The Recruitment Hackers Podcast

Hire better and faster with Talkpush, Let’s jump on a call and explore how Talkpush can help your team.
FEATURED ARTICLE

The AI Mirage in Hiring: Are Vendors Innovative Solutions or Industry Snake Oil?

March 6, 2024
5 min read
Read article

Receive the latest on recruitment automation directly to your inbox

Filters
Topics
      All Post
      Podcast
      5 min read
      Max Armbruster
      Max Armbruster
      CEO Talkpush

      Assessment Series 4 - Hiring based on merit with “on the job simulations” with Mel Macdonald from Vervoe

      Ready to level up your recruitment?
      Let’s jump on a call and explore how Talkpush can help your team.

      Episode 63 full cover-1

      Welcome to the fourth installment of the Recruitment Hackers Podcast Assessment Series. In this series, we'll be diving into different platforms that provide virtual assessments solutions. Today we’re talking about Vervoe with Senior Sales Executive, Mel Macdonald. In this episode Mel walks us through a demo of Vervoe’s assessment solution, which helps employers hire based on merit, rather than experience — through “on the job” simulations to determine whether candidates have the skills to perform.


      Watch the entire episode below.

       

      Don't feel like listening? You can read the entire transcript right here. 👇

       

      Max: Hello, and welcome back to the recruitment hackers podcast. I'm your host Max Armbruster. And I'm delighted to be welcoming to the show today. Mel Macdonald, a senior sales executive at Vervoe. And you may remember Vervoe for some of our past events and past shows.

       

      It's a company that is changing the way assessments are being done and allowing for... And Mel will correct me in a minute, but allowing for recruiters to design assessments that are very skill-specific and that allow to remove bias from the initial stages of the selection process for a very wide variety of job types.

       

      And so it's fascinating to imagine what they were able to build that has such range and today I hope to find out from Mel a little bit how this was built and what this looks like to give you, the audience, a visual experience as well. We'll be putting this show on YouTube so that you can see and get a feel for things and decide if you want to know more.

      So thank you for joining us, Mel. Welcome to the show. 

       

      Mel: Thank you. Thank you for having me very excited to share what Vervoe is and what we do. So thanks for having me. 

       

      Max: Pleasure pleasure. And we found out about Vervoe at Talkpush a couple years ago when we realized we had a customer all the way in Costa Rica.

       

      And you're in Australia, I'm in Hong Kong, we've got teams everywhere. It's wonderful to see how technology works. It's really a borderless world. But the original story of Vervoe starts in Australia. Maybe we could start there and tell us a little bit you know, what Vervoe does, where it came from.

      First of all, did my introduction give it justice?

       

      Mel: Yeah, I completely agree with everything that you've said. So the whole premise of hire based on merit, not background. So someone's previous job title isn't going to indicate how they perform on the job. Ideally in an ideal world, if you could always see someone do the job before you hired them, it would be very clear if they can.

       

      And essentially that's what Vervoe is trying to do. So we create assessments that are more like on-the-job simulations. So putting a candidate through the experiences and the skills that they will actually need on the job so that they can showcase that. And then you can make an educated decision and choose the person who's going to be the right fit for the job.

       

      You mentioned being able to work across, you know, different job types as well. It's the flexibility of the platform that allows us to do that. So Vervoe doesn't sit here and tell a business what a customer service representative should look like. We've got all of the baseline measurements, but each job is very specific and we allow businesses to create an assessment that's truly reflective of that.

       

      Max: Yeah. This is like two conflicting philosophies. One people who believe standardization, calibration. We want to measure ourselves against a certain, you know, like SAT score or IQ score. And then the other position is no, every company is unique. Every person is unique and the self-assessment is what should drive the decision-making to create true differentiation. 

       

      So you're obviously... Vervoe is focusing on customers on the second half of that equation

       

      Mel: Yeah, because it's that second half that really indicates what the on the job performance will be like. A lot of, you know, if you look at psychometric, introverts, extroverts, IQ scores, it really doesn't tell you when you get that person in the role, how they're going to perform, and who's going to be the best fit.

       

      And we can't create a truly diverse workforce if we are wanting to fit everyone into a very specific box and that same box is used company to company because companies don't fit into boxes either and their culture is very different and you want to make sure you've got the right person for a culture fit for the role fit across the board. And that's what we're trying to do and removing bias from the process as well.

       

      So really just taking it down to have you got the skills to do the job, allowing the person to showcase that they do, irrelevant to whether they worked at, you know, a fortune 500 company before, or a completely different role. More about screening people in rather than screening people out, which I really love. 

       

      Max: Just to play the devil's advocate here. You could make the arguments that well, I'm really looking for people who are going to be happy here, happy with our culture and who fit in the psychological makeup of our team.

       

      And then I can teach them anything, any skill in the world that can be taught, but culture cannot be taught. Well, how would you argue against that, if you want to.

       

      Mel: I'm going to say we'd be digging into what that culture actually looks like and what those attributes are. So is it the ability to openly communicate and being able to deliver a good feedback?

       

      Let's look at that. Then let's put someone in a scenario asking them to deliver feedback and see how they do that. Generally, a business that wants to grow and be better and perform better. You need a diverse range of people to do that. So having someone join a business, that's just a perfect culture fit.

       

      You can teach people, but that's a lot of work. So even I deal well, that would be better to find someone that can have that culture fit those softer skills. And also look at who's got those you know, maybe the technical skills or the customer service skills that also go with that role. So it's more a holistic approach rather than just taking you know, can an engineer code.

       

      Or can an engineer also fit our culture or let's take them both together and find someone that's got the coding ability, the communication side and the company philosophy all into one. 

       

      Max: Yeah. Yeah. And then I think that the culture can be just a bag where people put in a lot of things, but they need to be analyzed bit by bit.

       

      How do you make an employee happy when they start a new job? You make sure that they're good at the job that they were hired to do. That's probably the highest correlation you can possibly find above culture above anything. And it's like, can I do the job well, do I feel useful? Do I feel like I am contributing? And with Vervoe that's one of the metrics you're driving I assume. 

       

      Mel: Yeah, well, especially since it is like an on-the-job simulation. So the candidate gets a very good idea of what the job will actually entail before they're even walking into it. So in the same way someone might choose to self-select out.

       

      If it's a really difficult role, that's on constant time constraints that needs a high level of resilience and time management. If we're creating an assessment that reflects that, that has, you know, quite difficult questions. One after the other that are all timed. Certain people will really thrive in that environment.

       

      If someone is not going to thrive in that environment, they're probably not going to be the right fit for the role. So taking them through and showing them what the job will look like before they even take it is ideal from a candidate perspective and from an employer's perspective. So yeah, you hit the nail on the head there.

       

      I think it's important from both perspectives, but for people to have a really good idea of the job and often the bullet points in a job ad it's very difficult to say what your daily tasks might actually be. So we can help with that and showcase. 

       

      Max: Well think this would be a great time to give the visual, the people who are visual like me, I feel for what that means. And Mel  I welcome you to guide us through a few you know, some visual experiences. I don't know if you want to start from the candidate side or the recruiter side. I'll let you take it away. 

      Mel: Perfect. So I'll start from the candidate side. That's a very important element to any kind of assessment that you're putting candidates through.

       

      Across the globe. There is a lot of issues with candidates shortage, so you're wanting to make sure it's very friendly. Experience and how you're positioning. It also needs to work with your own brand as well. So I've chosen a call center representative role. Every step of this process is very customizable.

       

      So use it as a visual cue and an idea what the assessment can actually look like keeping in mind that every question can be changed depending on what the job looks like, because we are wanting it to be you know, a simulation of the role itself. So we can integrate with different ATS systems.

       

      So depending on whether we use right at the top of the funnel or, you know, further through the funnel as a validation process an email can be sent out to the candidates. They don't have to create a new log-in or you know, I think everyone who's applied for jobs probably has a million logins to a million different places.

       

      We'll always take the candidates through a process to test their microphone and camera. If we've decided to use questions that require that. And then they simply come through and complete the assessments. So we have a variety of different question types that can be used that will be so that you can make this as relevant to the job as possible. So giving people situations, asking them to type what their answer would be. 

       

      Max: I like the fact that you can on the right, you can see how many questions are there. So you're mentally ready for how long this is going to take. 

       

      Mel: Yeah. And you can jump ahead and look at different questions if you're wanting to as well. If you're wanting to go back because you've decided to change question two, once you've gotten to question seven that's fine as well. It's very easy to just click to record audio. It's a very mobile friendly process as well. So it's all mobile optimized, which is very important because a lot of candidates now will do the entire process through their mobile device. 

       

      So yeah, you just take them through, we've got some audio questions. This is a phone based role, so that's going to be very reflective of what the role actually looks like. Using video questions would be perfect if they're needing to present themselves to people or explain things to people as well.

       

      I mentioned before having time-based questions. So if you're needing a candidate to be able to come to conclusions or read information, acknowledge it and come to an answer quite quickly, putting in time questions can be very helpful to see how candidates will perform under different time constraints.

       

      They're always given a warning that it's a timed question. We don't just throw them into it as well. So knowing that I've got five minutes to answer this it's a video that you can watch. And again, you can incorporate from an employer branding side, some of your own like branded content having income, watch the video and then they have to type their ads and there's a little clock that's counting down for them here. 

       

      So each step of the process, it just really should reflect what that job looks like and what skills you're needing this person to have, because the output, which I'll show you in a moment  that's where the time-saving, that's where it becomes really clear what the value is from a company side as well.

       

      Max: And so you know for people who are seeing this for the first time. Is this coming out of the box  or do your customers typically design their own questions and start from scratch or what's, you know, how does it look from an implementation? 

       

      Mel: Nice simple ways  that we work with different customers on this. So we have got a library of assessments that are there ready to go. You can use them, as you said, straight out of the box and use it question for question exactly as. Or you can go in, if you're wanting to add company names or specific scenarios that are more relevant to the job you can go in and edit any questions.

       

      It's very easy to go in and create an entire assessment from scratch, and then you can choose what question types and specific. The questions that you're wanting to go in. And then we also have an in-house team that builds assessments for our clients as well. So we've got a psychologist on board as well.

       

      We work very closely, both with the TA teams and the hiring managers to find out what does a really great employee look like and why, and then we can build an assessment to make sure we're measuring what's most relevant. So there's a number of different options for businesses of how they want to do this.

       

      Max: It's a very customized approach. It's not, you know, it's not one size fits all. 

       

      Mel: Exactly.

       

      Max: Okay, great. Well, that sounds great. Too good to be true, too good to like actually generate meaningful insights. If everybody's got their own assessments, then how do you measure performance? It's an impossible problem to solve. Isn't it? 

       

      Mel: Well not with Vervoe. So any tests that directly mimics what a person will do on the job would be considered validated. Our AI is very unique as well, and this is where we take that customizability. You know, another step further. So we have a three-way machine learning model.

       

      So how and what, so how is essentially like a one-way mirror watching how I'm completing this assessment, erratic mass movements, jumping around the assessment, typing, deleting. So any interaction. The what model is all of the contents. So keywords sentence, structure, spelling, grammar, and so on.

       

      One thing to note that's you know, very important as well. So we don't use any facial recognition technology in our video or audio recognition in our audio-based questions. Unfortunately, there's alot of studies that have shown it can kind of include bias in the process ability to understand facial expressions of people, different backgrounds and so on.

       

      So we transcribe what the person has said, and our AI is only at a scoring the text. However, the video is obviously available for the team after to be able to watch. And then the third model is preference. So again, this comes in, what's a 10 out of 10 for a customer service representative at Talkpush, for example, that might look very different to a customer service representative at Vervoe or for retail or for, you know, any different business.

       

      So you can actually come in and score questions and say, this is a 10 out of 10. It feeds back into our system. And then we will know. This is what you're wanting to see and can continue to score candidates accordingly. And so when you've optimized that AI, essentially our AI will be scoring the same way that the TA team would.

       

      Max: All right. Could you give us maybe a look at what, you know what you get on the other end

       

      Mel: Definitely. So when you log into an account again, we've got our assessment library that you can choose different assessments that are ready to go. As you said, straight out of the box we'll create an assessment from scratch.

       

      It's very easy to get an idea. So if you're looking for a sales consultant, you can get a view of what skills in this particular assessment of being measured and what question types. And then if you're like this one looks good, you can add it to your assessment library. So creating assessments and sending them out this process is very easy and straightforward.

       

      What is generally most interesting is the output once a candidate completes the assessment, so I can jump in and show you that. 

       

      So we've got this call center agent. We've had 25 candidates that have been invited and 23 people who have completed the assessment, you can get an easy view of what the average score is.

       

      And it's very quick, even before I explain what everything means on this page, you can quite quickly see which candidates you would likely reach out to. Who are your kind of middle performers? And then maybe the people that don't quite have the skills that are necessary for this particular role.

       

      Max: The score is a much-demanded feature. I know people always asking, can I get a score? But if each of your assessments is designed, custom-designed for specific jobs within specific companies. How do you arrive at this score?  Considering the level of customization you’re promising.

       

      Mel: So in this particular assessment, we've got five different skills groups that we're measuring. So using our how and what model and the skills that have been selected and all of the data that we've collected from the thousands and thousands of assessments that have been completed it's already very accurate to be able to give this AI score.

       

      When creating the assessment as well, you can give examples of what a 10 out of 10 answer would look like. So you've already started teaching our AI exactly what it looks like. So it will take the questions. Generally, we would say three questions at least per skill group. So that you get a well-rounded, it's not like a fluke, you know, they've happened to answer this question correctly.

       

      And then using that, how and what model, we can identify how they've performed in these particular skills groups. And then again, through optimizing the AI and teaching us what a zero out of 10 and what a 10 out of 10 looks like. We've got clients that now have this score that is within a one to 2% variance of if a TA team member went through and individually scored every single question.

       

      Max: Yeah this optimization cycle I suppose it's at the beginning of a client engagement. How long does it take, you know how much time does it take to optimize the engine to give a score that's representative of our recruiters’ sensibilities?

       

      Mel: Yeah. So right out of the box, it's between 85 and 95% accurate because we have so much data behind us anyway. The optimized process can come through in several different places. So when you're creating an assessment, like I said, giving examples of what a correct answer would look like.

       

      If you're actually going in through the candidate which if I click on here because Cali, you know, performed very well across all of the different skills groups. If I go and look at some of her answers I watched this recording that has been left and I'm like, wow, this answer really blew me away, you can also manually grade down here. The ideal place to optimize an assessment is here throughout optimized tab. So the reason we have these set up is because we're all about removing bias from the process. So what we will do is take answers real answers from the candidates who have completed.

       

      You'll have no identifying information about them, just their answer. And then we'll ask you to tell us whether this is a zero or a 10, and then we can then monitor. So you've told us this is a 10. We agree. Like we've also scored it as a 10. Perfect. We're on the right track. These persons, it doesn't take very long. Usually, within the first, you know, couple of days to a week, depending on how much time someone wants to spend on it. We do have clients that don't use optimize it all because they find it REI. Accurate regardless. So this is just if you're really wanting to get you know, something quite specific to your business and quite specific to this job role.

       

      Max: All right. Great. And to illustrate the level of specificity that we're talking about, and I saw you had your different categories, sales, customer care, and others that are more high volume positions are there some really sort of weird in here most examples of really weird, specific jobs that ended up getting assessments done for that?

       

      Mel: Yeah, so essentially we can help create an assessment for any role, really it, depending on how regardless of how specific it might be. So even if you take a very general role, like a sales person, but it's specific to the medical field, so there will need to be an element of clinical knowledge that's also required there.

       

      Or a data entry job, but they need quite specific knowledge of Excel. We've got Excel-based questions and we would make it reflect that. So the customizability be the fact that we don't say here's a set question that you have to use, you can, of course, choose from our question bank which is about 300,000 questions.

       

      But if you really want to make it specific and it's quite a niche role then of course you can create something. Give us examples of what a great answer will look like. Teach our AI exactly what you're wanting to see from that assessment. So really that the limits of what job roles we can create assessments for is almost limitless.

       

      Max: I would say the ultimate diversity tool you could say cause you're really enabling for companies to build their own assessments, their own profile. And you could have, like you said, like a salesperson that is very niche and one particular industry, we know that a sales professional will look different from one company to the next of course

       

      Mel: I think that the ability to customize. And even create assessments of different lengths means that we can be used right through that the hiring funnel entry post-hire as well. So, you know, right at the top of the funnel with higher volume type roles, creating a shorter assessment so that you're getting a large volume of candidates through validating a smaller number of skills. Through to you know, very senior level roles where there might be an element of headhunting that goes into it. And you have a much smaller pool of candidates that you're wanting to really make sure have an in-depth skills fit an in-depth culture fit and to be, yeah, testing that and can be used then on a smaller pool.

       

      And then we also have a number of companies that use us post-hire so much longer assessments to identify where there might be skills gaps to offer additional training so that your existing workforce can be you know, improved and looked after in the right way. That's for them, rather than sending everyone through the same training where it might not be relevant.

       

      Also for internal mobility, I think promoting the person who's right for the job, not the person who's positioned themselves in the right way for the job. So yeah the flexibility of the platform is definitely a plus. 

       

      Max: Oh, it's so nice to be able to say, talk to the machine before we have that, you know, really strategic conversation about your career, because you know, those can be very emotional discussions and at least the machine can give you a good indication of whether you're heading in the right direction and about the training portion and the career mobility. I know that some employers we'll use Vervoe to set a minimum threshold but then of course there's the near higher and the near higher bucket can also be directed towards a near hire recruitment program, which is something that's used a lot in the call center space where like, ah, the communication skills aren't quite there, but with a little bit of help, we can get there. And so it could be used in that sense. I think that's how our customers are integrating the Vervoe score to move candidates to different buckets, you know, top potential and near hires and no hires. So that's...

       

      Mel: Right. It's based on real skills. So it's not just asking someone, can you multitask, which everyone says yes to it's giving them, like actually seeing how they multitask and then you know, using data to make these sorts of decisions rather than an interview, which also you know, very much not reflective of what the job might end up looking like.

       

      Max: Yeah, well, Mel I'm inspired. I hope our audiences as well. If they want to reach out to you or what's the best way to get in contact 

       

      Mel: Just drop me an email. So mel, M-E-L, @vervoe.com 

       

      Max: You got to spell the Vervoe for our audience. Then it's on the t-shirt V-E-R-V-O-E

       

      Mel: That’s right.

       

      Max: Thanks alot, Mel.

       

      Mel: Thanks so much, Max.

       

      Don't forget to subscribe on your favorite platform for new weekly episodes.

      👇

      SUBSCRIBE