Episode 56

The Future of Job Search & Enhancing Job Matching with Dylan Buckley - Founder of Directly Apply

TIM: Dylan Welcome to the Objective Hiring Podcast. Thank you so much for joining us.

DYLAN: Thanks for having me. Tim, great to be here.

TIM: It's absolutely our pleasure. You are the first founder we've had on this podcast, so I'm really excited for our audience to hear some of your insights, like directly from the coalface, because I think you have a lens that not many people would have. And so I'm really pumped to chat with you today, and maybe you could start with just a brief overview about how you started the business and what the product is all about.

DYLAN: Sure, so I'm one of the co-founders of Directly Apply. We started the business back in 2019, and Directly Apply is a job discovery platform, so we connect job seekers with employers in the United States and North America, and we do that across all sectors, so we're not sector specific. We connect job seekers from delivery drivers and cleaners up to lawyers and software engineers. We do the entire job market our big point of difference, and when we looked at the market back in 2016, we saw a real opportunity to improve job matching and connectivity between job seekers and employers. We've all used the big legacy job boards and platforms, and often they really struggle to do anything beyond sort of keyword matching. and finding a job is one of those really important decisions that we make in life; it's up there, and most people only change jobs five or six times in their life, and every time they do, and it doesn't matter if you're doing it for 14 bucks an hour working at Walmart or if you're going for half a million dollars a year. The same decision-making process goes into it: you're looking at whether it's a good role for you, whether it's a good company, what sort of lifestyle it is going to be like in terms of commute and benefits, and all these types of things, and we felt that what most job platforms really lacked was the ability to give job seekers that insight and that data to then make an informed decision, so that's what we've been building, and we're a tech company at heart; 90 percent of the company are software engineers. We are very much just focused on building out that tech and making those connections.

TIM: And how does the business model work? Like, how do you make money?

DYLAN: Yeah, so we receive we've got around five million jobs across the US hosted on our platform, and they come to us from either other job sites, they come to us from employers, they come to us from media agencies who manage budgets on behalf of large employers, and typically we're paid on a performance basis. So that could either be for a completed job application; sometimes it's even for a completed hire, or it could just be a click where we send a job seeker to the applicant tracking system to review that job, so it's all performance-based. The old model of how job boards and these sorts of platforms worked was you'd You know you pay for a monthly post and that post would stay up there for a month and what you got is what you got but particularly in North America and Europe we're moving towards a performance based model which is similar to Google ads or Facebook ads or something like that where employers are paying for the results they receive and as time's going on we're getting far more into jobs yeah into employers wanting to pay for quality and not just the volume as well when we really focus on that matching element and actually making sure we've got the right job seeker to the right job that means we've got higher quality and then employers are willing to pay us a premium for that exposure

TIM: And I'm interested in the matching piece, and so you mentioned that's like the key point of difference between yourselves and a typical old-school job board. How do you do that matching without giving away the secret sauce?

DYLAN: https://otter.ai You know what? A lot of people have tried to do this, and actually when we started this business, we were looking at trying to do something like a niche, so we're actually looking at the accounting and finance space and thinking how could we build almost like an artificial recruiter that could act in the same way that a human recruiter would in making those connections? And when you're doing it for a specific niche, you can actually drill quite deep into specifics around the job potential qualifications you might have and all that type of thing, and you can create quite a cool, quite an accurate sort of process. The challenge that we face in building directly applies; we've wanted to try and provide that recruiter-like experience or that headhunter-like experience, but we want to do it en masse with basically every role you could possibly imagine across the US. So our approach is multi-layered. It starts with having a huge, as much as it's incredibly data-driven, right? So you need huge amounts of data on the roles you're looking for at their companies, and then when we ingest all of these jobs, we try and get as much enriched data out of it as possible, and then once we've got that, we're really, I guess, an event-driven platform, so we're looking at Hey, we've got this job. Let's look at all of the people, or all of the job seekers, that have looked at this job. rejected this job, clicked apply on this job, got hired for this job. What insights can we pull from that? And how does that relate to the next job seeker we might bring on? Whenever we have a job seeker that comes in and they say, Hey, my name's Tim. I'm looking for this type of job in this area, and this is what's important to me: we have a huge number of data points that we can look at from other job seekers that look similar, other job seekers that look similar that behaved in different or similar ways, and we were able to pull all that data together and say, Hey, here are five or six jobs that might be of interest to you, and sometimes on the surface those jobs aren't what you think is a job seeker. You think you're looking for what we've got: so much data around slightly adjacent jobs, or perhaps that next step in your career that you didn't even know you were ready to take yet, and because we have all that, we can provide a lot of suggestions that just go so far beyond you typing in a job title and getting some results, and that's really the value add that we provide, and so then job seekers see that as a really valuable add-on to their job search, and they often discover companies and jobs that they wouldn't have otherwise found.

TIM: That is really interesting, and that must be quite an adjustment in the process because they're probably used to going to the more traditional job boards and going, I'm going to search for, I don't know, a head of data analytics job. Oh, this head of data analytics looks good. This one that's only a data analyst is too junior for me. I'll dismiss that, but you're saying almost you're surfacing other roles you feel like people like me are also applying for, interested in, have applied for, have been hired for, and so I might actually uncover a new career almost.

DYLAN: Yeah, it's the way I like it. It sounds a bit silly, but it does feel a bit like magic when you're using the product because you go from having to do that sort of Even if you don't realize you're doing that thinking as you're looking through different job posts, we've done a lot of it for you. And so what we tend to see is we see longer engagement because as job seekers are going through the different options that we present to them at different parts of the journey, and this can be in different mediums, right? It could be on a search results page; we could send them an email alert. we'll say Hey what do you think of this job there's lots of different places where we can surface this And what's really interesting is that we find that when we go when we put a job seeker through this sort of automated discovery process they spend a lot longer engaging and then ultimately once they find a job that they'll interested in the their intent to apply and complete that application process is far higher we while we think that lots of people try and solve problems for people using tech and so we don't want to try and replace that decision making process that a job seeker goes through we think it's really important that a job seeker looks at multiple options reviews them and then makes a decision but where we can provide real value is bringing in some ideas that might have taken them a lot longer to get to through the traditional process and speed that up for them but also when they do go to click apply they feel like Hey I've reviewed the market I've seen all the different options that are available to me I really like this job. I'm going to go and complete the application, and so we can create a more well-rounded sort of search experience.

TIM: It's really interesting you mentioned that the platforms in the United States or North America have you considered expanding to other regions. Are there any kinds of differences in the way that recruitment is done or the way that some employers act in other regions that make it challenging or potentially compelling?

DYLAN: Yeah so as a company we actually operate or we were based in the UK so we build the product here in the UK but we don't actually operate in the UK currently we only operate in North America that's for a few reasons one is just from a business perspective we'd rather do one thing focus on that rather than trying to spread ourselves too thin we're a small company we haven't raised lots of VC money or anything like that so we run the business profitably and that means we've got to pick our battles but from a recruitment perspective there's absolutely big differences across different regions and we do have experience in different regions as well. Some examples: there's obviously the easy ones, language and currencies, and different types of workplace benefits, and what we get here in the UK and Australia is very different from the US, and health insurance and all of that sort of stuff around jobs is quite different, but then there's also very big differences in workplace culture. We get a lot of people coming to directly apply that are looking for their second and third jobs, and in America it's far more normalized. For example, coming into the festive periods, people will take on additional jobs to then fund the holiday period. And that's quite a specific thing to think about doing a job search for a job that you intend to only have for three or four weeks to help you finance some of the things you want to do in your life; it is very different from looking for a job that you intend to do for the next two or three years. and we're very conscious of the different types of job seekers in different regions and different markets, and it makes the American market quite interesting compared to other markets, but where we've seen platforms not do as great a job, particularly in that matching piece, is when they try and apply their thinking about how the world works or how job search should work. and they try and do a blanket across the entire world, and you typically then you just get a watered-down product, whereas if you can create if you can really understand the market, we believe you can create a really good product, so that's been our focus: never say never in terms of expansion. and we think that, for example, your market in Australia is quite ripe for disruption, particularly on the performance marketing and programmatic piece; there's not a lot of that going on down there, and that could be interesting in the future, but I'm sure that the way Australians approach work, for example, in jobs and job search, is quite different from that of America. I don't think it's a sensible decision to do a copy-and-paste type thing. I think it's better to go deep and just create a really good product.

TIM: And what about on the candidate side of things? Where do your candidates actually come from? Are they finding you through social media, through SEO, or through SEM? Like, how do you actually get the candidates? yeah

DYLAN: Yeah exactly most job seekers find us organically they then will either do a search sign up to our job alerts get involved in the directly applied sort apply sort of product and and then we get a lot of referrals off the back of that as well we do search engine marketing we do some social stuff but what's super interesting is that we see job seekers that signed up a few years ago and found a job and then they start reengaging with us two years later as they look for their next job and what's really cool about that is obviously we've got a huge amount of data from their previous search we can connect a few dots around what they've been doing for the last couple of years and then the suggestions jump to that next point we've been growing really steadily like that and It's been good

TIM: And what about in terms of the data points you have on an individual candidate that you then use to do the matching? So, as you see, you've got all their search queries; let's say if they're logged in, you've got maybe a bit of their job history, you've got their CV, and other things. Are there other kinds of structured or unstructured data sets you collect on the candidate to then do that matching?

DYLAN: Yeah, so all the things you mentioned, and so a big thing that we released in the summer this year was a resume to job search, so where we're asking job seekers to upload their resume, and so this goes far beyond just typing in data analyst or something like that. We're able to take their resume, take their career history, take the specifics of what they did in individual jobs, take their overall qualifications, and take their skill sets. And create quite a large knowledge graph around that and then produce job matches off the back of it in terms of the less explicit data we're doing a lot of things around. We basically track every movement that happens on the platform, so we know how long someone's eyeballs have been on different job cards and different parts of the job description. If somebody spends a lot of time looking at the benefit section, for example, then we know we're able to interpret that as there's something within the benefits thing that they're looking for. And then based on the next actions we do, we can actually figure out what benefits they may be, likewise things around skills. It's interesting when you start tracking what people are looking at in the job; you can see we can actually almost figure out they probably don't have the skills that are listed here, and then it's a case of are they actually mandatory school requirements or not, and maybe we could prompt the job seeker to say, Hey, it's still good to apply for this job; if you match 80 percent of it, there's so much there, and that's been a huge amount of what we've been building over the last couple of years: how do we There are two ways you can get feedback. you can put up a form and say tell us why this doesn't match or why this does match and we do an element of that but there's also so much information that you can pull just from looking at somebody's behavior and that's where that magic element comes into it is that we're able to see that you're ignoring all the jobs where the salary is below 20 bucks an hour and you're only looking at jobs at 21 bucks an hour so then automatically we can we've got that as a structured data point we're able to stop showing those jobs show more of the ones in the higher pay bracket and then engagement increases so I would say in terms of matching at the moment I would say almost 80 percent of the data points we use is not data that's been explicitly entered by a job seeker So the 20 percent is things like resumes, job searches, that sort of thing, and then the 80 percent is their interactions with the platform.

TIM: Yeah, their behavior doesn't lie, does it? That's the beauty of product analytics, I think. You could survey something, anything you like, but they could come up with any answer, but when you look at what they actually do, that's the truth I find.

DYLAN: Absolutely, and when you take that behavioral data and then you compare it to hundreds of thousands or millions of other job seekers, then you almost get like a, it's a pretty foolproof methodology. It's fascinating how predictable humans are and how much overlap there is in how we behave and do different things. and you can draw pretty strong predictions based on that one individual's data, which by itself is useless, as in you can get some things, but the real power of it comes when you can do it at scale and then also compare those behaviors within outputs, and for us, outputs are completed job applications or hires or whatever that end result might be. somebody telling us I've got a job, and so we can almost predict from the top of the funnel right to the end based on those behaviors and then based on seeing it so many times from other job seekers

TIM: I'm wondering if you've ever done any analysis into what makes a good quote-unquote job description or job ad because you must be able to tell what people are going to apply for or not based on various factors. obviously salary and things like that are going to matter, but like I've seen so many times these kinds of rules of thumb or research that I was honestly slightly dubious of around Oh, if you include more than three dot points in the requirements, it's too long, and now no women are going to apply because they're going to be deterred from applying and these kinds of things, which I found like it seemed to exaggerate to me. It was my gut feeling, but I don't have data directly either way. Have you ever looked into some of those factors?

DYLAN: Yeah there's so much noise around this and so much so many quote unquote experts saying these are the things you've got to have and it's going to be this number of words long and you're going to have this many things and all the rest of it I think we're We see a lot of people going wrong as making assumptions about job seekers particularly job seekers in the lower paid categories or like the more blue collar workers There's this almost unspoken idea that job seekers who are working in a store or a warehouse or that end of the employment market don't care as much or aren't as invested in what they're looking at, whereas people that earn lots of money are more invested in this type of thing. and so that's just complete nonsense, and so what you ended up seeing is like a whole lot of, say, blue-collar job descriptions becoming smaller and smaller, where there's basically no information, and then jobseekers didn't have enough information to go on the best job descriptions that we see, or they get the highest response. Very simple: they explain what the company is, what they're trying to achieve, what the job is, and what you're going to be doing on a day-to-day basis, and then If there are any requirements for that job, e.g., a license or a certain educational background or anything like that, what that is and also ultimately what you'll be offered as the job seeker, it doesn't really matter what format you put that in if it's bullet points. If it's an essay, it doesn't really matter; it's more just having that information there. If a job seeker reads a job description, then it gets to the end and has questions; that's what's going to prohibit them from actually clicking and applying. If you can put yourself in the shoes of a job seeker and answer those questions, answer the very basic questions, then it'll work out. I think everything beyond that is noise; loads of people have done studies. I think you just mentioned it about if you include too many requirements, then you're going to get a drop-off of this from different demographics and all the rest of it. That might be true, and like a closed sort of test environment where you try and almost get that outcome that you're trying to write your paper on or something like that, I think for the majority of the population, that's not how it works. People just want to know if they're looking for a job; they want to know who's hiring me, what am I going to do, what am I going to get paid, and how does this compare to other jobs that are similar that I could also do and make a decision off the back of that.

TIM: One company we've worked with in the past that I thought was really good at being very forthcoming and transparent with information to candidates is a business called Get Your Guide. They're based in Germany, and I noticed basically like a continuation of what you're describing but throughout the process. So they added a job ad like anyone else, but even at that application stage, they were very good at giving candidates extra information, the kinds of things they would normally ask in that first interview, so they would give them the actual detailed breakdown of the share option structure, exactly how much the share option would be worth now and in different scenarios. They gave them like a one-pager and a day in the life of this role, links to the profiles of their teammates, what metrics they'd be measured on, actual products the team had built, and like descriptions of why they went through the process they did, like all these details, and what I found fascinating was how much Canada is like that. and then how much higher the continuation rate was for those candidates through the process because we basically de-risked it for them, they would then be happy to go through subsequent hiring steps because they'd already had all their questions answered anyway, so it's interesting to hear you say that even at that right at the top of the funnel stage, if candidates just get the basic things that they want to know, they're more likely to apply, which I guess is common sense. but it's a good reminder for businesses to try to include that on the job description.

DYLAN: Yeah, and I think too many companies look at hiring as all businesses are is a collection of people working, building something, or making something to make money, and that's all businesses are, and so when you start treating future employees and then hiring as just some sort of numbers game sort of mess, you know, it doesn't matter if you're at Amazon or Walmart or Target hiring hundreds of thousands of people a year, or if you're just hiring two people a year. You know that ultimately you want somebody to come into your business and then help you grow and all the rest of it. So you know it doesn't take much more effort to get them invested from day one. And by day one, it's on platforms like DirectlyApply, Indeed, LinkedIn, Seek, or whatever it might be, because that does help in the process going forward. Back to the point on the job description, it's just having a good job. the quality of your job description makes a material difference, or the amount of information you provide in that job description provides a material difference to the outcomes you get and for us the most important outcome is a completed job application if you start skimping on that as people are going through what can often be forms that ask loads of questions you're going to pull stuff out and big pain in the backside if you as people are going through that if they sit there and think I don't actually know much about this company I don't really know much about this job or I'm not even really sure how much they're going to pay me when they get distracted or another job pops up on their email what's the likelihood they're just going to quit off that one and go look at another one there's definitely a big relationship there and one of the things that we've been working on is providing that sort of I think encouragement is the wrong words but as people start their job application process they give us their resume for example and then we can say Hey Tim it looks like you're a great match for this job because of your skills doing a B and C and this company is looking for A B and C and almost just that sort of acknowledgement that we can provide as a platform to help people get through that process we've seen make some material difference as well so yeah I think for all companies to It really doesn't take much for the company you described in Germany to put all that together. It would take a couple of days, and then you can run it on every job seeker that you ever get on the top of your funnel going forward. and I would imagine their outcomes are considerably better than those that don't do that.

TIM: Yeah, and you are moving a lot of those; you're moving that information forward in the process so that you don't have to go over it again and again and again with every candidate. Surely that's worth the investment in time because not only are you going to not have to repeat yourself 50 times, but surely you're going to get more applicants in the first place to your point because you've de-risked it for candidates, and they're more

DYLAN: Yeah you'll get more candidates with higher intent as well you might get some people dropping out during that process and in some ways that's a good thing right because if something doesn't match with their values or their their expectations around compensation or whatever it might be to have them knocked out at an earlier stage before you've got to pay one of your internal recruiters or whatever to have that conversation that's great a lot of people you hear about this I think and before salary transparency became such a big thing you'd hear of these sort of stories where people go through a seven stage interview process only to be told the salary at the final sort of hurdle and it was 40 grand below what they were expecting and obviously we just wasted everyone's time. The same could be said for things like culture, what you're going to be doing day to day, progression, and all of that sort of stuff, and providing as much of it up front, I don't think Is that expensive to do from a talent acquisition team perspective? It just means that the funnel—or the point of the funnel that you as an employer step in and start interviewing and answering those questions—you've got a group of job seekers there that are super primed and ready to go and want to work for you.

TIM: Yeah, that's such an underrated aspect, I think, because I don't know you, but hiring for me could be drastically improved if both parties were just like awfully truthful to the point of almost being rude as soon as possible because, as you say, there's no point waiting seven rounds to then say, Oh, by the way, I want 200K." Oh, we're only offering a hundred or vice versa. Yeah, because that's just wasting everyone's time, and it's not just that; it's other things, as you say, like the career progressions and what you're actually going to be doing in the job. It's all those things that I feel like sometimes companies are trying to put on the sales pitch and trying to jazz everything up and make it sound better than what it is. Like, why not just be honest? And for the candidate, the same thing is just don't inflate everything, and it's easy just to bullshit on the CV, but it's probably just going to come on. You're going to come unstuck eventually, surely.

DYLAN: Yeah, for sure. I think it's called recruitment marketing for a reason, right? And not all marketing is brutally honest, and ultimately every workplace has its pros and every workplace has its cons, and so there is a balance to be struck there. I truly believe that the more data you provide, the better the outcomes you get. We see that as a data-run business: the more data we bring in, the better outcomes we can get. and we think that's the same at a high level for employers as well, but yeah, I think in my personal experience, whenever we hire people here, the first interview I'm brutally honest about what it's like to work here, the type of culture we have, what I expect of people, and what they can expect from me, and for some people they're just like, Okay, that's not going to work for me, and that's absolutely fine, and then others are like, That sounds amazing, but yeah, I think designing a funnel that, you know, allows job seekers to learn more and get excited is a great thing, and that starts on platforms like ours with your job description.

TIM: Have you seen companies go beyond a job description, or have you contemplated adding job descriptions plus kind of features, for example, Get your guide, shed all these extra dot points, like here's a day in the life of the role, here's exactly who's in your team, here's your KPIs? Here's share option information. A lot of the stuff is obviously going to vary drastically depending on the company, the role, the country, etc., but do you ever see any organizations really going above and beyond in those early stages and providing lots more information to candidates?

DYLAN: Honestly, not as much as we would probably like. One of the big problems in HR tech is the disjointed nature of all of the products. So you've got all of these different recruitment platforms, like us, Indeed, and ZipRecruiter, and you've got all of these platforms up here, then you've got all the applicant tracking systems, then you've got your internal HCM system, and you've got all the sort of external talent pool systems, and none of them really talk that well to each other, and there are companies out there trying to solve this problem. But what you can often see, which is sometimes frustrating for us, is you'll go to a company's career page, and it's fantastic; it lists out all of this information, it's got videos, it's got employee testimonials. It's got all of this called everything you described, and then you look at the job posting that they gave us to advertise for them, and it's pretty weak, and because of that, I find that the talent acquisition tech space was 10 to 15 years behind other industries from a tech perspective. And if you look at programmatic ads and all the rest of it, we're so far behind what Google and Facebook and stuff were doing in the 2000s, and I think that disjointed nature of all of these products means that it's not While they might be doing it somewhere, it doesn't permeate through their entire sort of recruitment marketing strategy, which is a shame. Sometimes we see people embedding videos or pictures or that type of thing in their job descriptions, which I guess is starting to move towards it. But I think for brands or for employers that want to really make this a focus, then trying to figure out how you can take the strategy that you might be doing amazingly well on your corporate career page And then how do you get that in with all your other vendors and stuff to then make it so that no matter where the job seeker finds out about you, they get that same experience? Currently, it's everyone's just trying to get everyone to the career page and then take it on from there, and that's one approach, but I think if we can figure out how to make everyone talk to each other and share all of this great content, then platforms like ours become far more exciting and interesting and valuable, and then the outcomes for the employer become more valuable as well.

TIM: One other thing I'd love to cover off with you is AI, of course, to most said letters in the history of the world, probably AI together. And what I've heard repeatedly over the last few weeks, I'd love to share with you to get your perspective on it to see whether or not we're on the right track. So many hiring managers have said to me they're getting inundated with applications, particularly in the United States but also in Europe, so inundated with applications they're noticing the applications seem to be more and more similar to one another, seem to be ChatGPT-written or optimized. And they're finding it difficult to differentiate among them, and so now they're basically looking at this pool of all these quote-unquote amazing-looking CVs that seem to match the job description perfectly, and now they're stuck with thousands they don't know how to deal with, and it feels like that screening step is even more broken than it has been in the past. Because now CVs seem to be even less relevant or connected to the person than they were even a few years ago, are you seeing that trend? Have you done any analysis? Do you have any sense of whether or not candidates are using AI for either applying to roles or to augment or write their CVs?

DYLAN: Sure, this is quite a complex and obviously very timely sort of topic. There are a few different parts to this here, so I guess at a high level we are currently seeing less than 2 percent of applications through DirectlyApply that we're detecting as being using one of these AI tools. and these AI tools do lots of different things, but one of the There's that kind of split into two buckets one is the auto apply which is this idea that these tools can go and apply to jobs on mass platforms like ours block this type of behavior you can't do it our system will detect it and block it we're looking for human levels of interaction with apply processes and if that suddenly starts happening in a few milliseconds then we understand that's not a human doing it that can block it down the second part is this idea of job seekers using tools and they say here's my resume here's the job I want to apply to fix up my resume to make it a perfect match so we've detect we built detection systems for this and it's sub 2 percent of total and we process tens of thousands of applications a day of that two percent though it is very much skewed towards students as in university graduates and also more technical roles and particularly software engineering and anything tech data that's where the focus has been we've detected almost nothing for like blue collar workers and things like lawyers and that type of thing so there is a focus and that kind of makes sense in terms of the early adopters of this potential stuff And so then in terms of the reports around huge numbers of applications, there's no denying that is happening. if you go to LinkedIn you can look at any job post and it'll tell you how many applicants it's had and some of them are wild right it's 900 applicants in the last day I think there's a few things here I think one is you've got this idea of easy apply and so LinkedIn is a big one and deeds and other ones that recruit to do it as well where basically once you've got a profile on these platforms you can just click apply and I think that does actually make up a large part of what we're seeing here particularly within sectors that over the last 18 24 months have seen a lot of layoffs if you look at any talent acquisition job in a company in America that are getting hundreds of applications within the first day same with tech roles sort of entry level data roles all of these types of things are getting huge amounts of volume the other one is obviously if you advertise remote or hybrid you're going to get a huge number of applications I we don't think that job seekers using these AI tools is necessarily a bad thing I use ChatGPT and stuff almost every day in my profession to help me write emails or do things and that type of thing, and it's like a better spell check or Grammarly, and I think there's value there. We did some work looking at basically the hallucination aspect of us and how many of these resumes are actually being submitted where they've probably modified it to the point at which you're now lying on your resume. and that was the most concerning part there, but we don't see it at the moment as mass panic or anything like that. If hiring managers want to limit this a little bit, I would put in maybe stop using the easy apply systems, making sure that your ATS is set up really nicely so it's not asking a thousand questions and you've got to create an account and all that sort of stuff. You can have a nice ATS process that makes a job seeker actually click and actually answer a couple of questions. Upload their resume, click go, and that'll drop it down massively, and just be conscious that obviously in some we've seen huge amounts of layoffs and that type of thing. and there are a lot of people, particularly in the United States, particularly in the recruitment tech space, who are looking for work, and so application volumes will be higher, but yeah, we keep it under constant review and see what's happening, and I think a lot of other platforms similar to us, like LinkedIn and stuff, probably identify that this could potentially become a bigger issue in the future, and I think you'll see a lot more sort of interventions when you go to apply and that type of thing to make sure you're a good match. What we're doing currently is if we think that you may have used GPT to enhance your resume, we highlight this and just say, Hey, Tim. No worries if you want to apply, but just double-check that this hasn't actually made something up that's going to make you look silly down the road. and I think a lot of tech vendors can do a lot of stuff like that to mitigate the problem, and almost understand that's probably the new reality of how people work, not just in recruitment or job searching, but we're going to be using these tools a lot more instead of thinking about ways to just stop it or block it. How can we work alongside it and still create a good recruitment experience?

TIM: Yeah, I feel like that's the only real mature response, yes, and I get the sense that hiring managers are now thinking like that throughout the rest of the hiring process as well, so on testing, on interviewing, on anything, and yeah, particularly in data and tech, it would be hypocritical, I think, to tell candidates Hey, don't use this groundbreaking AI technology that we need you to use in your job; otherwise, you're going to be nowhere near as productive as what you should be. It would be a bit of a hypocrisy, but I guess just because it's developed so quickly and it is shattering the way hiring is being done, if you can just get a score of 70 percent on a test by chucking something through Chachi Petit, then obviously that's going to shake the system up a little bit. But I guess this will just, yeah, this will sort itself out eventually once we come to terms with how to deal with them.

DYLAN: I think we've and if you look at the progression of applicant tracking systems and basically online resume collections for the last 30 years, there was a time when everyone would write all of the keywords and then put it in white text, claiming that they did it with the SEO, because then the ATS would be like, Oh, he's a… it's nobody actually got hired incorrectly off the back of that. and we have to remember that for the most part I don't think there's any fully automated recruitment from start to finish you don't speak to a human and this is just where the value of a good recruiter either internal or external comes into play is no matter sort of what trickery is going on your job is to identify the right job seeker ask the right questions get the right response and then make the right hire and I think that there's always some sort of thing going on that I see a lot of very interesting claims around what ATSs can and can't do in terms of automatically rejecting people and making a decision about your life based on and 90 percent of it's nonsense from what I can tell anyway And I think also it's around education for job seekers as well as and that's what we've tried to do is if you use these tools you have to look out for their downfalls this is where they can go wrong and you're going to look really silly in that interview when they ask you about how you did A B and C and you look blank because you didn't even realize that was written on your resume and ultimately a job seeker is only going to make a mistake once I think we're in a bit of a hype cycle of it at the moment combined with in some sectors of the economy quite poor employment prospects and it's created almost the perfect storm I think we can definitely revisit this conversation in 18 months and say, I think it'll blow over, and we'll think Oh, that's funny. We thought that was going to be a big sort of surge. I think, like everything, it'll be a few percent, and we'll just learn to adapt and move on.

TIM: Yeah, it's interesting to hear your thoughts because you have such a broader array of job seekers and jobs, whereas I'm very much fixated on the data and tech, which is where you're basically saying

DYLAN: Exactly, yeah. I think that's definitely probably the number one place where we're seeing at the moment, and that makes sense, right? Like, you see all these cool tools on GitHub and the types of people applying to those jobs so in tune with the latest progressions in tech, and they want to try it out and all the rest of it. and in some ways that's quite cool, but yeah, as of yet we haven't seen it move into the wider employment market.

TIM: So again you're making the distinction between okay use of these tools to improve, optimize, and create your CV and some issues with that and then also with the mass applying to different roles, and you said for you guys it's pretty easy to tell if someone's trying to use one of these tools because if they don't rate limit it, they're just trying to bombard you. You can tell that you're a human kind of quiet three times in a second. What about other platforms? So do you know LinkedIn? Indeed, seek other big platforms like this preventing the use of these tools or

DYLAN: I believe so. I think I saw one the other day that sort of went a bit viral. It was on GitHub, and I think it was a LinkedIn-specific one, and looking at the source code, it was actually trying to rotate user agents around to try and bypass the mechanisms that LinkedIn had in place to do it from memory. This was a few years ago, so it might well have changed, but I'm pretty sure that indeed we'll block you after a number of applies per day. I can't remember what that number was, but I do know that there was a limit, and some people during the COVID sort of thing were starting to hit that limit. And we're complaining about it online so I I think and I think that's probably why you haven't seen especially these big job sites come out and really say anything about it because it is a bit of a non issue for them currently I think they can handle it obviously you're going to get some tool that will figure out a way to it's essentially hacking it to bypass the protocols they have in place and get through I would imagine that's relatively short lived because like us we come in the morning we look what happened overnight oh there was one job seeker that managed to apply to a hundred jobs in a minute it's it wasn't a human was it oh what did they do block that and then we move on so I think and I think in general like it's while lots of platforms like the big ones do make money on performance like we do it's bad for business if you're just sending through loads and loads of unqualified robot appliers you're going to get very unhappy customers at the end of it So there's really no incentive to allow it to happen. You've got to balance creating an easy experience for job seekers with also an easy one. You don't want to make it so easy that bots are then able to exploit it, and it's finding that balance for us. It comes all the way back to that sort of original matching and the sort of marketing element of it is that if you've got a job seeker that you've shown a bunch of options to, they really want to apply to this job. When it comes to the application form, you don't actually need to make it one-click apply; you can make it a five-step apply because the intent is so high they're going to want to apply. When we start reducing this apply process down to one click or whatever, which then becomes very easy for bots to simulate, you then have to question You know, if jobseekers aren't willing to do a couple of extra steps, have we actually done a good enough job as a platform to match this jobseeker and make sure they want to apply before putting them through this process? I think just getting that balance right solves a lot of these problems quite quickly.

TIM: And if I think back to my days in online travel, the bots were the bane of our lives, and it was this endless game of cat and mouse of blocking the bots, letting them through, filtering out the traffic, and dealing with this and that back and forth, blah blah blah. I know for LinkedIn, I feel like this should not be a problem for them because if you just are logged in, then you know who the user is, and you could block them after applying five times in a day. and it would be pretty rare that anyone would have more than one LinkedIn profile. I feel like it's for the platforms where you're logged out most of the time, and you're just a cookie that the chance for abuse seems a lot higher to me anyway.

DYLAN: Yeah, for sure, and I think when these things become impactful on those platforms businesses, I think they'll intervene pretty quickly. Losing money is a pretty good driver to getting stuff done pretty fast. It'll be interesting if over the next six to 12 months we do get almost statements or product announcements from the LinkedIns and Deeds that recruiters, etc., or the Sikhs say. Hey, we've identified this as a problem, and this is what we're going to do about it. It'll be super interesting. It's highly possible that they don't actually say anything verbally, but they just make changes as well. I think that would be interesting because while we're quite big, they're huge, right? They're very much getting a lot of data run every minute at the moment around what's happening with the stuff, so when we see them start to take action, or if they ever do take action, I think that'll be quite an interesting point.

TIM: Yeah, and for them I feel like it is potentially an issue because they have, or at least last time I checked, a few different commercial models. Like, you used to be able to pay for a job slot, and we pay X amount per year, but now they also have is it per impression or per application pricing? Something that's almost like that, which means that you as a customer would be pretty pissed off if you got low-quality, spammy applications coming in. So maybe there's something they're going to have to look at that

DYLAN: Yeah, for sure, and like I say, as soon as it becomes a commercial problem, the technical solution will follow pretty quickly.

TIM: Dylan, tell us a bit more about AI Talent Agent, which is something you guys released relatively recently.

DYLAN: Yeah, so we're actually rolling it out at the moment, and what we were trying to figure out is what the future sort of interface for job search is and how are we going to interface with our career and our job search 10 or 15 years from now. What's that going to look like? and it's hard to predict, obviously, but we probably feel like having a website where you type in a job title and a location and clicking search is probably not where we're going. We think that what we've tried to emulate since day one is that sort of recruiter headhunter experience for everyone. We want you to feel like you're You directly applies to your career, knows your progression, knows your skills, knows what you want and don't want, and is able to make intelligent suggestions off the back of it. Most people don't. Most people in the economy have never been headhunted or had a recruiter reach out to them and say, Hey, would you like to apply for this job? But we can provide that experience through technology, so we think that voice is a very natural and intuitive way to interact in the same way you'd pick up a phone and have a phone call with a recruiter. Text is good, and we've built in text as well, but we've made voice the primary interface for the agent. so what it basically is as a user of directly apply in the bottom right of your screen you can click it and a Siri like interface opens up and you're able to chat with your talent agent and you can name it you can change what it looks like you can change his voice you can do all that sort of stuff and so initially we launched it in terms of just asking questions about the job you're on so you can ask does it have this benefit what's the salary show me similar jobs and all the rest of it as we've been developing it we've given it more context into the individual job seekers resume You can ask, Am I a good match? What skills do you think I'm missing? Where does this role sit in my career progression and all of that type of stuff? And it's able to provide really good contextual advice and answers based on all of that data, and it also pulls in, obviously, all of that data that we're talking about previously that we use for job matching as well. Because we're not focused on it, we only get paid as a business if we have a successful outcome for that job seeker, so we designed it to not be afraid to say, Hey, Tim, this job's probably not right for you, but this one might be because we're more focused on just trying to get that job to you. And we think there's loads more we can do here so when you get a list of search results you can say Hey I really want more jobs that do this or less of that and we're able to update in real time to to do that so we've been rolling it out to different sectors we hope to do a full rollout in December and it's been really cool watching the engagement from job seekers chatting away with basically our product in a way that you couldn't do before and it's far more natural than trying to have a whole bunch of check boxes and select tabs and text fields and all the rest of it so when you remove that and you just say talk to me tell me what you want and we can provide really high quality answers that's been super cool and it might not be the future of job search in terms of the interface but we feel like it's going to be something closer to that direction than what the industry currently has

TIM: Exciting times, and yeah, I can't wait to see what the future holds for Directly Apply and Dylan. It's been such a great chat and conversation today. I've learned a lot, and it's great to speak to someone at the coalface with direct access to truth-level data. You don't get to speak to someone like you every day. So I really appreciate you sharing all the insights with that audience.

DYLAN: I appreciate you having me. Thanks, Tim.