Interviews have always been the central tenant of the hiring process and the one area where it has been most challenging to collect meaningful data. The mass move to video interviewing in the last 12 months has potentially changed everything. So, what data can we now gather, and how can it be analysed to drive actionable improvement in the interview process.
My guest this week is Siadhal Magos, the Co-Founder and CEO of Metaview. Metaview is a technology start-up that is helping some well know employers gather data and act on interview intelligence in new and innovative ways.
In the interview, we discuss:
▪ Using data to make sense of interviews at scale
▪ Two techniques to improve interviews
▪ Will technology replace traditional interviews?
▪ What are the critical interview data points?
▪ The correlation between hiring funnel efficiency and interview rigour
▪ Bias and inclusivity
▪ Addressing privacy concerns
Listen to this podcast in Apple Podcasts.
Transcript:
Matt Alder [00:00:00]:
Support for this podcast comes from Metaview. Metaview is an interview intelligence platform that helps growing forward thinking organizations run amazing interviews. Interviews are the most important part of the hiring process. Unfortunately, they are also the least reliable. Metaview fixes this. Metaview uses cutting edge technology to provide talent acquisition leaders with insight about what’s happening in the interview process and provides interviewers with automatic transcripts and personalized tips on how to improve their interview technique. All this means you can conduct better interviews with fairer, more reliable outcomes. Want to learn more about how Metaview can help you conduct amazing interviews? Visit Metaview AI now.
Matt Alder [00:01:05]:
Hi everyone, this is Matt Alder. Welcome to episode 337 of the Recruiting Future podcast. Interviews have always been the central tenant of the hiring process and the one area where it’s been most challenging to collect meaningful data. The mass move to video interviewing in the last 12 months has potentially changed everything. So what data can we now gather and how can it be analyzed to drive actionable improvement in the interview process? My guest this week is Siadhal Magos, the co founder and CEO of Metaview. Metaview is a technology startup that’s helping some well known employers gather data and act on interview intelligence in new and innovative ways. Hi Siadhal and welcome to the podcast.
Siadhal Magos [00:01:56]:
Hey Matt, how are you doing?
Matt Alder [00:01:58]:
An absolute pleasure to have you on the show. Could you just introd introduce yourself and tell us what you do?
Siadhal Magos [00:02:02]:
Sure. So I’m Siadhal, I’m one of the co founders and the CEO at a startup called Metaview. Metaview helps companies run amazing interviews. We use some pretty bleeding edge and some pretty cutting edge technology to help them do that, take a really innovative approach to that. And yeah, we’re super obsessed with the value and the importance of interviews in helping you build great teams and that’s where I spend most of my time.
Matt Alder [00:02:27]:
Fantastic stuff. And we’ll dive into interviews and the importance of interviews as we go through the conversation, but just give us a little bit of backstory about how the company was started.
Siadhal Magos [00:02:37]:
Sure. So going a little bit back, I’ve been working for and in technology for a little over a decade, usually for high growth tech companies. So most recently before I started Metaview I was on the product team at Uber. So don’t have a recruitment background per se, nor does the rest of the founding team. We are technologists and product folk, but what we were exposed to was just an Incredible amount of hiring. I was at Uber at a time where it grew from 7,000 to 17,000 people in about 18 months. So it was everyone’s job to be part of recruiting. And I ended up getting really nerdy, essentially about the importance of hiring the best people if you’re going to achieve your objectives. So even though, yeah, as I say, I really spent most of my time on product challenges, I became increasingly obsessed with hiring. And essentially from these experiences at Uber, where, again, we were, we were hiring so much, came away with a bit of an observation, which is that it really felt to me as a, as a hiring manager, that the place where the decisions were getting made, the places where the, frankly, the quality of my team was being determined was in these interviews. And I also had a sense that it was one of the most unreliable parts of the hiring process because I didn’t really know if other folks on the hiring team were great at interviewing or sometimes there’d be sort of sideways glances between the various members of the interviewing team at each other because they came away with a different understanding out of an interview. It was unclear who was asking what and whether they were getting the right signal. And so this combination of really being exposed to the importance of great interviewing at the same time seeing how unreliable they were, got me thinking around how technology could help solve that problem. And that’s really, I guess, the kernel that started Metaview is. Is technology now at a point where we solve that problem, and we think the answer is yes, which is of course, why we’re building the company. And the answer is yes for two reasons. One is speech to text. So natural language processing technology is now pretty darn good. So you can make sense of conversations at essentially infinite scale, which is something that Metaview leverages. So we record and transcribe interviews in order to build a lot of applications to help people conduct better interviews on top of. And the second thing is that many conversations now, especially interviews in this, in this, well, you know, current, you know, I wouldn’t call it a post pandemic world just yet, but during the pandemic right now, a lot of conversations are being had into microphones, you know, which is really fundamental shift, right, that, that, that now you’re in a world where the majority of business conversations and certainly the majority of interviews are already being had into a microphone, which means you have this, this, this native ability to capture them. So what we realized was with NLP being pretty darn good, and with a lot of conversations being had into microphones already, there is this Opportunity to start to make sense of interviews at scale and empower both hiring managers, but also talent leaders with data about their interviews which they couldn’t have access to before. Obviously we’d love to speak more about some specific use cases of that data or specific insights that we realized, but that for us was the starting point. Interviews, super important, super unreliable, though no one has any data about what’s happening in them. Can we change that? And that, that’s what we work on.
Matt Alder [00:06:12]:
That makes perfect sense. And there’s, there’s lots of things I want to ask you about that. Before we do though, it’s probably worth just talking a little bit more about interviews. As you say, quality of interviews in, in recruiting, you know, particularly in corporate recruiting, has always been something that, that people have talked about and tried to improve. The last 12 months, everything’s gone to video, which is really shaken things up. And I’ve noticed in the, in the last few weeks that there’s a, there’s a lot of conversation and debate going on about interviews and what makes for a great interview process and how you train people to do interviews. I actually wrote an article a couple of weeks ago about quality versus its quantity of interviews. So it’s a, it’s a topic that is very much under discussion at the moment. What makes for an amazing interview in your view, what is it that employers should be looking for and executing on?
Siadhal Magos [00:07:02]:
Yeah, really interesting question. And as you can imagine, it’s something that we’ve sort of been through the idea maze on because obviously throughout my professional life been part of conducting interviews. And usually when you get into it, it’s just part of the process. Right? But now sort of having thought really deeply about it over the last three years or so, since we started Metaview, there are two fundamental goals of an interview process. And these two goals are actually to some extent in tension with each other. Not completely, but they can be. One is to identify great talent, so find people that are going to do great work with your team within your organization. And the second two is attract, is to attract that talent. So make sure that when you interview them by the end of the process, they actually want to join your organization. So again, not completely intention, but there can be times where they’re competing with each other. For example, if you rigorously interrogated every candidate that came through, never told them anything about your organization, to push it to an extreme, you very highly likely identify whether they’re high quality or not, but very low likelihood they want to join your company. So you’d fail on the attraction side. So you have these two competing partners identify and attract. Now the reason interviews are so important is because they are the most signal dense, the most data dense part of any interaction, part of the hiring process. They are where 95% of all the data you ever sort of get exposed to on a candidate exists. It just so happens that for the last however many hundred years, hardly any data is being captured from those interviews. But that doesn’t change the fact that it’s where all the data lives because they’re incredibly, as I say, rich rich conversations. What’s more, of course they are fundamentally where you as an organization or you as a hiring manager make your decision. Because that’s so signal rich. You know, whenever you’re part of a a hiring decision or a hiring debrief or a wash up, whatever you may call it in your organization, it’s always the content that you uncover during interviews that is under discussion when you’re making your hiring decision. Which says to me, really the interviews are the most outcome defining point of the hiring process. And the same is true on the flip side, if you’re a hot candidate who can work for Facebook or they can work for Google or a startup, you’re really going to make your decision based on your experiences during that interview. So, you know, the war for talent is a bit of an overused cliche phrase. But there is, you know, there is this, for want of a better term resource that we’re all competing for, which is great talent and they have a lot of power at the moment. And so making sure you have an attractive process is key as well. So just from a high level, that’s obviously what you’re after identifying and attracting and really interviews are where both of those things happen in the hiring process. In terms of what great looks like for organizations, we see a couple of things. One thing that we see work really well is it’s very simple. But you’d be amazed at how few organizations do it effectively is truly competency based interviewing. Now the reason this is harder than it sounds. And by competency based interview, I mean having specific competencies in mind when you’re interviewing a candidate for each stage that you’re interviewing them. So very easy to make a plan for this sort of thing. The reason actually many companies, we find many companies don’t do it is because truly understanding what you’re hiring and who you’re hiring, you know, what sort of competencies you’re looking for, is actually part of the challenge because it’s quite hard to refine it down to those sort of specific five or six competencies. So, but being really rigorous and really, really specific in what you’re looking for at each stage of interview is something that works really well. And one of the sort of the telltale signs that a company is doing that well is often the discussion that they have after each interview stage is not a hire or no hire discussion. It’s a does this person have this competency yes or no discussion. And then you leave it to the hiring committee or hiring manager to decide whether given they do or don’t have that competency in aggregate, are they a higher or no hire. Where you see companies that have less effective interview processes is where every stage is a hire, no hire decision. And even just making that verbal change in how people talk about their interviews, we’ve seen it be really powerful in changing hiring outcomes. Other things we’ve seen work really well in making sure you’re identifying and attracting consistently is having obviously very thoughtful training processes. The best training process that we’ve seen outside of Metaview and something that actually metaview enables as well is shadowing. So that was the de facto way that we trained as interviewers at Uber, for example. Some of my friends who work at companies like Amazon and Palantir, same thing. Shadowing is the way that you learn. The reason shadowing is so powerful is because it’s highly credible tuition. Now by shadowing, what I mean is as a new interviewer, when you’re onboarding new interviewers to your organization, maybe they’ve been there for three months now they’re going to become an interviewer. The first thing they would do is sit in on the interviews of some of your top interviewers. So they literally shadow them. That’s where the phrase comes from. They sit in on the interview and listen to the interview. It’s really just a listening brief. And then they might flip roles and in the next interview they’ll conduct the interview and the experience interview will be there to shadow them and give them feedback afterwards. Now the reason this works really well is because rather than sort of classroom based broadcast tuition on how to conduct an interview, you’re actually seeing what a well honed interviewer within your organization who is looking for the attributes you’re looking for looks like. And you know, it’s the best way to learn from this is from examples. And on the flip side, when you conduct it yourself, you’re getting feedback from someone who you know is great at this task and who you know understands your interview style in detail rather than Again, just sort of broadcast generalities about what great interviewing. So I think those are probably the two. You know, if I was at an organization that was struggling with interviewing and Metaview didn’t exist, the two things that I would implement right away would be really structured competency based interviews so you can get consistent outcomes from for each candidate and a shadowing program that enabled me to make sure more of my interviewers were conducting great interviews and I was burning out my top interviewers less than I would be otherwise.
Matt Alder [00:13:37]:
So as we’ve already mentioned, it’s a very, it’s a time of fast disruption in terms of technology and recruiting processes. And one of the big topics that we talk about on the podcast is recruiting automation. And a lot of the conversations that I’ve been having around the sort of, at least the very, you know, the start of the interview process and how traditional interviews can be replaced by automation and AI, isn’t that a trend that’s set to continue? Are we still going to be interviewing in the same way in the future?
Siadhal Magos [00:14:14]:
Yeah. Again, another thing that we think loads about and automation is AI and automation is a disruptive force. And so it’s definitely the right question to be asking oneself in any industry. Is this something that would actually be better served as an automated process? The fact is that is not going to apply to knowledge worker creative collaborative jobs where the power sort of rests with the, with the employee, as it were, with the candidate. If you’re, if you’re a really strong candidate, you are going to want to know the folks that you’re going to be rubbing shoulders with, the folks that you’re going to be having sort of rapid fire slack exchanges with or you know, multiple zoom calls per day with to collaborate and towards a creative outcome, it’s just not realistic that AI is the right solution for that. Now there are elements of this that I can help with which Metaview does look into. Like, is there ways that we can use AI to coach interviewers to improve? Because that is, there is slightly more. There are, there are ways that you can predict what great interviewing should look like, but to actually predict who will be a great fit for your company when you’re a fast growth, fast moving, creative organization. No time soon are we going to be putting our faith in AI to make those decisions. But I’ll concede that it’s different. And in fact, obviously we’re very much focused on, as I say, knowledge work companies. But if you’re talking about a filtering problem, so a scenario where you have, say thousands of applications and every single one of those applications went through the same process and is in a very structured format, then, yes, getting from thousands of applications down from hundreds for you to review in more detail. That’s a great use case for artificial intelligence because you’re essentially automating a rote task that one of your team would have to do, or more likely, that you wouldn’t bother doing. You know, if you can move from a world where, well, I’ve got a thousand applications, so I’m only going to look at the ones who have a degree from Oxford or Cambridge because there’s still enough to choose from there, then using automation to actually look at that in a bit of higher fidelity is a good thing. But where we are in the funnel, which is more towards the decision end of who do I actually want to work with, who do I want to spend my time with? The human touch is vital and we believe that will be the case. So AI for the application here, the application for AI in that decision level is less to do with actually making the decision on the candidate, much more to do with giving the interviewers or giving the people involved superpowers. So one of the things that people sometimes refer to is rather than thinking about this as artificial intelligence, think of this as augmented intelligence. We are augmenting the folks on the front lines that are doing the work to do an amazing job. And that includes interviewers, and that includes talent leaders who are making decisions about the interview process that they own. We’re looking to empower them with data and augment their intelligence rather than replace them, because it’s not the way that is going to result in the best outcomes for them.
Matt Alder [00:17:36]:
One of the things about any discussion about interviews, whether it’s the sort of discussion that we’re having now, some of the debates that go on in the recruiting community or within organizations where they’re trying to improve the way they interview. One of the problems in the past is the conversation has, has by necessity been based on sort of anecdotal and subjective sort of views on, on the interviewing process and the, and the quality of it. Obviously you’re collecting data on all of, all of this. And I’d be really, I’m really interested to find out what kind of things are you learning from the data that you capture.
Siadhal Magos [00:18:14]:
Yeah, absolutely. And just, just before I dive into that, just let me give two minutes on what Metaview does and how it works, because this will explain a lot about how we have this data. Otherwise it could sound a little bit like it’ll Be a little bit unclear how we have this data. So as I mentioned, we focus a lot on helping companies run amazing interviews. We leverage speech to text technology to help them do that. What that means is we have a product that records and transcribe interviews from that transcript. We then pull out a number of data points. There’s too many to name, but to give an idea, we pull out things like the percentage of time that the interviewer is speaking as opposed to the candidate, the number of questions the interviewer asks during the interviews, the number of those questions that were open ended questions, the number of interruptions during the interview. Did the interview start on time, did it end early or run over? So you can imagine there’s all these characteristics of an interview that once you are applying what we call interview intelligence to your interviews, you can start to understand. So your question about what data points are most compelling is a great one. And there are two buckets for I guess, data from interviews. One bucket is what data do we use, do we see to actually help individual interviewers improve? So this is our way of providing best in class interview training and what are the factors there? And then the second is what are the aggregate level insights that you see? So just to start with the first bucket, which is what insights from what data points from interviews do we see that help us identify who is a strong interviewer and who is not? So what we’ve seen from interviewers on the platform is, is that interviewers who conduct rigorous interviews tend to be part of more efficient hiring cycles. So the way that we measure rigor is based on a few things such as question count and how descriptive candidates answers are to interviews and how many open ended questions are asked. What we say is if you conduct a rigorous interview, you’re part of hiring funnels that are 30% more efficient. And what I mean by that is there are 30% fewer interviews per hire when you are part of, when you conduct rigorous interviews. So again this is a great example of you say, you know a lot of, a lot of thought, thinking around interviews is very anecdotal and people have ideas around what great interviewing looks like. And you know, it’s really based on feelings about what goes well. But what we’re starting to see is the hard data that actually interviews that we can coach to become rigorous interviewers meaningfully reduce the amount of energy that is spent interviewing. Because of course you’re making better decisions at each stage. Each time you have more data on the candidate, more insight on the candidate with which to make a decision. Which means the decision is better. Which means when they go through to the next stage, they’re more likely to also be a yes at that stage, which means your whole funnel becomes more efficient. We also see that 15 to 20% of the time, depending on varies by organization. That’s on our platform. Interviewers change their decision based on reviewing the transcript, based on reviewing the evidence. So as part of metaview, after each interview, the interviewers and the hiring team get transcripts and recordings of the interview. They can collaborate on various moments and highlight sections that are seen as key moments of the interview. What we see is when this occurs, they change their decision 15 to 20% of the time. So what that means is in a world where you don’t have all this evidence, one in five times, you’re making a bad decision and not knowing it. So, yeah, another great example of where actually having the data makes a difference. Moving on to the more aggregate level. So what’s also really cool is once you have an interview intelligence platform at play, you can start to understand your interviews at scale. Right? Not only are you capturing these data points at each interview, but you can see what that looks like across the company. So you can know if the interviews being conducted in this department are more or less rigorous or more or less consistent than this department over here. Some of the really interesting things we’ve seen probably the most, well, the one that really got us excited and gets our customers excited too, is around consistency and fairness. So what we see across the population of companies on our platform is that female candidates, on average get 12% less time to speak during interviews than their male counterparts. So, of course, quite rightly, lots of organizations have diversity and inclusion goals. Some of those goals focus very specifically on gender of folks in their team. And one of the things that’s always been a blind spot, of course, has been, well, are we interviewing these people the same each time? Because it’s always been impossible to measure. So the place where really the decision is getting made is also the place where you have the least data. That’s no longer the case. And what transpires is that it is indeed the case that in general, female candidates are treated differently to male candidates. So, yeah, again, that metric was 12% less speaking time. Another sort of really interesting, I’m part of a few recruiter communities, and this is always something that comes up and that I sort of weigh in on is what number of roles should a recruiter be recruiting for any one time to be able to do a good job. Again, what we can focus what we can bring to the table is an understanding of how your interview style suffers or changes as you hire for more roles. What we see is that recruiters that screen for more than five roles per week. So within any given week, if they are screening candidates for more than five different roles, their interview rigor drops significantly and to the point where many of their 20, 20% of the time they’re conducting what we call low rigor interviews. So again, if you’re conducting five different screens for five different roles, you’re more likely to be less rigorous. That results in actually a much less efficient hiring funnel. So you end up putting through weak candidates to future stages and of course, therefore taking up slots, taking up people’s time with weaker candidates. So, yeah, lots and lots of this data to nerd out on. And I think the way you put it was excellent, which is this has been an area that everyone has always understood as important. People talk about it a lot. It’s been rife with anecdotal data and now we’re at a point where we can actually turn it into a bit of a science.
Matt Alder [00:25:13]:
As a final question, we haven’t sort of really talked about the privacy or almost the, the Big Brother implications of this. We’re seeing obviously, you know, everything’s on video, people are talking into people talking into microphones. Data is giving us fantastic insights that can be really used in a positive way. Are we not getting slightly scary though, with. With all work conversations now potentially being.
Siadhal Magos [00:25:40]:
Recorded, this needs to be treated thoughtfully because we don’t want to move to that world. The best way to avoid that contention is to really give the control to the participants. So I can only obviously speak for how we work, but we rely on dual party consent. Dual party consent means in a conversation with two people in it, both sides have consented to there being a recording and both sides understand why the conversation is being recorded, what the data is going to be used for. That often means that there’s essentially very little problem with recording conversations. Now, to speak to the point, the broader point, will all conversations be recorded? I think the way this will go is that the operators, the people who are in the workplace trying to achieve great things, their preferences will win. And what people will realize is because actually of some of the regulation that has occurred over the last couple of years, so gdpr, for example, lots of critics, lots of ways in which actually for some startups, like startups at our stage, it can make, it can make, you know, it’s definitely something to consider and think about, but overall for the sort of general sophistication around data privacy and compliance. It’s been a great thing. And I think as consumers or even, you know, in the, in a B2B context, as employees or as candidates, as you start to understand your data rights more and more, actually some of this recording becomes less and less scary. What I mean by that is anyone who’s part of a conversation that is recorded is a data subject in that conversation, which means they have certain rights to that data, such as, you know, the right to request it being deleted, the right to access their part of the conversation. All these sorts of things actually mean that the increased means there’s higher fidelity and higher levels of control that participants have in their data, which I think should, should well end up meaning that yes, more and more of our conversations are recorded because people understand their rights over this and actually they realize, well, it’s in my benefit 99% of the time. It’s actually completely beneficial for this to be recorded because it saves me right now shed loads of notes, which is a real pain for an interviewer is to write down so many notes or it makes sense for me to have this captured automatically because I don’t fully trust my memory to be able to make a non biased decision come the end of this. So there are lots of reasons why it’s in people’s interests and the real hurdle to get over is making people understand that they’re still in control and that these platforms really are just there to help. So yeah, it’s definitely an important talking point. It’s one that operators or players like us in this space need to manage proactively and be on the right side of the argument, as it were. We feel pretty confident for as long as we’re getting great reviews from interviewers saying that this has really helped them, we’re happy that it won’t be too much of a problem.
Matt Alder [00:28:50]:
Sial, thank you very much for talking to me.
Siadhal Magos [00:28:53]:
Cheers, Matt.
Matt Alder [00:28:54]:
My thanks to Siadhal. You can subscribe to this podcast in Apple Podcasts, on Spotify or via your podcasting app of choice. Please also follow the show on Instagram. You can find us by searching for recruiting future. You can search all the past episodes@recruitingfuture.com on that site. You can also subscribe to the mailing list to get the inside track about everything that’s coming up on the show. Thanks very much for listening. I’ll be back next time and I hope you’ll join me.
