Episode 149: How Meta Developed a High-Performing People Analytics Team (Interview with Alexis Fink)
In this episode of the Digital HR Leaders podcast, David is joined by Alexis Fink, Vice President of People Analytics and Workforce Strategy at Meta, to discuss the critical role of people analytics in shaping the future of work.
A thought leader in the industry, Alexis brings over two decades of experience in the field of HR analytics and insights to the conversation. Throughout the episode, she and David explore:
How the people analytics function has evolved over the years
The different types of people analytics team structures and how they can be built for optimal results
Alexis’s three-dimensional approach to building people analytics teams
The importance of a strong HR and people analytics relationship
The skills, competencies, and behaviours required for HRBP's and people analytics professionals to be effective as the world of work continues to evolve
The importance of IO psychology in the field of people analytics
Alexis’s valuable insights and practical tips for scaling up people analytics teams
How people analytics will continue to evolve and shape the future of work
Overall, this episode offers a wealth of insights and practical tips on people analytics from a true thought leader in the field. Click the link below to listen to the full episode. Enjoy.
Support from this podcast comes from Orgvue. You can learn more by visiting: https://www.orgvue.com/
David Green: Today's episode is one that I know listeners will enjoy. Joining me today is Alexis Fink, Vice President of People Analytics and Workforce Strategy at Meta and the new incoming President Elect of SIOP. I'm particularly honoured to have Alexis on the show as she has built an incredible people analytics team at Meta and is a renowned expert in the field.
Alexis will provide insights into how to structure a people analytics team, the skills needed to thrive in a people analytics role, and guidance to aspiring people analytics leaders to take their teams to the next level. So, if you're looking to understand how to structure your people analytics function for scale and to deliver value, this conversation is for you. Let's get started.
Alexis, welcome to the show. Before we dive into the conversation, could you please share with our listeners a little bit about yourself and your role at Meta?
Alexis Fink: Sure. So, this analytics work has really been my life's work. I was reflecting recently that my first job in this space was in 1992. I do have the privilege of having had similar roles at both Intel and Microsoft, and have had roles doing this kind of work going back, actually back into the 1990s, I had the opportunity in graduate school to do grant work for the Navy and NASA. Right now, my role at Meta lets me play across the whole people analytics space, so we get to play with data foundations, we get to play with scaled reporting, we get to play with deep research, we get to play with really close partnerships with our HR and business partners, and it is just really a dream and a delight.
David Green: So, really looking forward to the conversation today. Again, before we start though, I should first congratulate you on your incoming role as President Elect at SIOP. Could you share with listeners a little bit more about SIOP and what your role as President will entail?
Alexis Fink: It's one of my favourite topics! So, SIOP is the Society for Industrial and Organisational Psychology, and one of the things that's really beautiful about it is that it is a very strong, balanced marriage between those who are academic, really on the forefront of research, and those who are practitioners. And frequently, practitioners in a lot of professions, they leave school and never really re-engage, but there is a really thriving community of people; and given the advances in people analytics, there is a thriving cutting edge of research work that is happening in practice. So, we do have a really strong annual conference that brings together typically about 5,000 people going deep on new methods and new insights.
As President, I will have the opportunity to help guide the Society overall. We have a strong executive board, made up of people who attend to science and research and governmental advocacy for employment law and have the opportunity to make sure that we are distributing these best practices to our members, and also we're really about advancing this science, not strictly for our members, so the things that we can provide to the PA community at large. I will have the opportunity to spend three years helping to guide our strategy.
I have been an advocate for a long time of people analytics and of that practitioner aspect, making sure that our work is pragmatic, making sure that our work is really helping serve the organisations in which we find ourselves, as opposed to just being interesting, and so I am just really excited to get to spend more time with my colleagues and help advance solid science for better workplaces and better opportunities for workers.
David Green: So, before you got into analytics and in fact alongside some of the roles you've held in analytics, you've held operational and change management roles as well. How have these roles helped you understand people analytics in maybe a more intrinsic way?
Alexis Fink: You know, David, that is a great question and it's kind of a timely one. My opportunity to spend several years of my career focused on change management, and then another several years focused on real business operations, integrating acquisitions and doing business turnarounds, getting really deep into the financials and the mechanics of how work got done, I do believe makes me much stronger as a people analytics professional, because it helps ground me in what happens before and after our research, so why are we interested in this question; how do we frame it in a way that it will be useful; now that we've come up with an insight, what do we do with it, what's the vector, what's the channel?
There's a really common framework that I think my team is sick of hearing me spout about figuring out about what, and then so what, and then now what. The so what and now what is really all of my expertise that came from time in OD and change management, came from time in deep operations. Having a really great insight is like a cool mic-drop moment for television. But if you want to have impact in an organisation, you then have to translate what I have found into what we could do differently to make something else happen.
I used to joke that I got into people analytics as a bit of a power trip, because I didn't want to just see what was coming, I wanted to be able to bend the arc to make a different thing happen. And it's not enough to just say, "Look at that tornado headed straight for your house", we need to figure out a way to bend the curve so that we can have something better happen, and that's where people analytics gets a lot of its value, and that is more than just the statistics.
David Green: And so important, because actually insights are great; but unless you actually deliver actions and outcomes with them, as I've heard some of your peers in the community talk about, it's overhead. You've got to be thinking, I guess, as you're crafting the recommendations from the insights, "How can we get these implemented; and then how can we measure their impact?"
Alexis Fink: Exactly. And in fact, when I talk about a research cycle, years and years ago I got certified as a Black Belt, which was trendy, about a million years ago, tells you how old I am; but one of the things I appreciated about that methodology is that it always included going back to make sure that your recommendations had the outcome that you intended. And we know that broadly in science, the things that you can isolate in a lab setting degrade when you put them into practice by about 50%. So, let's make sure that what you promised will happen actually happened, let's get better at tuning those recommendations, and a team I had a while ago, made them go back and check on how their recommendations had landed and it was really fascinating that that "do" loop deeply honed their partnership with the climate group implemented, and it meant that we about doubled the number of our recommendations that landed.
I don't ever want to get to a point where every single recommendation lands, because then you're not pushing the boundaries quite enough, but it needs to be a pretty high percentage in order to be valuable, because otherwise people aren't interested in what you have to say, it's just a fun party trick.
David Green: And I guess, thinking about your experience in the people analytics field, it helps you separate -- because I think as analysts sometimes, we can easily dive into the stuff that we think's cool, but it may not be the stuff the business actually needs or even the workforce needs, to actually differentiate the cool from the work that's actually going to add value.
Alexis Fink: Yeah, I joke all the time that if you want to make me cry, tell me, "That's interesting", because it is a guarantee you're not going to do anything with it. It's some really fascinating research, maybe I had a great time going down a rabbit hole with some new maths that I thought was fun, and then somebody looked at me and says, "That's interesting"; I know it's going nowhere. So, really figuring out how to make sure that all the way through our research pipeline, you're really close with those partners in defining a problem, in figuring out what are the segments that are relevant.
In graduate school, I could not wrap my brain around Simpson's paradox and the fact that you could have one trend that goes up and to the right if you look at it with a particular denominator; but then if you segment it out, each segment within that up and to the right might be going down to the right, and really understanding what's a meaningful chunk so that we can deliver insights that are actionable, instead of doing harm by delivering something that is poorly conceived. And over and over again, it's those business partners, the leaders, the HR VPs who pick at you and say, "Yeah, but…" And, if you've waited until the very end to deliver an insight, you've lost an opportunity to build partnership, to build trust, you've wasted cycles, you've potentially really made yourself look like an idiot. There's just all kinds of things that can go sideways if you're not investing in that deep partnership.
David Green: I'd love to hear your reflections on how the people analytics discipline has evolved over the years.
Alexis Fink: So, when I started, it was a lot of things I would think of as very classic. I was doing a lot of job analysis kinds of work, for example, which is entirely unsexy. Usually now, if you're doing that, you're calling it competencies. But it was really buried down six layers in an organisation, wasn't terribly strategic, it was often to meet employment law requirements and was often quite focused on job analysis, often quite focused on performance management. Sometimes, you'd get to break out and do some work on leadership, or effective teams, or effective management. I even occasionally got to do time-and-motion studies to figure out how to price particular products based on the staffing that was required to do them. But it was really focused on utility and really focused on back-end operations of making HR work.
If you look back further beyond my time in the career, you can see this decision science emerging from those operational disciplines, and I described a lot of things that were pretty operational 30 years ago. Now, we are with the CEOs and delivering data to boards of directors on, "Here's what your most critical asset, or certainly in many cases your most expensive asset is doing", and it's really fascinating; I'd jump to hear others talk about this. We saw differentiated competitive advantage emerge from different kinds of decision science over the last, loosely call it a century.
The turn of the 20th century, you saw finance emerge from the operational disciplines of accounting to really create differentiated advantage, and you can't imagine a large company without a finance function now; the same way, mid-century, the sort of madmen era, you found marketing emerging out of sales. Certainly sales is still really critical, but marketing as a decision science, figuring out how to configure our products, for example, or think about to whom we might sell; and we're really seeing the same thing happen in the people space now.
It doesn't take anything away from the need to really do excellent HR still, but the decision science around people analytics, to really figure out how to make best use of the skills and capabilities that you have in your organisation that might be changing, how to solve the perennial problems of performance review that helps instead of hindering, all of these kinds of things are really emerging now as a decision science that are creating competitive advantage for the organisations that are good for them, or good with them.
Now, I think we're actually at a moment where it's hard to imagine a copy that isn't doing some of this. I'm seeing tiny, little 500-person start-ups that are beginning their journey with a focus on people analytics.
David Green: You've worked in you mentioned Microsoft, Intel and now Meta, you've worked for three of the leading companies in the people analytics space. I think most people in the field would recognise those three companies as leaders, I think, in this area, and you've seen various different people analytics team structures and operating models. What are the different types of structures that you've seen; and in your opinion, based on that experience, what would you say is the best way to build a people analytics team?
Alexis Fink: It is interesting that there are lots of ways to solve this problem and depending on the opportunities in front of you, different answers can be appropriate at different times. One obvious model is to be very distributed and deeply embedded with clients. So you see Amazon embracing this, they have lots of people analytics teams tied to lots of different parts of the business; and even within that, different functions within those parts of their business, and those teams are often pretty self-contained. You see this in lots of places, and there's really a lot to recommend it, when you have high-quality data infra, when you have solid practices, when you have good coherence in approach, then getting that data team, that research team, as close to the business it supports as possible, it really helps make sure that their work is deeply aligned to the problems at hand.
You can start to risk contradictions with the rest of the ecosystem, you can potentially miss some opportunities for internal mobility, and if the business conditions that make that deep partnership workable aren't there, you can also end up with some pretty bad things happening. But in general, that's a very workable model.
The other kind of index, or access, as opposed to reporting line, so is it centralised or distributed, is how is it anchored? It's fascinating to me the extent to which people will use the same term, people analytics, to mean different things. I will categorise them as people data, which might be much more reporting-centric, and there are some very esteemed folks who are basically running reporting shops and you know what? Their clients love them because they can get answers, and they can get those answers quickly, efficiently, accurately and often those reporting focus shops can do a lot of custom slices that does put the onus on the client often to do the sense-making, but the data is available and it's very satisfying.
The other access to index on is basically new information, I characterise as people research, where you're really doing new, original research. And there, you mentioned earlier the pandemic and distributed work; so there, rather than just counting how many people are coming into the office or not coming into the office, you'd be getting underneath, what's our productivity in each of these places; and what's their tenure and their job level; and how many of their clients are coming in; and how much are those other features predictors of their satisfaction and productivity, and all of these other things?
So the vector of, am I a reporting shop or am I a research shop, is another important one to consider and again, that goes to what the business needs; what's the tolerance; what are the assets that you have available? Interestingly, when folks are the first PA leader in a team or in an organisation and they come to me and they're like, "Hey, Alexis, can you help me figure out what to do?" I'll often advise them to start as a research team, because you can get there quicker, you can do some bespoke work and you're not as dependent on your HR data being in an analysable state. So, that can be a way to establish trust and start building value and start building relationships, while you do the fairly laborious work of getting whatever your HRIS is in a state that you can do the reporting that we need and get the data definitions, etc. Most teams will need to do some of both, but there's also kind of a complementarity and sequencing.
Additionally, you can use the reporting as a Trojan Horse to get to some of the research stuff that maybe the clients didn't think to ask for.
David Green: I think another thing I see particularly with companies that are maybe younger in their people analytics journey, they hire a lot of analysts, data scientists, hopefully they hire data scientists when they're actually doing data science, and sometimes it's quite an inward-looking focus on the data and not enough of a focus perhaps on the business priorities of the workforce as well. And as you think to build the team, whatever structure you've got, I'd love to hear your views on -- you need people in the team who are good consultants, I guess, who can really start to -- someone comes with a problem that might not be the question you want to answer, you might need to ask five questions to get to the question and the hypothesis that you'd want to test with analytics, yeah?
Alexis Fink: You know, David, you're exactly right and my last two teams both had a whole function that were really in those consulting roles, that were deeply connected to their businesses and the problems that they were facing and the quirks of their leaders and the structures and the history of how they got there, so that when someone asked a basic, "How many…?" question, they could figure out what's the real decision we're trying to make; what's the problem; what's the pinch-point in this organisation we're trying to get underneath?
Most people in PA roles, and frankly most people in lots of roles, have experienced a sort of "bring me a rock" problem, where someone gives you a poorly articulated need and you answer that need like, "That's not really what I wanted. Could you look at it this way? Could you cut it this other way? Maybe you could look over here?" and sometimes it's just fishing expeditions where they're trying to get away from something uncomfortable. But a lot of times they just haven't articulated it well. And if you just take that request at face value, you're going to end up going through many frustrating iterations, especially if each iteration requires that you do a new data pull, because they ask you to pull in a new variable that wasn't in your original set.
So, it's not a five-minute, "Let me add this", it could be, depending on the state of your data, it could be a couple of hours of extracting something and then joining it and cleaning it up and doing all of your transformations, and then you can run the 30-second analysis. But it just ends up sometimes consuming days or weeks of iterations if you don't invest in that deep business consultancy. And then if you hadn't built the trust upfront to also be able to say no, or actually I tell my team often, "You don't have to say no, just don't say yes yet". Unless you understand what decision I'm trying to make, I might not have time for you, because if you're just going to some kind of curiosity thing, or you're getting ready for a big meeting and you want to show off that you know this number but you're not actually making a decision, I've got a lot of requests here, people, we've got to prioritise. And if there's not a real business outcome, I'm not asking my team to work nights and weekends for this.
So, that laser focus on business impact is something that some teams do really well, and others maybe focus on just what's your SLA for answering a question, as opposed to, did something different happen in the organisation because of it. It's not to say that an SLA isn't important, it's good to keep your promises, but really thinking through utility is very valuable, and I've seen a lot more conversations about that since we're either in or flirting with a recession and in or flirting with staff reductions at lots and lots of companies; all of a sudden, we're thinking through how we demonstrate value in maybe a more urgent way.
David Green: What are your thoughts on the importance of a relationship between a people analytics team and their HR business partners and the Chief People Officer?
Alexis Fink: So, the meeting I have immediately after this podcast is with one of my HR VPs, so I guess they're really critical and I need to spend much time with them; all the context things we've talked about and the problem definition things. Also the HR VPs, or in some cases the COE leads if you're talking about talent management or learning and development or others, they're the ones who have to consume what you're building, they're the ones who actually make the magic happen. And if you're not (a) meeting their needs, and (b) delivering in a way that's usable, you're not actually terribly valuable, you're just a pretty ornament.
So, spending time with those folks, as we've said, to make sure we're solving the problems, and solving the problems looks like not just good insight, but good insight that's tied to a choice that they might actually make, whether that's a programme design choice, or a forum design choice, or whatever else, and then if we -- one of the things that's been really exciting in really primarily the last five years is seeing people analytics come out from underneath maybe a talent management or a selection or an HR operations or even an HR strategy role, or sometimes even a corporate strategy role. They were a couple of layers, in many cases, below the CHRO.
Over and over again, you're just seeing that people analytics leader report directly to the CHRO, and then that person can serve as a little bit of an audit function, that person can serve as a little bit of a strategic date to make sure that we're spending time on the right bit problems, and also can help create more robust narratives for that CHRO, as opposed to having everything filtered through. And while I believe that all of my partners, literally in this job every one of my partners is a sincere and wonderful and just brilliant person, there's still a game of telephone, and anything that I say gets filtered through that person's priorities and then, if it has to go through that person to get to the CHRO, then there's been an extra layer of priorities put into it and some key insights may, through no malice, but just through that person's lens, get dropped out. And joining those more closely I think is really, really helpful.
It's been gratifying in the last few years to see how much of our HR strategy, other organisations' HR strategies, are really filtered through or influenced directly by people analytics, and that is a change over the last decade. It also is a tremendous opportunity and a significant obligation. If you are really in a role of guiding strategy, you have an obligation to make sure that you are being pragmatic, that you are being focused on business value, that you're thinking about scale, that you're thinking about all these other things that when you're in more of an ivory tower-style research function where you're like, "Look at this cool finding!" you need to make sure that if you're going to guide your organisation on the basis of it, you need to be really confident that it's real and true and will deliver an outcome.
David Green: It's interesting because the research that we've done the last three years, each year the number of people analytics leaders reporting directly to the CHRO has increased, it's over 20% based on the 184 companies that participated last time, and over 80% that at least report to a member of the HR leadership team. So, I think as you said, it used to be buried down several layers in the HR function and it's difficult to add business value then, isn't it?
But if you're at the top table (a) it sends a message to the rest of HR that this is important, (b) it probably helps you to get access to the right stakeholders in the business as well, and potentially more investment as well to build that function out the way it needs to be built out to deliver the value it can deliver.
Alexis Fink: Yeah, and one of the things that I think has changed in concert with that over the last several years is I have seen a shift away from people analytics being largely a vendor management function where you're like, "Here's the vendor who does your selection stuff, and here's the vendor who does your performance management stuff, and here's the vendor who does your HR reporting, and here's the vendor you're using for learning and they're going to give you some learning data". And as we have brought that inhouse to get the strategic business value out of it, to get the intimacy between the data and our own business, that necessarily floated up.
When it was just, "Find the vendor who'll give us the services", that's not terribly strategic and it makes sense to be a couple of layers down. Where now, it's really setting strategy that lands a little bit closer to the top of the house.
David Green: Yeah. Again, we see a lot of companies that are blending people analytics and people strategy together which, as you said, you shouldn't really have a strategy unless it's informed by data, should you? On that note, again staying with the HR VPs for the moment, what skills and competencies and capabilities and behaviours maybe do you think HR business partners need to be effective as the world of work continues to evolve?
Alexis Fink: It's really interesting. I've seen, and I think you've seen as well, a bifurcation, or maybe even a trifurcation, in the kinds of HR VPs that are out there. There are some that are very execution-focused; there are some that are very classic OD, leader-coaching, change-management-focused; and then, you're starting to see the third leg that are really more data-focused. And the challenge is that the HR VP function needs all of those. We have to do performance management, we have to pay people fairly and adequately, we have to build hiring plans and build benefit plans that are germane, regardless of where that lands, etc; and the strategic OD pieces of that really should be yin and yang to the strategic data pieces of that.
When you think about those three legs, I think all of them are really critical, even for people in the first two legs, if you will. I think that given the way the world has changed over the last decade, at least data literacy becomes really important. It's been a while since I've had an HR VP look me in the face and say, "I just don't believe in data", and I used to get a lot of, in fact I used to do a talk about it, a lot of what I referred to as the "don't talk to my boyfriend" problem, where the HR VP really owns the relationship and you come in with something different. And depending on the business leader, sometimes they really like the something different you're providing.
Instead of seeing that as a partnership, there've been HR VPs in the past that have really seen it as a threat, and I haven't had that experience over the last maybe half a decade either, where there's more of a realisation that this is a vector for power and influence, this is the lingua franca of business. So, to the extent that we can make sure we, as people analytics professionals, are equipping and lifting up the HR VPs, we can make sure that we don't get into the "don't talk to my boyfriend" problem.
The flipside of that, the one that we have to own, is -- I'll talk about one of the quickest ways to get iced out of a relationship is by swaggering in there and saying, "Look how much better I am at your job than you are". That's not a way to win friends and influence people. I can see you shaking your head! So, really recognising that this is a partnership and it is our job, both of our jobs, the HR VP and the PA leader, to make sure the business is making the best possible choices, and we are both bringing things to this relationship.
I've used an analogy in the past that it's almost like the HR VP is the parent of the child, and I'm the paediatrician, and I'm going to come in with some tests and some data and some standards and, "Are they gaining weight?" and whatever else, but you know your kid. So, between us, we can make sure that kid stays healthy. But you're the everyday relationship and I have the kind of, for that infant, the every-three-months relationship. So, with that as context, data literacy, really good partnership and then all that same business and strategic acumen that we need in PA.
David Green: Moving now to people analytics, so as people analytics continues to scale and grow, and I know in the organisation, Meta, and in Intel and Microsoft, there are some pretty significant-sized people analytics teams, what are the key skills required now in people analytics, and what are your thoughts on some of the emerging skills in the field as well, such as maybe NLP or some of the other skills that we're seeing functions build?
Alexis Fink: Excellent. So, if you look back even 15 years to the way that data science was being defined, that's actually a good frame for us. There's a pillar around content and there's deep content expertise that we need in order to be able to form reasonable questions, to be able to know what kind of maths to use, what's the nature of the underlying variables. There's a whole bunch of stuff that's really about content. There's also a bunch of stuff, to do this well, that's about basic blocking and tackling of data. How do you tease data out of the cosy, little warm hidey-holes and get them into a usable format, and ideally do that in an ongoing way so that you can be efficient in your analyses?
There was a time when people talked about 90% of their jobs being basically data angling. As we get better and better at this, that percentage is going down, but it is not at zero yet, so that pillar of data work itself, data engineering transformation, etc. Then you get into stuff that everybody thinks about as people analytics, which is the statistical and analytical pieces; so, as you know, as we've talked about so far, all of your reporting kinds of things, all of your parametric tests kinds of things.
Knowing regression is probably -- Keith McNulty has a whole text book on regression, it's probably the most like the Swiss Army Knife of people analytics. But knowing when you're using linear versus logistic and some of the rules about how to do that well, there's been some super-exciting stuff in NLP. We can have a whole conversation about GPTs. But even apart from GPTs, there's all kinds of great stuff you can be doing with skills extraction, all kinds of wonderful things you can be doing with comments in a survey even. You mentioned engagement surveys and employee listening. I don't know how many people listening to this have ever done the mind-numbing work of reading 5,000 comments to try and make sense out of them, but holy smokes, a good NLP can zap that in order pretty quickly and it's deeply, deeply powerful.
I'm a big fan of optimisation modelling under that analytical stack, because it gets you closer to some of the recommendations, "If you want this outcome, here's a way to get it, here's how we want to move forward", whether that's a junior/senior mix, or whether that's a diversity question, or whether that's looking at some succession planning outcome; using those kinds of techniques I think can be really, really powerful, and often not taught in certainly higher psychology graduate schools, although they're beginning to be more so.
Then, people will disparage it as pretty pictures but the older I get, the more I am convinced of the power of good data visualisation, which I put in that same analytical stack. And again, people will disparage it, but there's something amazing when somebody can walk past a conference room, glance sideways at what's on the screen, and understand what you're saying. And the first time I put just a block chart up in front of an executive and within a second she was like, "Oh my God, that means X and Y and Z are true", and I was just like, "I don't know how you go this from this, but now that I tilt my head, I can totally see that you did". The speed and power of a really good visualisation is something I think we underestimate.
If I move past the standard way we think about content, data work, analytical work, there's something else that's really important for projects to not fail, and when I saw failures they were not usually in content, they were not usually in data, they were not even usually in, "Dude, you chose the wrong maths and you violated statistical assumptions", or whatever, they were usually a failure of influence, which goes back to that relationship with the HR VP, it goes back to making sure answering a question the business cares about, it goes back to making sure you move from insight to implications and then finally to a real action. That influence layer is sort of where great research goes to die.
David Green: As the incoming President of SIOP and an IO psychologist yourself, I'd love to learn more about your thoughts on the importance of IO psychology in people analytics. I guess in sits in that statistical and analytical layer, but maybe calling out why it's important, particularly when we're thinking that ultimately it's about people, I guess.
Alexis Fink: I'm going to have to buy you a drink later for giving me the excuses to talk about IO! Yes, it's very much in that analytical layer, but it's also really in the content layer. Robust education and IO psychology will get you into a century of research. Our flagship was launched in 1917, so literally a century of research about how selection works and how performance review works and how management works and the distinction between management and leadership and how power moves through organisations and how you make a team effective and what happens with different work settings, which we actually started studying in the 1970s. All of the content about how these things work, and even getting into cognitive biases and a bunch of these foundational things, they come to play with us at work.
So, I do think that IO psychology has some real advantages in the content space that then can be married with computer scientists who will have all the data engineering, and in some cases some statistical techniques that are really complementary to the ones that we would have learned in a more pure social science framework. And then, when you can marry that with the content expertise coming out of an IO psych background, you can really do some powerful things.
Similarly, every team I've been on has had folks who came out of an HR VP profile, who could really do a lot of that influencing, a lot of the pragmatics, making sure reports were usable. And so, when I think about the role of IO psychology, this is my own bias showing, I see it at the centre of this, and then really augmented powerfully with folks who know the business, who know influence, folks who really understand how to work with data and make it efficient.
I don't code any more but back when I was, I could do it but it was never great. And then, amazing code that somebody who's really, truly a data scientist, really, truly a SWE can make, software engineer can make can just accelerate our impact.
David Green: There's a lot of focus at the moment about upskilling HR professionals, particularly around data literacy and some of the things that we've talked about. But what I'm interested in is you, as a people analytics leader with a big team with a variety of different skillsets within that team that you've got at Meta, how do you develop your team to ensure that they are continuously learning and adding value?
Alexis Fink: There's decades of research that shows that most learning actually happens on the job. Even if you decide you're going to take a free class to learn how to do regression, you're going to learn it best when you have a real data set that really matters and you have actual problems to struggle through. So, making sure that we are structuring the work in a way that people get to learn and stretch generally hasn't actually been a problem, because we have new problems coming at us all the time; and take the opportunity to say, "You know what, why don't we try an optimisation model with this? It might not work, but you know a little bit about that, let's do it once, let's see what it comes out with, it might give us something interesting", I believe strongly that we need to always be interrogating data really, really thoroughly, and then presenting the simplest possible version of the correct answer.
So, I'm comfortable saying, "Let's try looking at this analytical method. This guy's really great at NLP, why don't you go and talk to him and see if we can apply some NLP to this problem", and that ends up creating a lot of peer learning opportunities, where folks are building on the strength of what they have internally. It also ends up creating a lot of practical learning experiences, where someone will say, "I did an NLP example at an old team". There was someone who ended up doing something that was for his school district, his kid's school district, it's like, "I want to learn this technology and I don't have a data set I can use here. Here's this other one I'm going to do to get really good at it", and then we found a way to actually take advantage of that at work.
Learning on the job in some pragmatic way, a lot of peer learning, we've always structured brown bags, etc, and then a lot of community work. Actually, there's a really vibrant community on LinkedIn. You're actually a stalwart of it, I love your monthly summaries, I read them religiously. And then, of course, there is a role for formalised learning. I have found that in our roles, there is more of a narrow role for formalised learning. We did do, in my team about a year ago, a class on data visualisation, just to bulk upskill there. But if you're thinking about ways to simultaneously enrich your team and also recognise and reward expertise on your team, some basic lunch-and-learns and brown bags and peer relationships and joining people together on a project is pretty powerful.
David Green: How do you predict, and I won't hold you to this, I promise, how do you predict the people analytics will evolve in the years ahead?
Alexis Fink: If I just think about the next three or so years, I do see us getting to more focus and more business alignment, really being clear about that impact, in a way that not everyone has been. I also see us focusing much more on scale, instead of doing a cool thing and walking away, figuring out how to replicate scale, etc. The technology landscape in people analytics is kind of a Wild, Wild West at the moment. It's had a lot of consolidation, but I think there's more maturity and evolution that I expect to see.
I'm curious about the extent to which we will, as a function, as a profession, evolve to thinking about workers and not just employees, particularly as we start to see different ways of working. We've been talking about the gig economy forever and it seems to have stalled, but it's a pretty unstudied area. There's some weirdness with labour law, we have to be careful for compliance and other things as internal practitioners. But as external practitioners, etc, it will be interesting to see us thinking about workforces as a whole, as opposed to only employees.
Of course, GPT, NLP, technologies like that will fundamentally reshape some things, even in the next year or two. I think already, it's fundamentally reshaped things in the last six weeks. To the extent that it replaces some of the didactic work, some of the "tell me about" work, I think it creates space for us to do more problem-solving work, and I think that that's actually interesting. I've never enjoyed writing action plans and if GPT can do that for us, God bless. Figuring out which action plan to use and how to implement it is interesting work, and I think GPT can't do that for us yet.
Then, the last, I was actually just thinking this morning about how the way medicine is evolving is a lot of the way I see PA evolving. There are new tests, there's new technology, there are new ways to customise at the epidemiological end. And then, there is a ton more quantified self, a lot more data that's available for the individual, and if we can equip folks to make good decisions without needing to touch us specifically and just use scale, I think that's really powerful.
Then, there's a ton of other allied professionals, where you've got physical therapists and these kinds of folks in the medical example. We have, I think, potential for lots of folks who are literate and specialised in one area, who can then take some of this insight and apply it to a particular business area in a way that then extends the impact of the central PA group. And I think thinking of it in terms of those three areas and maturing those three areas is an interesting model as we look three to five years out in the future.
David Green: Yeah, really interesting, and guess that kind of scaling, that productisation I've heard it referred to as well, that brings in maybe other skills that we haven't talked about yet about thinking about putting the user at the centre, thinking about UX skills, thinking about if we're developing products, just like obviously you do at your organisation; it does create products that people use, and they do use. An exciting time, I think, and more excitement to come.
So, Alexis, finally this is a question we're asking everyone on this series, and it's a kind of bigger picture; you may even bring some of what you've just spoken about into this. What do you think HR leaders, or CHROs, need to be thinking about most in the next 12 to 24 months, and I know it depends on your organisation a little bit, so maybe as a field; what is your biggest concern, and what do you see as the biggest opportunity?
Alexis Fink: My biggest concern is the extent to which, are we in a recession, are we not in a recession, what's going to go on, how do we need to think about those things. So, let's set that a little bit to the side. I do think we have huge opportunities, whether there were recessionary forces or not, to really get reporting right. It's still mostly, "Oh, that's interesting", in many cases, so really getting that right.
I think in the next 12 to 24 months, we have an enormous opportunity to be focusing on measurable impact, and that could be through this reporting, it could also be through the power of new techniques, not just the tools, but really making them pragmatic and joining them to business decisions. So, regardless of broader economic forces, I think that there is attention on the people part of the business that has been much, much more concerted attention in the last few years.
I think that that will continue, and I think that we have opportunities, HR leaders have opportunities to really understand what's happening in their organisation through solid reporting, really understand what to do about it and how to bend that curve to the outcomes they need, through analytics tools and techniques; and make sure everything we're doing in the people space, analytics space, learning space, recruiting space, performance management space, the full deal, make sure that all of it is really focused on measurable impact back to the business and to the individuals that comprise that business.
David Green: How can listeners stay in touch with you, follow you on social media, find out more about your work, find out more about SIOP? I know you've published books and other publications in the past as well.
Alexis Fink: Yeah, I do have two books, one actually on people analytics and pragmatism utility, showing the value, called Investing With People. That's with John Boudreau and Wayne Cascio. And then, the other one is Employee Sensing and Surveys. Both of them are available on Amazon. If you just want to generally be in contact, LinkedIn is the best way to do it. That has an amazing, thriving people analytics community and I'm not as active as you are, but I try to stay up with that. So, I think that's probably the best way.
David Green: You're pretty active, Alexis, which is good, because I think it's good that people share together as a community. So, Alexis, thank you very much and hopefully at some point this year, we will see each other in person as well.
Alexis Fink: That would be lovely. I'll be at SIOP in Boston in April if you want to hop across the Pond.