Episode 71

Douglas Silverstone on Balancing Cultural Fit and Diversity in Recruitment

TIM: Douglas Welcome to the objective hiring show. Thank you so much for joining us.

DOUGLAS: Good morning! Nice to be here; looking forward to having a good conversation with you.

TIM: Yeah, it's great to have you with us. It's a pleasure, and I'd love to start our conversation today with something that I've been thinking about a lot over the past few years working in data recruitment, and that is that there's a bit of a, I feel like almost an oxymoron out there between companies that are saying they want to have this kind of cultural fit-based interview process where they're looking to fit people into an existing culture. Make sure the candidates can come in and be part of that ecosystem, but on the other hand, seeking diversity and these two things seem slightly at odds with each other. I'd love to get your views on the topic.

DOUGLAS: Yeah, that's a really good starting point because it's something that we're a social housing organization, so we look for—we have a social purpose, and things like that really matter to us. We have to represent the people that we serve and give houses to, and we work with them. So if we don't reflect that population, it becomes really difficult for us to operate culturally, but equally we have a culture we are driven by the customer. We need people to focus on that, and it's really easy to look for people that look like you and look like the organization. Actually, we also know we want diversity. So there is a friction. I think between looking for someone who's going to bring you something extra and something additional that builds on that culture and also something that can keep you going and keep you moving over the same way in the same way, but if you don't look out, you can't get there. and I think it's really interesting clash, especially when you start talking about things like neurodiversity when Quite a lot of those differences are hidden, and different approaches are hidden, and the way people's minds work doesn't always come out in an interview; they just come out as different. and if you don't bring those in, you're not able to see whether they bring the culture, whereas the culture, once you're in, I've also had people where we've brought them in, and they just haven't matched the culture, and it's been really disruptive to the rest of the team, so I find it really difficult sometimes to work out why that cultural change is different and why you need to bring that in and how you actually adapt to that. It's a really interesting conversation. I certainly don't have all the answers, but it's something that I've certainly come across and built it on.

TIM: I wonder if there's also something to be said for whether you are in a steady state where you want to maintain the culture like you feel like your culture is very strong and anyone coming in needs to be able to adapt to it, or maybe you come into a new team and you think Oh, I got to shake things up a bit. like we need to actually change I want someone to come in and be almost aggressively different from my current team; one of theirs is almost an element of that as well.

DOUGLAS: Yeah, certainly. I don't know because all organizations are the organizations they are, and they reflect those organizations, so you hear all the sayings about that an organization is only as strong as the people in it and things like that that you hear all the time. I think there has to be something fundamentally wrong if you have to come in with the organization if you have to come in and completely rewrite the culture from scratch. but equally organizations that don't evolve the codex fall away, and if you don't bring in new people, you don't see those codex moments; you don't work through there, so I think it's about balance, and it's about finding the stage that you don't have to revolutionize your culture; you adapt it and evolve it. Now what's really interesting there is almost implicitly in that you're saying your culture can stretch from here, from A to B, but it can't stretch straight from A to Z, and there will be some people that are just too culturally different to come into that. Now that's very close to saying, actually, I am going to discriminate. and I'm not going to discriminate, but it's not on personal traits; it's not on individuals abilities; it's on where their career paths are and are they ready for the organization that's going to be in, and are they going to bring the organization to the next level, so I think it's that that makes the difference. It's got to be objective on what the organization is, where the organization is working to, and what the people can bring into that, as long as you're quite open with your own biases and understand that everyone's got biases and you can understand how to bring those in and evaluate that and give people an opportunity to talk through how they'd approach things differently so that they can have the space to demonstrate that they can enhance the culture.

TIM: I wonder if it also has a little bit to do with the current competitiveness of the industry you're in. So I've been to see just today some comments from the relatively new Manchester United owner, Jim Ratcliffe, where he was just smashing the current culture of Manchester United in terms of a lack of data analysis in recruitment. funnily enough, that was his biggest pain point; he thought we were like 50 years behind we're 50 years behind what a Brighton or a Liverpool or whoever else is doing and needing to change that; otherwise, they simply won't be able to compete. I wonder if in something like sport, where the difference is so obvious because there's a literal league table, like you can measure it directly. that maybe there's a motivation there to change things radically when things aren't going well, whereas maybe in any other normally more normal industry that maybe doesn't change as quickly, maybe it isn't when things aren't going well that it isn't as obvious or as immediate. I wonder whether there's also something to be said there for kind of not rocking the boat, as you say, stretching the culture but not going for someone who's completely left field.

DOUGLAS: There's no one ever got hired or fired for bringing IBM kind of thing, isn't there? But eventually IBM wasn't the right tool for the job, so you do need to move on eventually, but you've got to move on at the right time and move through that. I think it's also there is that what is it to be an organization? and from a data analysis point of view I have Keep coming back to organizations and people that say, We're a data-led organization. We are data," and I've got a long background in data, so what I'm about to say may seem counterintuitive, but data can only take you so far, right? If the data shows that Man United should stop playing football because they're that bad, then instead they should start playing something else. They should play a different game. They start selling fish; it is a thing they should use. Man United is not going to give up football and start selling fish even if the data shows it's more profitable, so there's something intrinsic about Man United as a club, as a team, as an organization, that they are there to be the best football club in the world. not doing so well, so how do they meet that game? Data can only take you to the point where the organization wants to be, and that, I think, is the difference between data and culture because the culture bit is, yeah, we could sell fish, but that's not us, and the same for us: we're a housing association. We provide houses for about 88, about 80,000 properties. It's a large group of people that live in our properties, and that's our purpose. We don't make a profit; we're there to make their houses and their communities as strong and as lovely as they can be. Now there's always monetary constraints, but that is what drives us as an organization. So I think coming back to the point of culture and diversity, you've got to buy into the aims of the organization. If you buy into the fundamental aim that Man United needs to be the best football club in the world or Southern Housing needs to provide decent homes for residents, that's the starting point. and if you can buy into that and you can show how you can support those aims, then the rest of it is fine. It shouldn't matter about diversity, where your background is, what your situation is, where you come from, or your social housing background, whether you're gay, straight, or neurodiverse. All those other things shouldn't matter because you can use that to your advantage and the organization's advantage to drive to that goal. So maybe that's the balance that we're looking for.

TIM: I just can't shake the image of Bruno Fernandez flogging fish and chips down at the Quay. I feel like there's a meme in there somewhere. The Liverpool fans would love that one. One personal challenge I have with fully embracing the idea of hiring for a culture is the fact that it gets into very subjective territory in terms of actually deciding if this person is a good or bad person. So the way that we did this ourselves was we spent a couple of years hiring people and realizing this isn't working very well; we need to better define who we want to bring into the business. We then came up with our core values that we thought we partly already had but also partly were aspirational values of the company we want it to be. and then we incorporated that directly into the hiring process to ask questions that were somehow going to try to evaluate these things, like do you have a growth mindset, do you make it happen, and are you a no-bullshit person? That is one of our values as well, but it's still quite subjective because those things are not black and white like a technical skill. and so I just find it difficult as a data person to not think that despite my best efforts, I'm probably still at some level selecting for people that I like or that I have some favoritism towards that might not necessarily even be connected to those other values. What do you think about that?

DOUGLAS: Yeah, we have the same right, so we've got Southern Service, which is about how we deal with customers, and we've got our Southern Southern Housing Values, and they are both front and forward, and they do really perforate through everything we do as an organization, but in the recruitment process sometimes it's easy to forget because you've got the technical skills, and you've got the social skills, and you've got the business skills. the culture value, so I think when you're working with recruiters or recruiting internally, getting those values out to the people who might apply to show what you're looking for is half the battle so that it's almost self-selecting. You're saying to the organ because recruiting is a two-way thing: you want people to want to work for you, and you want to find the right people. So if you can get out there, this is what Southern housing is really all about. This is providing great homes is what Southern housing is all about. This is you want to help with that; you can do that. These are our values: we're honest, we're trustworthy; they're heart values for us, and you can Google them, and you can find them all there. but they're so honest, trustworthy, efficient, and accountable, that kind of thing, and if you get out there and people really believe that's what the organization is about and they see that through talking through the networks that we all have, then hopefully the people you attract, it becomes a retention thing, and they will buy into that. and then it can be actually wherever your background is, as long as you can demonstrate those; that's the door into the organization. Once you're in that door and you've said, Yes, I buy into those values; I really understand it, the rest follows. I think where it gets really hard to test, as you say, and I've seen, I'm sure we'll get on to interviews in a minute, but where it gets really difficult in the interviews with this sort of stuff, you end up asking six or seven technical questions. What's your how do you write in R? What's this sort of thing? How would you convince stakeholders that kind of and at the end you say Hey, oh, and how do you demonstrate the heart values, and that makes it very different? Really, the heart values need to be demonstrated in how you write R; they need to be demonstrated in Python. and like some of the questions, typically when I'm interviewing someone, normally what I ask is how can you put the heart values—how can you put equality and diversity into the work that you do as a data analyst? So if you're coming to be interviewed for a data analyst, it's not enough just to say that the normal Oh yes, I believe in equality and diversity. Yes, I've done this; I've done that, but actually, as a data analyst, how does equality and diversity work in your role? When you're writing reports, are you writing those reports in mind that you're actually some people might be colorblind? Are you writing reports that make it easy to read? Are you thinking about that in the way that you write those reports? So you're testing then, and questions like that, you're testing both the technical skills and the cultural approach skills, and that, I think, does start to give you something that you can tease out, and that's about actually doing my job just the way I work, always valuing work into the values, and that's a much better way than just keeping it entirely separate. I'm a nice guy, which is quite often how those questions can come across.

TIM: I'm wondering the way you describe that then You could almost infer maybe by the candidate's absence of connecting to those dots in their technical answers that they're not the right fit if you don't even lead them down that direction and they don't even touch it. Is that part of the evaluation itself? or would you not expect them to make that leap necessarily?

DOUGLAS: The best candidates will, but I think, again, you're coming down on another thing, so if you're hiring for new diversity and if you're hiring for people that maybe they're not roles that need to be forward fronting and people-focused roles, then in an interview, expecting them to be able to understand all of that and put it all into context is going to be hard for them. So you do, in that just in that judgment, you filter out a layer of people, so if you've got—if you're on the ASD spectrum, that might not be something you think about, so as an interviewer, you've got to pull that out, and as I say, interviews work for interviews; they don't necessarily work for how people do their jobs. after the interview, so I think it does, and if people come out and they'll say exactly that they can demonstrate it, they put all the dots together, and it's, Wow, you've really proven you've really understood it," but I don't think, as someone doing the interviews, you can stop there and just rely on that. think you have to probe, and you have to get again I think a lot of this comes from pre-interview to how you present yourself. Like, I try as much as I can to give interview questions to all candidates for roles in advance, so it's not about someone having to say, I want an adaptation. It's actually what does Keeping an interview question secret, what does that benefit me? Why does that add anything to the interview process if I give them to you in advance, and I say on the application form and I say on the advert you'll get these in advance, and then I do give them? I'm demonstrating that I'm honest; I'm demonstrating that we're open; I'm demonstrating that we're effective, and you as an individual can start seeing what we're about. and in those questions you can start seeing actually each one of these questions has a cultural element to it. The culture is really important. If I'm lucky enough to be working with recruiters, those recruiters know what's important to me and what's important to our organization. They can steer that before they even get to the interview. So when you get in the interview, you've already done that, and then it's making people comfortable. If you make people too comfortable, and you can, I've been in interviews where you can clearly see they're using ChatGPT and other things like that in the background, and we're going to maybe get into that because that's not necessarily a problem. but my point is once they're in the interview, you've made them as comfortable as possible; you've given them the best ability possible to show how they match your cultural skills, and then the technical skills come as part of that, and I'm a big believer that yes, you need to do the job, so you need the technical skills, but you can train a lot of technical stuff where the culture where it falls down When I brought people in and the culture hasn't worked, it's a lot harder.

TIM: One thing I was thinking about recently because I traveled to a whole bunch of different countries that were vastly different from where I live in Sydney is that I went to Thailand, and I went to Saudi Arabia. I went to Germany, which is all very different from here, and one thing I noticed as I was going to these different countries is that you pick up on the local customs and culture pretty quickly. So, for example, I arrived in Riyadh, and my friend picked me up at the airport, and I had gym shorts on. He just put some pants on for starters, so I went and put some pants on, and then it was okay. It was a little bit more acceptable where I was in Bangkok or anywhere in Thailand. If you, let's say, I don't know, buy something from a shop or pay someone for a coffee, you wouldn't just pay and then leave; there's a moment of almost a bow. almost a moment of reflection where it would be quite rude just to immediately dart away, and you wouldn't know that until you went there, and then you just start doing it without noticing, so I wonder if we don't necessarily give candidates the credit for being adaptable to fit into an environmental culture that they're coming into because most people are adaptable at some level to an environment. like you're going to have to be a very outlier, aggressively disagreeable type of person to say, I'm just going to act exactly the way I want to act no matter what the circumstances, so is there an element of over-indexing for almost like how they behave in an interview or what current values they demonstrate, and we might not give them enough credit for being able to adapt to whatever culture they're joining?

DOUGLAS: Yeah, I think so, and when you were saying that, it was like thinking, so if you have AST or something like that, that's the same thing, so that culture in Riyadh is going to be the culture in Riyadh is what it is. You land, you don't know you're not supposed to wear gym shorts, and if no one said to you, Put some trousers on, you'd have carried on wearing gym shorts. Now you'd eventually realize no one else is wearing gym shorts, but what if you didn't realize that no one else is wearing gym shorts? You might be the best recruiter in the world, but every time you go in for a meeting, you get kicked out the door, and it's not because of how good of a recruiter you are; it's just because you haven't noticed you should be wearing trousers, right? and I know that's the but that could be the same when you put it back in as like interviews for technical roles where ASD or other things can be a real advantage. Technical and data problems from a different mindset can get you solutions, but you're never going to get to the point to try and solve that because you never pick up on those cultural cues that get you to be asked those questions. So how do you level that playing field? How do you, because in an interview, I've never met that person before; I don't know what they need; they don't know what the culture is, so how do I, how do you broker that to say to adapt to a level where the culture still works but things do? And that starts getting really interesting because it does start toying with lots of different laws, lots of different practices. So you've got to be really open, and I, and sometimes interviews and best practices of interviews, stop you scrolling down to those routes to find out what a candidate's needed because you've got to ask all candidates the same questions. You've got to be able to demonstrate it. You've got to be able to document why this is and people It's not what people are like all the time, so it can be difficult to skirt around to find it. Why are you not answering that question the way I'd expect you to answer it? We've all had interviews where you're interviewing someone and they're 80 percent of the way there, and it's a pretty damn good answer, but they're just missing that crucial last bit, and you just want to say, Why aren't you just saying this last bit? I know you can do it, but you haven't said this last bit that would just tick it over, and you can't give them that because in an interview you're supposed to be testing them, and yeah, they're not saying it, so how can you nudge them in that direction to give them the opportunity to do it? Are there things about the way that their brain works that are going to stop them from doing that? and in a real situation where you've built up a relationship with them over three months or something like that, and then you'd know Oh, the reason you're not getting that you can just say Hey Bob, this is what you need to think about here: next time you do that, do this. You can't do that in an interview. So how do you understand if you're missing really good candidates? I think that's the point on that.

TIM: Yeah, it's so tricky, and I feel like maybe part of the solution would be to unpack these cultural values, these slightly subjective things that the company is hiring for, unpack them into more detail, because if they're vague or open to interpretation, I feel like that's where the problems lie. So as an example, even something as soft as skills also falls into this category. So let's say communication So you might say, Okay, we want to hire a data analyst with good communication skills. Everyone probably in the world would have that as a criterion for the role. But what is good communication? You could have a candidate have an interview, and they have a very direct style of communicating, like they're very precise and concise, almost like in some cultures it would come across as slightly rude, so then depending on where the interview is from, they'd either interpret it as, Oh geez, like I can't handle that person. or great, they just answered quickly, concisely, and got straight to the point. I love that, and yeah, having done sales with companies across the whole world, it's very different dealing with the average American or Australian. Americans are a lot more direct on average; it's a broad generalization; they just get to the point. They'll cut you off mid-sentence if you're going on, whereas an Australian will maybe use more roundabout terms, use more fluff, okay, and so without really having that clear definition of the get-go around what good communication skills are, I feel like it's just open to interpretation, similar to some of the other cultural values.

DOUGLAS: Yeah, it is a good point, and communications are, I think, a really good thing to draw it out because every organization is already a cultural mill, right? We're already full of different people. If you're multinational, you've already got Americans and Saudi Arabians and Israelis and all the nations that even for a UK-based company, we still got lots of different viewpoints of those. and each of those needs different communication styles, so if your data team only has the I'm going to write a report here; it's done; tested; tick box done style, that might work for you to say some of the customers, but some of the others might leave. I actually need this graph this way; I need it balanced out here. I don't want to be on that page, and there's more two-way communication. That does it, so I think maybe that is where that cultural balance fits, and it is about where you fall down if you ask for the same people all the time, so you've got to think about who is on my team and how does this person compliment it. There was something else you were saying about that as well, but it's gone from me now, but we were talking about

TIM: And communication skills and interpretation of what good or bad communication is

DOUGLAS: What's good, and that's what so what what the question also drives me to think about is how do you approach a problem because you're right; the communication style might be actually the end result is I'm going to give you an email, I'm going to document it all, and it's going to be fully one It's all proofed, so I don't need to do much there, or it can be the more fluffy side, and both work in different styles, but what I'm trying to get from the candidates is what is their approach to that? Are they consciously thinking about how I need to communicate? Are they thinking about this is how I work, and therefore this is where it worked well and this is where it doesn't work well? So I need to work on this bit; I need to adapt this bit, but actually this is where I'm hot on it, so I guess part of the recruitment process is about looking at where that fits in and thinking about how the candidates approach it rather than a black-and-white answer, like when you're doing the technical skills, there is a right way of doing X, Y, and Z. When you're doing cultural things, there are many different ways to get to the same answer as long as they can be honest and open, and I can see how they approach it. I think that's a good starting point for me.

TIM: If you think back to candidates that you've interviewed who have fallen down at, let's say, the cultural step or the cultural aspects, not the technical aspects, can you think of the most common reasons why or common patterns where candidates would have fallen down at that stage?

DOUGLAS: Yeah, I guess it's hard to put it in those terms, but the most common thing is lack of openness. There's something that hasn't come through, and I don't trust the candidate's answers. It sounds like they are paper-based answers they've read in a book, and they're just telling me there's no proven evidence or real depth to their answers. and when you're talking about culture, how you approach things, it's really hard; it's really easy to miss that entirely and not have any depth or understanding, so I think that kind of if you go to interview closed for me, that doesn't allow you in; it doesn't. I'm not interviewing you; I could interview anyone. So that, for me, is where it falls down quite often, and then there's the other bit around which is really subjective, which is the field that you're just not getting it; you're just not getting what this role is all about. You're really good at all of this, but it's not coming across that you really understand what you've got to achieve in this role. and the reason I say it's really subjective is because whose responsibility is it to make sure that you understand what the role is? Is it you as the interviewer or me as the one who is? See, and ultimately, at an interview, you have got to prove you can do the job that bit. Not understanding the function of the role can be where it falls down a lot.

TIM: The first bit you mentioned is that I'm seeing a bit of a theme here that you're really looking for a level of authenticity because you're trying to be really honest and transparent with candidates in terms of what the role entails, and you almost expect the same from them. Jeff, I personally think, is the right way to go because, God, how much time would we all save in hiring if we were just as brutally honest as possible? none of this Oh, we get to the office stage. By the way, here's the offer: it's 20K less than what you're expecting. That's just wasting everyone's time, so the sooner we can get all these things up front, I feel like the better it is for everyone.

DOUGLAS: Yeah, and I'm not exactly buying. I am buying technical skills, but I'm not buying those. I'm buying an outcome of people wanting better housing and buying. Actually, we've got to improve residence, so even though it's technical skills, that's what it's for. So if you don't buy into that, if you don't on a basic level want to provide good customer services, even if you're never going to talk to a customer, if that's not what drives you, it's not about making the world better. Then we're not the organization if you want to go home and drive a Ferrari. Go and work for an investment bank; start your own company. Really push that. That's great, but you're not going to get a Ferrari working for Southern Housing because it just wouldn't be right when we've got people who are working in social housing and we're paying you enough to pay for a Ferrari, right? It doesn't mean that balance isn't there culturally; those people aren't going to be there, and if there was, I see you. I like bankers bonuses and all of that, and a few years ago I saw there was one that was getting multimillion-pound bonuses, and I thought, God, if I got that much money in an annual bonus, I'd quit the next day, and I'd go and start a charity. And that made me really bad; it was all about authenticity that made me really think that actually I'm never going to get that scale of bonus because I wouldn't take me that far before I quit and did something with social value, so it's intrinsic to the role you get that would not be the right role for me. and it's not that there's anything wrong with that role, but for me, that's not where I'd be, and again, equally for social housing, if you're not driven by passion, it's not going to be the role for you because you're going to keep coming across those things where you really need to do something for individuals rather than for the cutting-edge technology of things for that So if the organization culture doesn't fit, I think it's as bad for the candidate.

TIM: Yeah, so again, there's that self-selecting process. As long as you're upfront as soon as possible with the candidates, you make it clear the culture, the role, the tooling, everything like that, then that's for the best. I would be remiss of me not to recount my first two attempts at hiring people this 10 years ago now. So the first two people I've ever hired as a hiring manager both quit within one week. It's about as sketchy an outcome as you could ever hope for, and I swear to God it was not because I was the boss from hell. I don't think, but we were introducing them into an Excel hell data environment where everything was just in spreadsheets with macros, and just, oh my God. like I imagine no business intelligence tooling or warehouse at all, just a mad rush to somehow combine a bunch of spreadsheets every month, and once they realize what they're in for, they were on their bikes, but I think the takeaway was I probably over-pitched the role to them. We said, Oh, we got a lot of interesting data. It's exciting or whatever. We didn't say, You know what? The first three months are going to be hellish. You just have to automate away this shit, and then we can start doing something. interesting if I'd sold it like that, then they wouldn't have been surprised on day one, but they were probably surprised on day one and then horrified on day two, and on day three they were out the door. So definitely good to be upfront and open.

DOUGLAS: Exactly. I've done it. I've hired for like roles in an agile environment where they're agile, and we've got people in to work through that because they understand agile; they understand where it works. If I brought that same person in and then said, Here's a Prince2 exam; go do it, how would they react to that? and the same way there are people, believe it or not, that love sorting out messy, dirty data and will come in and get really into the weeds, and their reward comes from that. There are other people that want the clean data to do the innovation to move on. I think that maybe that's the cultural fit when we're that line we're trying to pull the thread of what it is that means it's that alignment maybe more than cultural fitness. So this is what that role is. It is going to be all these spreadsheets from hell, and if you can get them from spreadsheets to hell to being spreadsheets for Manchester, then you're fine. Apologies, people from Manchester, but I'm picking on Manchester today.

TIM: Now, speaking of cultural fit and alignment, I'm not sure AI can help with this problem any better than humans can. I feel like there's probably lower-hanging fruit for AI to hack away at in the hiring process than trying to do something that's almost inherently subjective and very human, like cultural fit. But broadly speaking, with AI, have you tried to double at all in using AI in hiring? You mentioned candidates probably using it as part of the interview prep or during the interview, so imagine you've seen a little bit from the candidate side.

DOUGLAS: Yeah, I don't know. We don't use AI. I don't use AI. I think we're not in an industry where you get far too many candidates. It's quite easy to pull out the ones that have got substance in the shortlisting stage to ones that you want to interview, so I don't think it actually adds too much value. I think an AI is any automation process, so there are different sorts of AI, obviously, but any automation AI is only going to get you to a point you know you want to get to, so it can be very good at saying, Have you got the technical skills? Have you got this? But actually we've already got tests; we've had tests around testing technical capability for decades. So they're just better at that. You're right; the cultural stuff It's got to know what I'm looking for, and it's got to know how to get that answer out, so if you're going to be interviewed by an AI robot that's got neuro-linguistic programming, can do sentiment analysis, can do whether you're happy or sad, logic, and all of that, then it can do it. but that in itself says this is what the organization's about, so you're going to come into a joyless organization where we expect you to talk to a computer all your life and might like people, but that's what the end result is going to be, and it's going to get that culture because the candidate experience is going to be I didn't actually talk to a person until five interviews in, and then when they did talk to me, they just ticked the box. So I think you're right. It's not so much that the AI couldn't do it; it's that it drives its own culture, and is that what you want?

TIM: Yeah, it's such a tricky one because, yeah, at the moment, if I went for a role and got interviewed by an AI robot, I'd probably not find that too enjoyable. It would seem dismissive, a little bit patronizing in some ways, like I'm not sure I'd enjoy it, but is that just a cultural thing? If we go five years down the track and now 99 percent of at least early-stage interviews are with an AI robot, then I'm just accustomed to it. Of course they'll get better; they'll get more accurate and less stupid and what have you, so is it just a matter of time, though?

DOUGLAS: I mean, I hope not because I think there's something intrinsically human about the human endeavor of that of a purpose, so I don't think I could because if AI is good enough to be better at interviewing than not to blow my own trumpet, then it also follows that it's probably good enough to do the job itself. So why am I interviewing for that person, and then so everyone The belief of AI is that everything goes up so AI can do this bit so I can now do higher quality work, and if I'm doing higher quality work, I need higher quality interviews, and I need higher quality ways of doing it, and the AI needs to keep it up. So you're on this ladder, and I think the human endeavor would always be on the cusp of the wave beyond that is my hope, because, like I say, it'd be a joyless organization if, because then you fall into if you've been brought into the organization pretty much by a robot, your HR processes are managed by robots, and all of your processes are managed by robots. What's the point of what's my incentive for doing the work for that organization? It's not that I'm not making the world a better place, and that's why I think fundamentally that's why AI will only improve us, not replace us, because it's all allowed us to do better, and it's the same thing, so I really hope not is the answer.

TIM: I have to say I've recently become quite a techno-optimist and an AR optimist as the large language models have improved. One area I think it could make a big improvement in, at least in theory, is fairness, so there's really interesting studies from different countries, one from the University of Sydney a few years ago that I could share. So basically they got tens of thousands of different CVs and then put them into three buckets, so the first bucket of CVs had an Anglo first and last name, and the second bucket of CVs had an Anglo first name and a Chinese last name. third bucket had Chinese first and last names other than that, the three sets were similar So they controlled for all the other variables, and then went and applied for tens of thousands of different jobs in Sydney, Melbourne, and other cities, with different seniorities of roles and in different industries and different domains. They then measured the rate at which those applications got a callback or not. And so across those three groups, the callback rates were 12, 8, and 4 percent, respectively, so in other words, if you applied for a job in Australia with a Chinese name, you had one third the chance of a callback as if you did with an Anglo name, all else equal, which is unfair to put it mildly or just racist to put it another way. and I feel like a well-programmed AI that did the CV screening step well, and I think there are some ways to do it. Surely it can't be more unfair than that. You know what I mean; is there not a step forward there, do you think?

DOUGLAS: I think, yeah, I think it can be really good at pointing out biases because we all have them, and we all, whether they are social class or whatever it is, we all have them built in. The way that we like to document code is even a bias; there are all these different ways of working, so anything that can strip that out and present you with that bias of saying This CV looks very similar to this CV, but you shortlisted this one and not this one. Take another look at this one. I think it'd be really good, and I say yes, I absolutely agree with that. Yeah, I think fairness is that creating a level playing field is yes, there's a real advantage to it.

TIM: The other angle where I think there's a bit of a quick win is in feedback at the moment. For a typical process, if you got rejected at the CV stage, you get at best what I like to call a sorry, not sorry rejection email, so the generic Hey, you didn't make our cut, blah blah blah. At least if you get some closure, that's helpful. Like some companies don't even send that, but let's say you get a no, but you don't really know why; that's the real challenge. You don't know why your CV has been rejected, and it would be very unscalable for any company to provide customized feedback as to why your CV was rejected or not, so that's why they don't do it. but if the larger language model was doing that evaluation themselves and they were doing the grading and ranking and scoring, they could easily pass some of that feedback on to the

DOUGLAS: Yeah, I think you're on a bit of a self-fulfilling prophecy there that you've got a candidate needing to fit in this form in this structure to get past a large language model. The large language model instruction, whatever it is, it's not judging that person on that person; it's judging that on the world of other stuff that's been successful in this process. So there are exact right words that this person can fill in, and it's not filling them in right, so the language model says these people get through; these ones don't, even if it just recommends it, and me as a recruiter then say, Actually, no, move this one up; move this one down. The end result has been evaluated by that. and then you feed that back to the language model and say, Go and tell these candidates why they didn't get that job, and it goes, Oh yes, because on the internet, people that say red rainbow Never get this job, whatever it is, right, and then so the next round they know they apply this person rather than being authentic, just knows to get past the large language model they've got to just not use the term red rainbow. So they're so then they're now using large language not to demonstrate their own skills but to get through a gate, so they're changing who they are, they're changing what they're writing, they're changing about who they are just to get through the gate, and everyone's got to do that because if they don't get through that gate, they can't get the feedback; they can't get through. So it becomes self-assertive. There's the old adage of in a theater, if you're sat at the front of the theater and you stand up, or you're sat, I'm quite tall; I'm six foot three, so the person sat behind me in the theater can't see over me. Sometimes if they stand up, they get a better view, but then the person behind them can't see. So in order for them to get a better view, they also have to stand up. Suddenly the whole theater's standing up, and everyone's uncomfortable. I think that's the same thing with Relying on large language models to tell candidates why they haven't got the job that large language models have decided It makes it unauthentic, and it makes it dry.

TIM: Yes, I agree. If that is the method that developers of such tools took, and I think Google is famous for the Google or Amazon one of them, it famously had some automated machine learning-based processes about 10 years ago looking for similar people who had been hired in the past, and it was just perpetuating the existing bias. However, I think the way the tools that we built would be in a slightly different way, which would be to say it's a CV one pass first to chat GPT or whatever model to then extract the skills, extract the experience, extract whatever is considered the information, not the noise, and then just match it against the job description to come up with some kind of score. so that it would basically be doing what a human should be doing or at least what a human says they're doing but in a very organized, objective way so then it would be feeding back, You've got a match score of 70%. Here's the strongest match is the weakest match. You were in the top 20% but not in the top 5%. Therefore, that's why you didn't get a callback. I assume that's how the tools will be built out, which then I assume would solve that problem. What do you think?

DOUGLAS: The feedback is still going to be B: You didn't get a callback because you didn't justify this skill and this skill.

TIM: And they'll just add it, you're saying basically.

DOUGLAS: Yeah, and then, and that's still because a machine has done it, and because it is not based on that, it is based on the large language that machine has access to. Yeah, it's going to be black and white at the end of the day, and I know large language models are when they're presenting text; it comes out slightly different, but essentially it is to get a score of seven; you need to say X, Y, and Z in some form or other, so then they need to just come around that cycle. and I guess the truth of it is that no, you're right, that's no different from interviewing and recruitment processes. It is because if the only difference is we don't give the feedback to let the loop carry on, so it's counterintuitive, but I do think I do worry that it actually just makes it harder and makes it more likely that the right person, because how does a large language model know that a novel response is going to get a better result because large language only looks historically So if you come up with, actually, I'm going to take that challenge. I'm going to do it entirely differently than you've never thought of before, but it's going to work. That's never going to get through a large language model because that's never proved to work before, so why would it work this time? So you can only do what you've already done, and that's my fear of that feedback loop. From there I think you're right. The fairness approach, like HR professionals will always say, is to ask the same questions based on competence, look for evidence, and everything like that, and again, that makes it really rich, and large language models can make that rigidity and the same approach really good. But what about the difference? Then we come right back to where we started about diversity and things like that, and like I say, neurodiversity, particularly if you do things in a different way, maybe only 7 percent of people have been successful in this role by taking the approach of doing things in a different way. A large language model would say no, they're not the chance of success for that 7%, but if that 7% also goes with a match to the culture, then that 7% can be your key differentiator of Wow, you've changed everything; there's loads of candidates come into roles; most of them are pretty good at the job. The ones that change an organization are the ones that really jump in and get it and do things beyond the role and will come in and say, Oh, that's great; I'll do that, and now I'll do X, Y, and Z, and in five years time they're not doing the job you brought them in for; they're doing a different job, and they've been promoted, and they've gone through, and everyone's working around different ways. If you sanitize it so much, I think you risk losing those game changers, and you need some game changers in every organization.

TIM: Yeah, you're right, and if AI, as you say, over-sanitizes it, you go down that path. I've heard also that as a general criticism for overly structured interviews with humans, if you just ask the same questions you're just looking at, you miss the outliers almost by definition because you're just forcing them down this track asking these specific questions. And yeah, that certainly seems productive, but what about thinking bigger picture? Now, if you had, like, the proverbial magic wand, be it AI or some other kind of magic, how would you be trying to fix the hiring process?

DOUGLAS: I think there are a lot of things that I think differently. I think I question how good interviews are; I question how good shortlists are. I think they all have their place because they get you through, but here I've taken out interviews and shortlisting pretty much all the recruitment process. I question if it really serves a purpose. So I think what I'd do if I could snap my fingers would be to create some sort of flexibility that changes on its head so it becomes more I don't know, it's like for me, for speed dating, something like that, where you have a conversation of I haven't got a rigid role, I've got a problem I want to solve. How do you solve that problem? You get creative thinkers; you get those sorts of things, and then you work out the rest of it from there, but I'm conscious that if I've taken all of that structure out, you can get some really crazy outcomes, so it's balanced, but I do think we get too far down the road of everything having to be really structured. So I get this result, and I can prove it, and there's no risk in this because risk is what drives business.

TIM: On the spectrum of intuitive to data-driven, how would you characterize your own hiring philosophy? Where would you fit on that spectrum, do you think?

DOUGLAS: I would say my HR colleagues are killing me. I would say I'm much more intuitive. I think you've got to that; that is where it is, and I would like, as I said, a data professional background and a technology background. I actually value a lot of process and a lot of structure in the way I work. but I don't think that's what humans add, so in my interview approach, it's intuitive; the structure should take care of itself; the proof should take care of itself, but the decision, the understanding, for me, it's people, things that matter, and yeah, so that's where it is. I think for me, much more than I like what I was going to say. So I have ADHD, and if I had an organization of people working, we all had ADHD; it'd be a crazy place, so you need to balance; you need the structure, and you need people who will do things differently. For me, recognizing that yes, I'm an intuitive interviewer means that when I interview people, I've also got to bring in people that balance that so I don't go off on a tangent and hire willy-nilly, that I don't hire a fishmonger for a football club. Coming back to that one So I've got to have someone in the interview panel, and I think that's why you need a panel. It can't just be my decision, even if I'm the recruiting manager. Having someone there that's different from me in every way that can say Hey, No Doug, you've just gone down a rabbit hole there, and have you really thought about how that's going to think, and having the honesty with other panels that they can challenge and they can push is the Nirvana of it, then maybe

TIM: It's a nice tie-up then to you having that diversity of thought and diversity of skillset also in the hiring process to ultimately make a better hiring decision. As a final question, Douglas, if you could ask our next guest one question about anything in the world, what would that question be?

DOUGLAS: Do you balance diversity and cultural fit? Let's go right back to the start, right?

TIM: Yeah, that's a good one, and I don't think I have the answer, that's for sure, but I will level that question at our next guest maybe later this week, so I'm looking forward to hearing what they say, and I'll share that with you. I'll share their response. Douglas, it's been a great conversation today. We've covered a lot of different topics. It's been very relaxed and informative, and I'm sure our audience has really enjoyed it. Thank you so much for joining us.

DOUGLAS: Thank you. It's been a very enjoyable day. Thank you.