer.  If we're in data, we are very, very quickly in the realms of statistics.  And once we're in the realms of statistics, there are no certain 100% answers.  So, you might also get an answer where you say, "That new product we're trying to introduce, there's a 63% chance of success", which is highly annoying if you're a decision maker, right?  What is 63%?  Is it a yes?  Is it a no?  Do I round up?  Do I round down?  Do I have the security now?  But in many cases, that's as far as data can get you.  And then as a leader, you have to say, "Okay, what am I going to do with these 63%?  Am I going to go for it or not?  But that's on you, that's not on the data anymore.  

[22:48] David Green: As you said, you still need to make the decision.  The data can help inform the decision, but probably, you still need to have that judgment as well.  

[22:57] Sebastian Wernicke: Yeah, otherwise it's just a calculation, it's not a decision anymore.  

[23:01] David Green: So, you mentioned green data earlier.  I enjoyed your comments on that and in the book as well.  So, you're basically saying that green data doesn't necessarily mean that we're fine, we can relax.  So, what questions should leaders be asking when a dashboard appears to be consistently positive, other than what you said earlier about, are we actually measuring the right things or reporting on the right things?  

[23:26] Sebastian Wernicke: Yeah.  Well, we should probably clarify by green data, we mean the green dashboard, not environmentally-friendly data.  

[23:32] David Green: Yes, yes, of course!  

[23:34] Sebastian Wernicke: Organic data, you could come up with!  I think if the dashboard is green for too long, the main question to ask is, "Are we looking at the right things here?"  And also, maybe a related question, "What might be happening in the business that is going wrong and we're not seeing it?"  These are two facets.  One is probably a bit more pessimistic than the other one.  So, there's a law called Goodhart's Law, which says that, "Whenever a measure becomes a target, it ceases to be a good measure".  And what that means is if we're looking at numbers to steer the business, which in the case of green data basically says we're looking at numbers and saying, "As long as it's green, we're doing fine", then at some point, it's not a very accurate measurement anymore of how the business is actually doing.  And we have some very famous cases of where this has gone horribly wrong.  

So, coming from Germany, of course, I have to talk about cars a little bit.  Everybody has heard about the Volkswagen Dieselgate scandal, where a team of engineers was given this task: develop an engine that has very environmentally-friendly measurement values.  It wasn't really develop an environmentally-friendly engine, even though that was the outcome that everybody was hoping for, but it's essentially, once it's on the test stand, here are the numbers this engine needs to deliver.  Well, the engineers, of course, managed to deliver an engine that delivered all of these values.  But as soon as it was off the test stand, it didn't turn out to be an environmentally-friendly engine at all.  And not every green dashboard will turn into a Dieselgate scandal, but I think it should always make us very careful of saying, you know, I mean, the base assumption should be, "Things are going wrong somewhere anyways".  So, if everything's green, that is a little bit suspicious and not necessarily a sign of everything going great all the time.  

There's of course also a long-term effect, because if everything's green, again we are not really receiving any impulse for change or for transformation or for thinking about what's the next step.  It can become very comforting and bring us into exactly that state where the data we are looking at is making us very good at staying the same, but we're missing the next big development and the next evolutionary step that we should probably take as a business.

[26:07] David Green: And you can imagine, can't you, the leadership meeting and the dashboard's green, "Great, we don't need to talk about that.  Let's talk about something else", rather than that enquiry, curiosity, as you said, "What could be going wrong that we're not even aware of?"  Because there probably is something that's not going right.  I mean, it's very rare that an organisation is going to be 100%, everything going well.  There's always going to be a few challenges, or maybe risks that could become worse if you're not actually on top of it.  So, you definitely need that curiosity, that enquiry, as you said, otherwise you're potentially in trouble.  

[26:49] Sebastian Wernicke: Yeah.  I mean, if you say, "Everything's running 100% great here", that is a very suspicious statement, I would say.  I mean, how can that be?  

[26:57] David Green: Yeah.  It sounds like a vote in a non-democratic country, where 99% of people vote for the leader by sheer coincidence.  We're going to spend a couple of questions now sort of looking through an HR lens, really.  So, again, you've probably experienced this yourself, Sebastian, with some of the companies that you work with.  HR functions often focus heavily on data quality, maybe too much sometimes, and governance.  Not saying that they shouldn't, but they focus on data quality and governance.  And sometimes that can handicap them a little bit from doing the work.  How can HR avoid providing or producing reliable answers to the wrong questions, which again, I think that probably talks to the dashboard piece of it there as well?  

[27:43] Sebastian Wernicke: Yeah.  And I mean, I fully agree, by the way.  I mean, data quality and governance are super-important, especially of course for the HR data.  I mean, it's highly sensitive data, it's data that drives some critical processes, such as payroll.  

[27:59] David Green: Can't get that wrong!  

[28:00] Sebastian Wernicke: Yeah, exactly, you do want some things to be right here.  But if you want to look beyond that, I think it goes back to that purpose discussion.  And so, then the question, of course, becomes, "Okay, so how do we have that purpose discussion in a structured manner?" Because if we say, "You need a strategy", I think that is something that sounds very nice, but strategy is probably one of the most misused and misunderstood terms, right?  It can become very, very abstract.  And many strategy sessions that I've been in, they're essentially just a brainstorming where people will put out the ideas that they've always put out.  So, what I would recommend is a very structured way of looking at that purpose, and it starts out by identifying your challenges going forward.  So, you take a very clear and honest look about saying, "What's going to be difficult for us as a company, and as a consequence for us, as the HR question over the next three to four years?" you don't look farther ahead.  I mean, who knows what's going to happen in five, right?  And you identify these issues and they may be different.  I think some of the discussions, of course, today are quite common of saying, "We need to think about what AI is doing.  We need to think about churn.  We need to think about do people still have the right skills?  How do we know if they still have the right skills?"  But it will not be a generic list.  I think if you're doing that discussion right, you find out very specifically what are the hard problems for you as you look ahead.  

Once you've identified that, you should develop what, in strategy research, and I'm relying a lot on Richard Rumelt here, whose book, Good Strategy/Bad Strategy, I can highly recommend, you develop what's known as a guiding policy.  And the guiding policy says, "Okay, we've now identified all of these big challenges for us.  How do we want to deal with that?  What is our general approach here?"  So, again, you need to be making decisions here, because there might be many options on the table of how to deal with it.  But one of these, or a part of that guiding policy, can then be to say, "And we want to use data to do that.  We want to use data to better understand what's going on with respect to these difficult challenges that we've identified.  We want to use data maybe to also alleviate some of these, to tackle the challenges".  And what that does is, again, it puts purpose first and data second, and so you know what you want to be doing with data.  Does that mean you will automatically have all the data?  Usually, it doesn't.  So, usually what will happen in these discussions is that someone then says, "But we don't have that data.  Why don't we just do something with the data that we already have?" which is the best way to run into the trap of, again, staying the same, because you're relying on the data you already have.  

So, you then need to move, once you've identified, "Here's what we want to do with the data, this is the data that we need", then you can go into action mode and basically say, "Okay, so how do we get that data?  And if we can't get that data, are there any proxies?  Is there any way around that?"  So, I mean, usually it's not a situation where you will directly identify, "Okay, we do two experiments, we buy three surveys, and then we know what's going on".  This is where complexity comes in, and you have to say, "We probably need to triage this, right?  We cannot directly measure it, but here are things in the vicinity that we can observe".  But you now have a very clear path ahead that will tell you, why are we doing this?  And what does a well tackled challenge, what does success, what does that look like?  

[31:49] David Green: I want to take a short break from this episode to introduce the Insight222 People Analytics Program, designed for senior leaders to connect, grow, and lead in the evolving world of people analytics.  The programme brings together top HR professionals with extensive experience from global companies, offering a unique platform to expand your influence, gain invaluable industry insight and tackle real-world business challenges.  As a member, you'll gain access to over 40 in-person and virtual events a year, advisory sessions with seasoned practitioners, as well as insights, ideas and learning to stay up-to-date with best practices and new thinking.  Every connection made brings new possibilities to elevate your impact and drive meaningful change.  To learn more, head over to insight222.com/program and join our group of global leaders.  

And I think we see that occasionally, HR can be a little bit too hamstrung.  Obviously, data quality is important, but it's unlikely, and you might tell me I'm wrong, but it's unlikely you're ever going to have 100% on your data quality.  Sometimes, it needs to be good enough, and other business functions, like finance, would operate on that as well.  And they'll make some assumptions, as you said, or they'll find some data that helps them with the proxy, or they'll collect some new data, or they'll bring some data in from external data to support maybe some academic data that says, "This is what several academic studies have found", those sorts of things.  What would be your guidance to the senior HR leader that they're saying, "Look, our data isn't good enough"?  

[33:47] Sebastian Wernicke: Yeah.  Well, again, first of all, why isn't the data good enough?  So, my first question would always be, "What are you trying to achieve?"  And if we find out that the answer there is a bit fuzzy, I would always encourage to go a bit deeper and to have that discussion first, even though it can feel like a bit of a slowdown or a bit of a distraction of saying, "No, let's just do the data, let's stop having that other discussion".  And I think you need to have thought that first part through.  Then, if you say, "I'm not satisfied with my data", you know why you're not satisfied.  So, is it the data quality, right?  Do you actually know what you want to do, but the data just isn't clean enough, helpful enough to do these analyses; the results are wrong, things don't fit together?  And in complex organisations, that can very well be the case.  I mean, even simple things like saying, who's an employee here?  The larger the organisation gets, that can actually be a complex question from a data perspective, right?  You will also have things like virtual employees, for example, because there's some user in a system that is required and all of these little details.  So, but okay, then you know you have a data quality issue and you know where the data quality needs to be in order to answer the question and tackle the challenge that you've identified.  

If it's not that and you, for example, just say, "Well, we have a lack of data", then you need to think creatively about where am I getting the data from?  And you already mentioned it.  So, there's internal data gathering that I can do, I can get it from surveys, I can get it from academic research.  The important thing here is I think to allow a very rich picture of data.  So, don't look too narrow because you will rarely find the data that answers exactly the question that you're trying to answer.  Rather, you should go broad and basically put together all these little puzzle pieces and facets of what you're trying to find out and look for the answer in the cumulative data set that you're gathering there.  Now, there's another source of data that many companies feel uncomfortable with, even though they shouldn't, and that is to do targeted experiments.  So, to basically say, "Well, here's a piece of data that I don't have, so we probably need to try something out.  Maybe we need to try out a new working model.  Maybe we need to try out a new home-office policy, something like that, in a representative department for a while.  We need to decide what we're going to measure about that.  And that is going to be our data-gathering exercise".

My feeling is that too many companies still feel uncomfortable about experiments.  They seem like they take a long time, they seem like a bit of a waste of time, they're expensive.  But sometimes, an experiment really is the best way to gather high-quality data that is very close to what you're actually trying to answer.  And so, if it's a question that is relevant enough, where the purpose is clearly defined, you should dare to do experiments.  

[36:54] David Green: How can HR leaders, or any leaders for that matter, design workforce measures without encouraging gaming, excessive control, or a narrow view of performance?  So, one of the examples I'm thinking about here is engagement, employee engagement scores.  So, leaders can obsess with getting the score from 72 to 75.  But I would argue that there's much more to it than that, and that doesn't necessarily mean it's good if it goes up from 72 to 75, there's probably lots of other things underneath.  So, how can HR leaders design workforce measures that actually lead to the right conversations, the right decisions, and the right outcomes at the end of the day?  

[37:35] Sebastian Wernicke: Yeah, I've been in these situations myself with it.  So, that just resonated very much, where all the leaders are having these engagement discussions with their teams slightly before the survey is filled out, and these sort of things.  We've all seen it, right?  So, in project management, I'm going to start with project management and then bring it to HR.  In project management, there's something known as a triangle.  And that is when we want to have a project done, we would like it to be done fast and cheap and at high quality.  And the reality that everybody who's ever managed a project knows is you will only get two out of the three at most.  So, if you want it done fast and if you want it done cheap, it's probably not going to be very good.  And you can go through all of the other combinations as well.  

With data, I believe there's a similar triangle, which is that we want three things from data.  We want it to be accurate, so really tell us what's going on; we want it to be universal, so be applicable across the organisation, especially when we're talking about these HR measurements; and we would like it to be simple, so have that one number that's telling us, is everything going okay or not?  And I believe the data triangle is just like the project triangle.  You can only ever get two out of the three.  And you have to decide which one is the one I'm going to deal with in another way and I'm going to accept.  Now, with data, of course, we want it to be accurate.  And I think when it comes to HR, we want it to be universal.  We want to really know what's going across the entire organisation and not just with two people or one team or department, right, we want to know it across the organisation.  So, there's a very straightforward consequence from that, which says it can't be simple.  

So, if you're thinking, "We're going to measure these two numbers", and whether that's engagement or some overall performance or however you define it, we're going to measure these two numbers or let it be three numbers and really know what's going on, that's an illusion, you're not.  And that is part of the whole gaming the system, because two numbers, three numbers, they're always going to compress reality so much that you're losing all of this nuance.  And then, people will just be focused on reaching these numbers.  The second part that you need to pay attention to is, and this brings us back to Goodhart's Law, is to separate the numbers that you use to measure the organisation, so to understand what's going on, and the numbers that you use for incentives and bonuses.  So, if you want to really know what's going on, how healthy is our culture, how good is our leadership, and you directly tie that to the bonus of the leaders, then of course you invite gaming.  I mean, that's just human nature.  And it doesn't, by the way, I mean, it doesn't even have to be malicious.  It's just something, you're just setting this very, very strong incentive to make the numbers look good.  

So, in order to avoid this, there's two things you need to do.  One is you advocate for a really rich and multifaceted data foundation.  So, you're not measuring one thing or two things or three things, but you're really saying, "What could be 20, 30, 40, 50 KPIs that we can measure and that will shed light on all these different aspects on how HR is doing and how our organisation is feeling?  What's the health of the organisation?  This is one way to prevent it, because when then one of these metrics is gamed, it usually affects other metrics as well, right?  And so, you really see what's going on.  And the other is to simply also separate these two things.  So, I think that research is telling us we should not be attaching manager bonus directly to engagement, for example, but rather say, we are measuring engagement, and once we measure it and we find out things in the data, then we'll have a conversation.  And it's something that then the manager of that manager should actually engage with and have conversations about, and rather than saying, "Okay, your number is 7.5, I'm going to multiply that by X and out pops your bonus and hope for the best".  

I learned a very important saying very early on in my career, where somebody told me, "You really need to be careful about whether you're managing the numbers or you're managing the business".  And the two things are just not the same.  But too often, we mistake them for it.  And I think this is exactly the same situation here.  

[42:23] David Green: That's very good.  Yeah, I mean, it's just too simplistic to say that someone gets a pay rise if their engagement score goes up.  

[42:29] Sebastian Wernicke: And yet, how often do we see it?  

[42:31] David Green: Yeah.  I think we've seen it a lot recently in the last few years with, I don't know, female representation at senior levels.  And people on the board get bonuses based on the fact that if they've got the figure up to 30%, 40%, whatever they decide it should be, and you just wonder whether that's the right way of measuring it.  But anyway, we could have a whole podcast on that.  So, Sebastian, a lot of workforce outcomes like trust, belonging, leadership quality, meaningful work, they're quite difficult to quantify.  How should organisations combine measurement with respect for what can't be fully measured?  And I guess that sort of follows on nicely from that conversation about engagement, perhaps.  

[43:17] Sebastian Wernicke: Well, I mean, my very short answer to that is, yes, that's just how it is, you need to deal with it.  It's interesting.  So, one of the people that I've gotten to know a little bit better through the research for the book is W Edwards Deming.  Deming is a mathematician by training, or statistician, and he's one of the inventors of what's known as statistical process control.  So, that already tells you what he's all about, right?  So, he's one of the first people that basically said, "When we look at processes such as manufacturing in an organisation, we should really measure that and scientifically manage that towards quality, output, and all these things".  Now, Deming is often cited with a saying that says, "If you can't measure it, you can't manage it".  And there's variations of that that I'm sure everybody here has heard, you know, "What gets managed gets done", and all these things.  The interesting thing is, this is one of the best examples I've ever found of a blatant misquote.  So, Deming actually said that.  Deming said, "What gets measured gets managed".  But the full quote from him is, "It's wrong to assume that what gets measured gets managed".  

He came to the insight in his career that this is one of the seven deadly sins of management, to try and measure everything and manage everything by numbers.  And it's just such a convenient quote that he constantly gets misquoted.  He must be turning around in his grave every time that appears on a PowerPoint slide.  And we should take a page out of his book.  I mean, you can't accuse the man of not liking numbers.  He is the inventor of statistical process control.  He is a numbers guy.  And yet, he fully realises and comes to the conclusion, "Don't you ever think that management can be replaced by numbers, and don't you ever think that you can measure everything that matters".  And that is okay, because that is then the role of the manager.  That is the role of management to deal with these things.  I mean, if it were just numbers, then why need managers in the first place?  You could just have a very mechanical ranking system that then distributes the bonuses based on whatever it finds in the data.  That's not what we're looking for in leaders, right?  

So, I just think whenever somebody says, "This all needs to be measurable", or also, this is one of the pet peeves of mine, where people say, "Well, all goals that we define need to be SMART, because they need to be measurable".  No, there are some very sensible goals you can define that are not measurable.  And you just have to trust the people, the leaders, the managers to be able to look at a situation and say, "Have we achieved that goal, yes or no?"  That can also be very legitimate.  Now, of course, when you can measure it, go for it.  But again, don't think you can measure everything.  

[46:14] David Green: That's funny.  That misquote from Deming reminds me of the famous one, I wish I could remember the full Oscar Wilde, "Imitation is the best form of flattery".  It's actually, when you see the whole quote, clearly, he doesn't necessarily feel that way.  And I'm going to go and look that up at the end of this recording just to remind myself of what it is.  So, Sebastian, in the book, you describe psychological safety, which we've featured quite a lot in in numerous episodes of this podcast, particularly two that we've done with Amy Edmondson over the years, as a foundation for evidence-based decision-making.  What must senior leaders do when the evidence challenges their strategy, back to that counterintuitive finding, their assumptions or their own previous decisions?  And what is the role of psychological safety, I guess, in actually helping create that healthy culture?  

[47:05] Sebastian Wernicke: Yeah.  So, numbers can, of course, be very, very powerful.  So, let's say you start a meeting and someone says, "Your sales have gone down by 5%.  What do you have to say about that?"  Oh, my God.  You are in a very, very critical situation already, right?  Because, well, first of all, if you haven't seen those 5% before, it's almost impossible to basically have a sensible discussion about this.  It requires a lot of depth and understanding that number, where does it come from, you need to ask these five why questions, you need to look at all the metrics.  So, if you have a culture of confrontation, the numbers are just going to be used as an attack-and-defence mechanism.  And we know this from politics, right?  That's one of the reasons why politicians like to cite statistics or even invent statistics when they have to, right, because they say, "Oh, well, 2% of people believe this".  What are you going to do against that?  And so, the psychological safety is very important to create that environment where you can actually have that sensible discussion.  So, you get out of that attacking mode and also looking at a critical number is not an attack, and basically saying, "Well, I would like to point out the dashboard is green, but we might still be doing things wrong".

So, you need to have that safe environment where numbers can be discussed and numbers can be discussed in relation to the problem you're trying to solve, and it's not about the people, and people do not feel attacked by the numbers discussion.  As soon as they do, the numbers take over and again, you're steering by numbers essentially, right?  "It's minus five, so in the next meeting, it better be not minus five".  Where is that going to get you in the end?  

[48:55] David Green: A couple more.  So, people decisions require both evidence and human judgment, probably not just people decisions, actually, most decisions, but people decisions particularly, because it's obviously about people.  What would be your guidance to organisations on how to find the right balance in areas such as hiring, promotion, performance management, workforce planning, pay increases, those sorts of things?  

[49:18] Sebastian Wernicke: Yeah.  So, I think a couple of principles that I like to follow here.  I think the first one is, and it sounds very basic, but it's an important one, to have that discussion in the first place, to raise the question that you just raised and basically say, "Okay, obviously we want to use evidence.  Obviously, evidence is not going to tell us everything.  Where do we see ourselves here?  Do we want to be an organisation that mostly relies on the intuition of experts?  Do we want to be an organisation that says we will try to measure everything as much as we can?"  And if the answer, which would be the sensible answer, is somewhere in between, then you need to talk about skills and you need to talk about processes.  So, one of the companies that illustrates that quite well is Netflix, for example, where when they select a new show to produce, on the one hand, of course, they have a very rich data set to rely on, but that data set is not in itself going to tell them what's the next hit show.  And so, they need to think very carefully about, "How do we equip our experts with understanding of data, but also with understanding of decision-making?  And what's our expectation?"  Are they expected to overrule the data if they feel strongly about something, or are they expected to follow the data?  And the answer may be a bit different for every organisation, but it's important first to have that conversation and to be very explicit and deliberate about it.  

Once you have that, I think what's going to fall out of that is certain requirements in terms of skills, for example, where you say, "Okay, if we want people to act sensibly on the data, we need to teach them the standard data literacy, of course, but we also need to teach them about the psychological effects here, how do you make good decisions?  And what are biases that commonly occur once you are confronted with data?"  And decision-makers then, with that, can I think act in the way that you want them to.  Another thing is there is always a strong impulse towards more numbers and more measurement.  So, I mean, a lot of these HR decisions, of course, are difficult.  And so, whenever things get difficult, we're just so tempted to say, "Okay, next time we're just going to measure that".  So, for example, "We have nine skills that we define.  I'm going to rank you on each of these nine skills.  I'm going to add up the numbers.  And again, those numbers are going to determine your bonus".  And I've always been fighting against this.  I understand where it's coming from, but I always say this just cannot capture the complexity that we actually need to capture.

I mean, just take this case.  Maybe you rank very high on one skill and sort of in the average on the other skill.  I need to think about what that means.  Does that mean actually, I'm going to value your one strength here, and that may even be able to compensate for your average skill in the in the other section?  Or is that average actually a deficit where I want you to grow?  But that is not a decision that data is going to make for me.  That is something that I, as a leader, and looking at my team and what I want to achieve with a team, but also what I want the team culture to be, that's what I have to decide, and the numbers are not going to tell me that.  

[52:34] David Green: So, Sebastian, two more questions before we wrap.  Technology, data, culture, they're often managed by different teams in organisations, sometimes in different functions, invariably in fact in different functions.  Who do you believe should own the shift towards a data-inspired organisation?  And what operating model, if there is a single operating model, because there probably isn't, what operating model enables HR, business leaders, technology teams, and analysts to work together effectively, or what type of operating model, is probably a better question, allows those people to work effectively?  

[53:08] Sebastian Wernicke: I think AI is probably presenting us a good window of opportunity here, because with AI coming in as a technology, I think it should be clear to everybody that what we're looking at is not the introduction of a technology, but we are actually looking at a transformation here.  And the nice thing about a transformation is, and especially about if the transformation is driven by a technology that is all over the media, is that it's important enough to get on the CEO's agenda.  And I firmly believe that this needs to have space on the CEO agenda, whether it is driven by AI or also it is that drive towards the data-inspired organisation, and I don't say that lightly.  I know that CEOs are basically being told every 30 minutes what should be on their agenda and the topic that they should pay attention to.  But fundamentally, I think you raised the crucial point; this is cross functional.  

So, there are some organisations that are actually now saying, "Okay, we are going to somehow merge HR and IT because we're going to have people and AI agents".  That is very forward-thinking, and let's see how that goes.  But in most organisations, of course, these functions are here to stay, but you need to bring them closer together, and you need somebody to drive that collaboration and also to be an arbiter when there are conflicts, which invariably will arise.  And in my mind, that is the CEO, and the CEO essentially needs to define that agenda and then make sure that technology and HR go hand-in-hand, that they talk to each other, and that they jointly achieve these goals.  

Now, if you are in one of these functions, I mean, whenever I talk to Chief Data Officers, CIOs, CTOs, I basically suggest to them to expand their definition of what they're responsible for towards decision-making to say, "Well, you are right now in the technology field, but ultimately what you're trying to do is you're trying to enable better decision-making for the organisation.  So, define yourself as a function that is responsible for better decision-making and uses technology as a means to that end" and just to expand that sphere a little bit.  And of course, this can also go the other way around where we say, "Well, as HR, you're not just here to create a few dashboards about how people are doing, you are trying to make the organisation more effective.  And what better way to make an organisation more effective than to improve decision-making".  And so, decision-making is, I think, where we can find common ground and have these two sides of the same coin essentially being looked at.  And so, driven by the CEO, taking decision-making as the fertile ground on which to grow these connections.

[56:05] David Green: Yeah, I like that that link then between technology and culture drive better decisions, drive better outcomes, drive more value ultimately.  If that's not on the CEO agenda, then I don't know what is.  

[56:19] Sebastian Wernicke: So much, but yes!

[56:22] David Green: And Sebastian, do you think, I mean obviously, we're in the early stages of the AI transformation at the moment, and obviously there's lots of hype out there.  But one thing that we're seeing is there's more energy now towards the work that needs to go in to get your data foundations correctly, get your data governance in order as well, which aren't the most sexy topics, let's be honest about it.  We might get excited about them because we work in that space, but they're not the most sexy topic.  Are you finding that AI is providing new energy to that and actually the sensible companies are making the right investments in that area, because ultimately that will help them create those better decisions, that better value from their AI investments?  

[57:06] Sebastian Wernicke: That is a very, very good question.  I mean, I certainly see a lot of hope that this will be the case, right?  So, everybody's talking about, "Oh, now that AI comes in, you finally need to pay attention to the data".  So, I would hope that this is the case.  The reason why I'm a little bit cautious is that over history, this has not really played out that way.  I mean, this is not the first time that we have very powerful technology coming in that relies on data.  And we had the age of big data, if anybody remembers that.  

[57:41] David Green: Oh yeah, I do remember that.  

[57:43] Sebastian Wernicke: Everybody got excited about analytics.  And before AI, by the way, there was machine learning.  And machine learning also needs a clean, very well-governed data.  And somehow, always everybody pointed to the data and said, "If you really want to use this technology, you'd better get your data right and sort it out".  And yet, once the other priorities came in, once the operational daily business took over, I always like to say data has always been the second most important topic.  So, it is always on the list of priorities, but it's always on place number two.  And everybody just looks at place number one and tries to take care of that.  So, we will see.  But I mean, to be clear, the organisations that will be successful and the ones that are going to run probably circles around the competition are the ones that have good data foundations.  Because with AI, you're essentially handing a lot of the operations just to data itself, right?  So, you have data for AI to make its decisions on.  Once AI makes the decisions, those go back into data.  If you don't have good data in that system, I don't know how that's really going to work.  So, we are going to certainly see companies that get this right, and maybe then other companies that have not paid enough attention looking a bit jealously at what's going on there.  

[59:09] David Green: Well, thank you for providing a healthy dose of realism to my perhaps slightly over-optimistic view of it.  

[59:15] Sebastian Wernicke: I want to share your hope.  It's not that, right?  

[59:17] David Green: No, no, I get that.  So, this is the last question, Sebastian, and this is a question we're asking everyone in this series of the Digital HR Leaders podcast, and I'm actually really interested in your answer to this, obviously coming from someone from outside HR.  How should HR prepare people for jobs not yet defined?  

[59:37] Sebastian Wernicke: Yeah, this is a fantastic question.  I'm going to make this a bit about AI because that's what I'm thinking about right now.  I think there's a couple of fundamentals that are not going to go away.  And the first one is the ability to build relationships and connections.  I mean, business is going to be done by people.  And if you have the ability to understand what's going on with your stakeholders, if you have the ability to bring together stakeholders around topics that they might disagree with, if you have the ability to translate what they want into something that you can help them with.  I mean, there's something I'm confronted daily, right?  So, I need to translate business problems into data problems.  If you have these human skills, I think that is going to be something that is always going to be valuable.  So, prepare people and help them grow in these skills.  

The other one is, I think, all about judgment and critical thinking.  This is a bit of a fuzzy one for me because I'm not really convinced of where we say, "Oh, AI, we certainly need human judgment".  I think we have a couple of counterexamples to that by now where AI is actually showing better judgment and making better decisions than humans, and we just have to accept that.  But still, the ability to look at something and say, "Is this right?" from various points, you know, "Is this right from a logical perspective?  Is it right when I take a step back?  Is it right from an ethical perspective?" I also think that is here to stay.  And so, as long as we help people grow in these skills, and they have the connection-building ability, they have the critical-thinking ability, and then if ideally we provide a bit of safety and trust of saying nobody really knows what's coming, but we are going to go through this together, and we're going to make it to the other side, then I think you're already very well preparing everybody in your company.

[1:01:42] David Green: Very good.  Well, it's been a fascinating discussion, Sebastian, I really enjoyed it.  I've learned a lot.  Where's the best place for people to find out more about you, about the work you're doing at Oxera, and also to find out more about Data Inspired and follow what's next on the agenda for you?  

[1:02:01] Sebastian Wernicke: Yeah, so I'm quite active on LinkedIn, I regularly post about data and data-inspired issues.  So, follow me there.  Also, feel free to connect if you want to have a dialogue about that.  Always happy to meet interesting people.  If you want to find out more about the book itself, well, you can buy it wherever books are sold or go to datainspired.org.  That's the website.  There's the newsletter on there, a bit of more information if you want to know about it, and yeah, also, ways to reach me.  

[1:02:30] David Green: Perfect.  Well, what we'll do, Sebastian, in the show notes, we'll put the link to the datainspired.org and your LinkedIn and the Oxera site as well.  So, wonderful.  What a great discussion.  I can go away now.  My brain's full of stuff to think about afterwards, and I'm sure the same will apply for many of our listeners as well.  Sebastian, thanks for being a guest on the show.  

[1:02:50] Sebastian Wernicke: Thank you so much.  It was a great pleasure.  

[1:02:54] David Green: Sebastian, thank you so much.  It's been a real pleasure to have you on the show and explore how organisations can get more value from their data.  One point that particularly stayed with me is the distinction between managing the numbers and managing the business.  A measure can improve, while the problem it was intended to help us understand remains.  That's a useful reminder to keep asking, what sits behind the number, and whether we're measuring what matters.  I also like Sebastian's challenge to ask when the evidence last changed our minds.  For people analytics leaders, that's a revealing question about the impact of our work and the environment in which it's received.  So, turning to our listeners, when did people data last change an important decision in your organisation, and what made that possible?  Come and find my post on this episode on LinkedIn and let me know in the comments.

To learn more about Data Inspired, visit datainspired.org, and where you'll also find Sebastian's newsletter.  You can follow him on LinkedIn and we'll include those links, alongside Oxera, in the show notes.  And if you enjoyed the episode, please share it with a colleague who would find it useful, and subscribe so you don't miss what's coming next.  One last thing before we go, if you'd like to keep up with what we're working on at Insight222, follow us on LinkedIn or head to insight222.com.  You can also sign up for our bi-weekly newsletter for the latest thinking on HR, people analytics, and workforce transformation.  Right, that's us for today.  Thanks for listening and we'll be back next week with another episode of the Digital HR Leaders podcast.  Until then, take care and stay well.