HRExaminer Radio Executive Conversations Badge Podcast Logo

HRx Radio – Executive Conversations: On Friday mornings, John Sumser interviews key executives from around the industry. The conversation covers what makes the executive tick and what makes their company great.

HRx Radio – Executive Conversations

Guest: Drew D’Agostino, CEO and Greg Skloot, President of Crystal
Episode: 346
Air Date: November 15, 2019




Important: Our transcripts at HRExaminer are AI-powered (and fairly accurate) but there are still instances where the robots get confused and make errors. Please expect some inaccuracies as you read through the text of this conversation. Thank you for your understanding.

Full Transcript with timecode

John Sumser 0:13
Good morning and welcome to HRExaminer’s Executive Conversations. I’m your host, John Sumser and today we’re going to be talking with Drew D’Agostino and Greg Skloot who are the CEO and COO of Crystal. And if you’ve been following my stuff over the years, you’ll know that Crystal is a tool that allows you to do quick back of the envelope personality sketches, this is going to be a great conversation. So hang in there.

Unknown Speaker 0:40
All right,

John Sumser 0:40
Drew and Greg, introduce yourselves.  

Drew D’Agostino 0:43
Drew D’Agostino, CEO of Crystal. I have been at this for about five years since we built our first little prototype that could analyze personality

from writing styles, resumes, and any other text samples and since we’ve kind of grown into different spaces like HR but also in talent, recruiting sales, and marketing as well.

My role here revolves around product

technology. And basically the high level of what we’re doing, Greg,

Greg Skloot 1:08
Greg Skloot, I’m the president of Crystal. And you just stole most of what I was going to say. I think he said it really well. But Drew and I’ve been working together for almost 10 years now and a while in Crystal. So we go back a while. And similar to what he said, you know, Crystal is an app that can tell you anyone’s personality. And we’ve built that out as something as a pain point that we found in our in our last company, and it’s been a great journey so far.

John Sumser 1:31
Okay, so that gets Crystal with a nutshell why don’t you talk for a little bit about how crystal actually operates, right? Because crystal can tell you somebody’s personality, but you do it in line in a variety of applications. How do you experienced Crystal if you’re a user?

Drew D’Agostino 1:48
Absolutely. So when we think about crystal and giving someone a per personality insights, our goal is always to provide the most accurate personality profile possible with the data available to you. The first way we think about it is for people you don’t know. And for people you don’t know, a crystal can predict personality using AI and machine learning based on analyzing a text sample. So this could be a resume or a LinkedIn profile. And this is what’s commonly used by what we call professional communicators. So particularly people in recruiting, in sales management that are about to have a conversation with someone that they maybe haven’t spent much time with before. And crystal kind of serves as a coach whispering in their ear, hey, here’s how this person likes to communicate, here’s, you know, a prediction of how they may behave. So you can adjust your communication style to make a really good impression and build a good relationship. So that’s, that’s part one for people you don’t know. Um, it’s a prediction on the other side of crystal and how crystal works is through personality assessments. And these this is based on the personality framework desk when you’ve been around for a long time. And basically, anybody could come on to Crystal sign up for free and take that personality test. They can invite their co workers and friends to do the same. And anyone who does that can see each other’s personality profiles for free. So this is a really great tool that people use in their personal lives and and at work just to better understand each other. So really bottom line, the way people use crystal, two sides of people, you know, through inviting take a personality test, and people you don’t know by using the AI to actually predict their personality by analyzing a text sample.

John Sumser 3:29
So that means practically speaking, walk me through how I would experience Crystal in email.

Drew D’Agostino 3:38
So let’s say you’re emailing somebody new, what you probably do is pull up their pull up their LinkedIn profile or upload their resume into the crystal dashboard crystal would scan the LinkedIn profile for the resume and it would show you either in our Chrome extension, so a little sidebar that appears in LinkedIn or on the crystal website, it would show you some really tactical easy to understand insights about how to write that email. Now show based on the person’s personality profile and what you want to email them about. So perhaps you want to email them about an opportunity for a new role. That’s an option and crystals, you just click I want to email drew about a new role, and it’ll give them some really tactical guidance on how to do that. So subject line example, bulleted list of do’s and don’ts, how to open up the email how to close it with a good call to action, lots of little snippets that you can copy and paste them even some templates that you can use and rework into your own story. So recap. The way would be using an email is you quickly scan the person’s personality profile or through a resume or LinkedIn and then look at the insights crystal provide to shape how you how you frame that email in a way that will resonate with the receiver.

John Sumser 4:41
Interesting. So how accurate is that?

Drew D’Agostino 4:46
Yeah, so it’s really remarkably accurate prediction to the ones where we don’t know someone are typically about 80% accurate. So four out of five of those should be right on target are quite close to that person’s actual personality. And one out of five of them might be a little more And we measure that based on a couple of ways one through user feedback. So on every profile, we predict users rate it on a scale of one to five on accuracy and we compiled and measure thousands of those ratings. And then we also compare people’s predicted profiles to their verified profile. So the ones that actually have a personality assessment and for the accuracy on the personality assessment version so the one we’re we actually know the person and they thought the personality testers are typically around 95 97% accurate so very, very accurate.

John Sumser 5:30
Interesting. So how does the AI part of it work? I certainly somebody’s resume a new told me their personality. That’s that sounds catchy to me. How does that work?

Drew D’Agostino 5:41
There’s a few different five different learning that go into it. And then we’ll get into the technical details but there’s it’s not just like one answer about anytime somebody writing something about themselves by their jobs or something, you know, read famous in some way, every word choice and every free choice and all the little facts are actually little bit personality any one of those things like the fact that they wrote you know, ambitious lever in their work that, you know, the senior sales manager at a particular company, individually, no friends, when you look at it, you know, hundreds or millions of those data points. So in our algorithm are going to say, all right, one of these books that finding a resume or a LinkedIn profile or somebody has some kind of associated personality consistently, you put all of those again, you can get a pretty good pretty, you know, pretty good and you can actually visualize that on full and it sits on the disk map, big presence. So if we can get enough with a little we can kind of pinpoint where some most likely to be on that. That’s all just based on all of those little figure writing sample, Matthew, what are big, cool, disrespecting, and not finding the trend. Statistically what’s most likely

John Sumser 6:54
the underlying theory is that you can do a correlation of some kind between two sets of data and and arrive at some level of certainty that there’s a match between this is a correlation game. In other words, is that right?

Drew D’Agostino 7:11
That’s where the platform comes in, is we get it over 1000 people a day signing up and filling out personality assessments. You know, another thing goes before because we know we have all those valid, that the more the more we get, the more accurate our our training gets.

John Sumser 7:25
That’s interesting. So do you have a sense of how many people use resume writers to write the resumes and if I’ve hired a resume writer to write my resume, does it still reflect my personality?

Drew D’Agostino 7:36
We actually encounter this more often with people that are, you know, executives that have others, right, their LinkedIn profiles, but kind of more common use cases where it certainly affects it. However, it doesn’t invalidate it. It’ll pick up something from the authors because it because what you write is also just as important sometimes is how you write it. So in many cases, the watch will be the same, but the house will be different, you know, different phrases are used, and that’s really not like necessarily a human thing. We pick up But when you look at lots and lots of examples finishing can pick up on them. Certainly had an impact on Yeah.

John Sumser 8:06
Okay, so how likely is this to make a mistake? If I rely on crystal to make judgments about people in the hiring process? Am I stuck with an 80% likelihood that the judgment is going to be correct? Or is there some way to say if you want to apply for this job, you must go through crystal?

Drew D’Agostino 8:23
Well, I mean, that’s where we have to kind of draw a line of recommendation and guidance about how to use critical. So we don’t recommend that hiring decisions are made. predictions are really for the very top of the funnel. That’s not when you’re filtering out candidates more for communicating with candidates and setting meetings, understanding how people might relate to each other, where you know, accuracy is helpful, but it’s not critical for the system. Once you start getting through the funnel and you’re interviewing and you’re going through the process. That’s where we recommend hiring managers use the assessment because the assessments have a much higher accuracy. They’re much closer like 97%. So when you’re when you’re dealing with real hiring decisions is trying to map skills to killer roles and understand where people are Might be trusted, energized, deserve a tired physically and you also have the ability to get them fill out a short assessment because they’re already participating. And it’s just a lot, it’s a lot more accurate. And that’s kind of how we advise people when you go down the hiring funnel.

John Sumser 9:14
I wonder Have you ever run this to Employment Lawyers, the idea that the top of the funnel can be sifted have an 80% likelihood of success that seems Cavalier.

Drew D’Agostino 9:25
So the purpose of the predictions is not to drift out or filter out candidates. The purpose is for the communication with those candidates and then understanding the relational dynamics between those candidates and your team. That’s not it’s not necessarily for you know, saying, Hey, we need a good person for this ball and then rely on that.

John Sumser 9:41
Well, but if you tell me that I should talk to who in the following way you make recommendations that I should talk to do the following way and they’re 80% likely to be right at the wrong then I’ve given unfriendly face to do, which increases the likelihood that he’ll select out.

Drew D’Agostino 9:57
Oh, God, that’s actually really quite but the all alternative is not 100%, the alternative is actually about 25%, you know that they happen to be the same communication style of view, and you’re kind of going in blind with that. So that’s certainly a valid point. But where you’re coming from is kind of a position of not having any guidance. So going up from, you know, call it about 25% chance of being right to an 80% test is a big improvement.

John Sumser 10:19
But interesting, I think the challenge is that in the old way, there’s no data and the new way, every bit of the stuff that you do produces evidence, right. And so it may be that it’s better, but it’s also if it’s my mother, who’s on the short end of the stick here, and now there’s evidence there’s something for a lawyer to bite into that there wasn’t before. And so I’m not sure that an increased success rate means that you are viable because you’ve got this new pile of evidence all of a sudden,

Drew D’Agostino 10:52
yeah, so in an ultimate, like things like the lesson procedures, I mean, we’ve kind of follow the basic guidelines published by me co opted in Finishing out, you know, validation. And that’s kind of where our own I mean, we definitely want to layers and figured out how to best apply this in the talent and HR market because very different than sales, you know, sales marketing different story where you kind of throw better data, it’s always, you know, more data is always better. And inaccuracy is not as critical, because you’re dealing with percentages anyway, has people who responded, definitely run through that. And that’s kind of how we bifurcated those parts of our. So I mean, we Yeah, totally. That’s why you’ll see whenever we whenever we publish anything about how to use crystal in this way, it’s very focused on the assessment when you’re actually making the decision or filtering out candidates, because the reality is that AI is not it’s not perfect, therefore, you shouldn’t be making those kinds of decisions with it. Yeah. So it’s more of a product than a legal question at this point. Because the I guess, the whole world is way more established. And you know, when you have like reliability and validity studies and assessment, ensuring that there’s no bias and things like that, it’s a lot easier to use and less of a gray area. That’s gonna be

John Sumser 12:02
sure if you’ve got people who are who have taken the assessment. That’s exactly right. But we were talking about the point at which it’s strictly a set of AI recommendations. And that’s just as an interesting thing. We could explore that for a long time. So I don’t have a sense of how long you’ve been actually in the employment space. But it seems like you’re making a new splash about being in the employment space. What’s your general experience? How long have you been talking to people in recruiting? And what’s the customer base of Christ do? What’s your what’s your experience there?

Drew D’Agostino 12:31
I don’t think recruiting Well, I’ll put it all under talent. We’ve kind of always been primarily in sales, that is sales and marketing that was and when when we often don’t really intentionally jump into markets. We’re kind of a generalized personality platform. And then we find that there are certain people that just sign up and gravitate towards it, but we never really made it that much of an intentional decision to go into like HR just kind of happened. It is the second largest of our target customers. Its sales is the largest and then HR and management are kind of like between two and three for the very close. So far as the customer base goes, we’d like recruiting a lot, because they tend to be using communication product quite a bit, you know, like they’re using our communication tools for reaching out and conducting interviews and for preparing candidates and things like that. But then they’re also using the assessment, using advanced sports and right understand how to build teams well, so that’s why recruiting is good, sweet spot for us. But to answer your original question, I mean, recruiters and HR, people in different functions have always been attracted and gravitate towards critical. So really, the beginning, but we have never really, you know, we haven’t really necessarily need a push into it. It’s more just more and more of them have adopted the platform

John Sumser 13:40
could have it. Okay, so you’re not setting up a separate recruiting division, I guess, is what you say.

Drew D’Agostino 13:45
We try to make it a pretty general platform that is useful for a lot of different roles, as opposed to just having specific products or girls. Got it. So what do you think the relationship between jobs and personality is We look at it all in the energy. This is where our book was a lot of a big chunk of our book was written about. So we use this as our framework, which many people in HR and talent are familiar with. And we have a slightly different take on it. So we develop on the personality map. And the idea is that everybody has natural tendencies at work. So when they’re working, there’s a lot and you know, where your sweet spot is, you know, somebody is really energized or at home, working in spreadsheets all day, meanwhile, that might drive another person crazy. So we what we want to do is find out where your natural tendencies are at work, they’ll put you on a point on that personality map. And it might change based on the context. And it’s really important to know, and then there’s an exercise that you can actually do in crystal to point out all of the different responsibilities that you have day to day and look at that on the map. So you might be in the top right of the map but have a whole bunch of responsibilities towards the bottom left of it. And what that means is that think about it like a car, you’ve got to drive all across the map every day, lots and lots of time. And as a result, you can do the things. You can’t do them but you’re going to be using a lot better me losing you Using a lot of fuel, and you’re venturing into run out, then when you really need to get over there, you might run out of fuel and burn out or not have much not habit. And meanwhile, if you’re in a job where you’re mostly thing close to your home, so let’s say you’re on the top right of the map, and you’re staying right around the top right of the map, that way something else arises, you know, like, that is outside of your comfort zone, and you’ve got the fuel to get there and you can adapt, then you’ve got the energy left over to do it. The idea is to understand where you are, be really honest about the responsibilities of the job, not like the procedures on the actual responsibilities day to day. And then as leaders do our best to put people really close to their homes so that they’ve got lots of energy, and then they’re giving them energy instead of draining them all day. That’s the nutshell overview, like our philosophy on career. A lot of invariable that

John Sumser 15:45
that’s interesting. So So you think that a job has a sort of a hard and fast set of parameters that make it more likely to be good for a certain kind of personality versus another kind of break? audience, I wonder if that isn’t the kind of a 20th century view of work where where jobs are these repeatable things and happen like they do in manufacturing? Seems to me that 21st century work is often the intersection of the role and the personality and the personality changes that.

Drew D’Agostino 16:19
That’s a really good point. And I think another another important point about the modern technology economy is that many jobs have all these different aspects where I, you know, I look at my job day to day, there are certain times when I have to be super energetic, visionary kind of person who’s just created and come up with new ideas. And another part where I need to buckle down and dive into budgeting for a day, you know, which is just totally opposite within the same job WB complex. And that’s the crystals not really revealing anything profound in terms of like, new knowledge, just giving things away to actually visualize what’s going on in the job and the role that they, you know, you’re not going to make a stereotype Hey, a salesperson needs to be like this. It’s way more than that. At this company with the people in the room with this boss with this, with this process, they follow, like so many different variables in it. And we’re trying to get a good map of what those actually are.

John Sumser 17:10
So do you have support that helped your users make sense out of this thing that we just talked about, which is that there’s not an inherent relationship between personality type and job, but there’s some information there. That’s useful, right. And I think that I think that’s what you’re saying. And so you don’t want to overemphasize it. And you don’t want to under emphasize it’s a sort of a Goldilocks problem. Do you help people make sense out of that, because what tends to happen in recruiting is things get binary very quickly. either have it or you don’t have it. And so so if there’s an argument that personality type x is good for job, why, then if you don’t have it, you’re out. But what we just sort of stepped through is the reality that the relationship between a job and personality type tends to Be a kind of a fluid thing. So how do you help people make sense out of it?

Drew D’Agostino 18:04
That’s a really good point. So there’s a feature that is in every crystal profile, which about a quarter of our users take advantage of it called and they end up Victor, they’ll end up looking at their job. And when they map out the behaviors, they instead of saying like a binary yes or no, you got it, you don’t actually can show you this is a percentage of the job is likely the drain you think here’s the percentage of a job that is likely to energize you. And here’s the percentage that is kind of meet in the middle to go either way to try to break it down into the individual aspect of other ideas and that as well. Yeah, no job jump into the big one. I think the key is that it’s custom. So it’s not based on a job title. But it’s based on this is pretty unique survey that you’ve kind of come in and you answer a whole bunch of questions about your specific job responsibilities. And based on your answers to those survey questions, it’ll crunch the numbers and say okay, based on your personality, and specific responsibilities tied to your specific job here are the areas that are going to be most energizing here going to be the areas that are most draining. So for example, when drew is having to do a budgeting day crystal can give a really nice visualization of a business why you’re frustrated. And it’s because not that you can’t do budgeting, but budgeting is outside of your natural an energy home. So you’re gonna have to said, spend extra gas to get there and do that. If you’re going to do if you have to do that every day, that might be a good indication that the primary responsibilities of your of your role might be sucking up a lot of energy and as a result might not be such a good job fit.

John Sumser 19:36
The term job fit has often been used as a colloquialism for discrimination. I wonder this is sort of the ethics question. I wonder what you think the big ethics issues are in your work

Drew D’Agostino 19:49
out they have a lot about whenever we’re dealing with decisions about employment, and it was about communication. There’s gonna be a lot of these questions that arise a big one we already covered a little bit in usage in day Now that data is available, what is the proper way to use it? And then baked into that is the question that we also covered before that is, what, what accuracy level is data usable? So there’s two pretty big questions around this. One is doing accurate data. Where did this to you? And to that, you know, the answer is pretty well spelled out. That’s been that’s been around for a long time. And you pretty much have to just make sure that personality and other information does not unfairly correlate with other aspects of someone’s life, if you’re, you know, unfairly filtering out a singular gender or something like that. So that’s why, you know, things need to be validated and why we kind of lean on long accepted assessment frameworks and models instead of developing our own. And then on the other deeper question about accuracy, I kind of felt out a little bit that you need to be honest with yourself and the users of a product about what the accuracy levels are in a product. And if the accuracy levels are here, like for example, in our predictions, 80%. That just means that yes, we’re going to be super transparent about what backers the And also, you know, tell us this is how it should be used and how it should not be used in that case 80% accuracy level, it should be used in cases where you’re communicating or you’re using it as a as an enhancement for a conversation or a good you know, report pair for a meeting or an interview. It should not be used to make an end all decision about you know, whether someone to work somewhere not because the active levels is not there. That’s more about limitations with technology and how man overfitting technology for decision words can be applied yet.

John Sumser 21:26
Awesome. Awesome. So what’s in the roadmap? What do you what’s crystal going to do five years from now?

Drew D’Agostino 21:31
Five years? That’s the big question I’ll start with you in are, we in general have been really focused on delivering more and more insights around your personality. So once someone takes the personality test, we have a prediction and we have that data there is just so much you can do with it. So the profile page, which is where most people spend their time when they’re when they’re in crystal now is just full of all the different tools. So you know, job fit is one of them that launched fairly recently, but there’s lots of other little ways to compare your personality. Again, averages for a job title or other people on your team. There are a lot of tools now that we’ve added to Crystal on to see how you get along with other people on your team and how you can avoid conflict, how you can work better together tips and tricks for managers to delegate the right responsibilities to the right people all based around, you know, energy cost and personality show, our big focus has been adding more and more tools to slice and dice that those personality insights into useful applicable ways. Whether it be for your own personality team building, or on the recruiting side, the more and more insight for sending an email that will resonate more having a more effective initial call, details like that. And I guess on top of that, we’ve also expanded to other mediums. So a big one we’re really excited about is we recently launched a mobile app to people that are on the move can see all their personality data right from their phone. We’ve talked about that for a long time. So very excited that that one is now live to.

John Sumser 22:56
Fantastic, what a great conversation. I really appreciate you guys taking the time to drop By the day, would you take a moment and reintroduce yourselves and tell people how to get in touch with you?

Unknown Speaker 23:06
Absolutely, again, Greg, D’Agostino, cofounders of Crystal, and best get in touch with us, has all the details about Crystal and personality, lots of free resources. And then of course, our book coming out in four days now Predicting Personality, which you can find anywhere where books are sold.

John Sumser 23:25
Fantastic. Thanks for taking the time to do this, guys. It’s been a great conversation. You’ve been listening to HRExaminer’s Executive Conversations. We’ve been talking with Drew D’Agostino and Greg Skloot, soon to be published authors of predicting personality and the cofounders of Crystal Nnows. Thanks for tuning in today. And thanks again, guys. We’ll see you all back here next week. Bye Bye now.