HRExaminer Radio is a weekly show devoted to Recruiting and Recruiting Technology airing live on Friday’s at 11AM Pacific
HRExaminer Radio
Guest: Ravi Mikkelsen
Episode: 142
Air Date: January 8, 2016
Ravi loves using data and technology to make life better for people, specifically in the areas of the workplace, clean energy, and wellness. As the founding CEO of jobFig, he has shaped how they used behavioral data for their clients and led the shift to their new API for personality, allowing companies and researchers to more quickly discover correlations between a person’s actions and their behavioral traits. He has been involved with small businesses and startups for over ten years, and was recently awarded a small development grant from the US Department of Energy for software to reduce the cost of residential solar installations.
When he isn’t developing new business ideas he likes to spend his time outdoors – usually in the woods or on the water. Ravi received a bachelors degree with honors in Materials Science and Engineering from the University of Washington where he researched new materials for energy production.
Guest Website and Social
- Subscribe in iTunes
- Downloadable MP3 File HRExaminer Radio Episode 142: Ravi Mikkelsen
Transcript
Begin transcript
John Sumser:
Good morning. Welcome to HRExaminer Radio. I’m your host, John Sumser. We’re coming to you live from beautiful Occidental, California where it is pitch black. We’ve got a day-long break from the monsoon that’s been consuming California for the last week or so. Today we’re going to be talking with Ravi Mikkelsen who is the founding CEO of jobFig. Ravi, how are you?
Ravi Mikkelsen:
I’m doing great John, thanks. How are you?
John Sumser:
All right. You sound great. Why don’t you take a moment and introduce yourself to the audience?
Ravi Mikkelsen:
I’m a relatively recent transplant to the greater Silicon Valley area, San Francisco Bay, from Seattle. Before that New York, and before that India. I have spent half of my college … Well, most of my college career and then all of my professional career working on problems that affect humanity, whether that’s clean energy, better nutrition, or how to work better. I like using data and science to improve the world.
John Sumser:
Fun, and how’s life in the Bay area? If you’re a recent transplant, it means that you’re living in a cardboard box somewhere, because the housing costs are so extraordinary.
Ravi Mikkelsen:
Well, it’s relatively recent. It was a few years ago that I moved down from Seattle, and I do have … Now I’m in an apartment in downtown Oakland, which is beautiful and right in the heart of things; restaurants and theaters and a beautiful lake and nature within the city, and then 45 minutes you’re up in the hills and more state parks and things like that.
John Sumser:
So Oakland. A lot of people say that Oakland is the best part of downtown San Francisco. That’s the story from the folks in the know. How did you get to be s startup CEO with a company that does automated personality testing? How did that happen?
Ravi Mikkelsen:
Yeah. Well, it actually started when I moved down here in 2011, and I was pitching an energy modeling software that I had built in Seattle, and 2011 was a terrible time for anything. If you said, “clean tech,” all the response you got was, “Solyndra!” VCs and everyone just ran away from you. The other thing … One thing that I noticed in doing all of the different [sounder 00:03:18] matching programs was that many of them ignored the technical filters that I’d set, but then without doubt all of them ignored the behavioral aspects of it. All of these people that I would and say, “Hey, can we start a company together?” It would be huge mismatches.
Then I started looking a little bit in to the research and saw that there was some good academic research behind pairing people based on behavioral traits, and then how a lot of these companies started. I told my friends that I wanted to do this, I got introductions, and I met Dillon, who’s my co-founder of jobFig. Within about five minutes, we knew that we were going to be great friends and could work well together. I’d pitched him on this idea, and he was well aware of the problem and had been experiencing it himself as well, and so we were like, “Let’s do this!”
Silicon Valley love story. Two weeks after we met, we moved in together, he quit his job at a well-paying tech company, and we got going diving into academic research and doing interviews to build the data we needed to build a team set algorithm and get started.
John Sumser:
It’s interesting. It’s a … I find it fascinating that there are a significant number of people in the world who believe that you can tell something about another person’s personality and behavior with a few questions. I know … I heard a great story the other day which is about the person who designed the test to see whether Air Force pilots or people who aspired to be Air Force pilots were going to make it through flight training. As it turned out, there was one question that would predict whether or not you would make it through flight training in the Air Force, and that one question was, “How long have you wanted to fly?” People who wanted to fly since they were 5 years old always make it through the Air Force flight training program.
There are things that can give you insight fairly quickly, but the idea that the complex thing that is a human being’s personality and behavior can be disclosed with a handful of questions, it boggled my mind in the shallowness of the concept.
Ravi Mikkelsen:
Right. I think you’ll like this quote, and it’s one that I live by, is that, “All models are wrong, but some are useful.” When we do these surveys, they are all wrong. Some of them are more accurate than others, some of them are more useful than others, and our job as scientists and as providers of these assessments is to constantly work to making them more accurate and more useful. When you say … There are personalities assessments that are just five or ten questions or statements, in our case, but there are also ones that are really long. We give 120 items which is the minimum number you need to get to the sub-trait analysis, which is what we go for. If you’re just doing a handful of questions, you can get a more gross picture of a person which is, as you said, extremely complex.
The academic standard is the Big 5, or 5 Factor Model, which includes openness to new experiences, conscientiousness, extroversion, agreeableness, and neuroticism. Those five make an acronym called OCEAN. Well, it’s not really an acronym, so the OCEAN of personality. Within each of those 5 there are 6 sub-traits, so you have 30 sub-traits in total. There are some scientists that say there are even more than that, there’s even greater subtlety and complexity. Then you get into the question do you need to go beyond that to have a useful model of someone’s personality that you can then describe to you, help them understand themselves better, to have a better work environment.
John Sumser:
I’m going to drill at that a little bit more. The idea that you can … 120 questions with multiple choice answers, fundamentally. Scenes like … Well, let me back up a second and say what I’ve noticed is that the scientific method is being beaten to death right now, because in it’s origins, the scientific method was a [inaudible 00:09:26] approach. Data was so expensive for the last 500 years that if you were going to collect data, what you needed to do was be extremely precise so that you only collected the data that you absolutely had to have to do the minimum amount of work required, because data was expensive.
Now, data is free. Data is cheap. Data is easy, and we don’t need that discipline on the front end. What we’re doing culturally is learning how to look at data that’s generated from a million sources to get to answers that are vastly more nuanced than you could get with the traditional scientific method. We’re in a time where things are changing, and it’s becoming clear across the board that the answers produced by traditional academia are shallow and rooted in the expense of finding data.
When I hear that you can diagnose a personality with 120 questions, or an hour, or a day, or a week, I think, “Holy Moley! What idiots have we got here? Have they never met people? Have they never interacted with people? Have they never understood the things that happen in a life that radically change people?” It just … It strikes me as so mechanistic that to be on the edges of dangerous. What do you think?
Ravi Mikkelsen:
I agree with a lot of that. One, the very first one about the cost of data acquisition, in general I would say spot on, but in this area I would say good data is still expensive, even though the amount of data that we collect in 5 to 7 minutes on our assessment usually takes academicians 45 to 60 minutes because they use paper and pencil or some other web form that isn’t as well designed as ours. Getting somebody just in their day to day to complete a 60 minute assessment rather than a 30 second BuzzFeed quiz is very difficult, and so we’ve been working our way, “How do you get people to do this?” We cut out 80% of the time through your design and meta-cognitive focusing techniques.
Then also the predicting … There are several tools out there now that want to predict personality from other things, Twitter, Facebook, and there have been some good correlations. In Social Sciences, you don’t need a very high correlation for it to be a good correlation, which doesn’t seem …
John Sumser:
It’s another piece of evidence that the cost of data drives a lot of it.
Ravi Mikkelsen:
Yeah. At least the finer … For a lot of job performance [inaudible 00:13:22] we actually don’t need to ask all of … Present all of the statements. We use statements instead of questions, and then the person agrees, self-agrees, with those statements about themselves. “I enjoy surprise parties. I get angry easily. I would never go skydiving.” For a workplace situation, we don’t need to present all of those to make predictions about different work functions of team [inaudible 00:14:02], but to your point, we are leaving out aspects of a person which may have broader implications and greater value for them in other regards. We’ve kept it all in. It’s not directly useful in this HR sense right now, but as somebody mentioned in our Facebook conversation the other day, is that it’s like the consumer genetic testing. It’s interesting, but what do we do with it? It’s there, and now once everybody has …
Our goal is that everybody in the world will have their behavioral resume, just like they have a professional resume, but with that behavioral resume you can use it not just for getting the best job for yourself, but you can also get better movie recommendations from Netflix, better book recommendations from Amazon. You can get a more personalized life experience by having companies serve you based on who you are rather than as a general average of the entire population.
John Sumser:
It’s an an interesting thing. So I’m going to ask you one more deep academic question, academic-ish question. Then we’ll look out. I’ve been thinking a lot lately about what happens when people join and un-join organizations, and I’m pretty impressed with the human capacity to adapt. It seems to me that people are of a personality and behavioral level very chameleon-like, and that the thing that happens when two people connect is that the basic personality morphs in some ways to make the new organism that’s composed of two or more people, healthy and functioning, and that we do that very, very quickly, and that that’s something that has been beyond the capacity of the sort of testing that you’re talking about, which focuses almost exclusively on individuals, and then a model of organizations that’s like adding a bunch of individuals to a pile. I wonder if you’d talk a little bit about how you think about testing the organism that people go into and come out of as part of the work that you’re doing.
Ravi Mikkelsen:
Yeah. So there’s a couple ways. One is we can get a huge organization, like a Google size at 40,000 people, or an HP at 300. That would be so fantastic, because as you said, most of the time … I would say nearly all of the time … Is that we can change our general behaviors to work better, that tribal … That survival mechanism from thousands of years ago is that hey, we need to work well with these people that are around us so that the wolves don’t get us and we don’t freeze. We moderate ourselves, but there are instances.
Several recruiters that I’ve talked to, especially they say, “Well, how do I predict that start that I’ve hired,” their person’s a rock star, “brought them into our organization. They did horribly, and then they went to a competitor and they were a rock star again. What happened?” In that situation, I didn’t … We weren’t analyzing things from before they came and everything wasn’t a controlled experiment, so I can’t say for certain but most likely, that was a behavioral conflict either within the team or a manager or a general cultural thing that is even more ambiguous to define and study.
It’s a … A mound of data and time to spend on it is what we would need to study a giant organism like that, because within the different departments, the different roles, the different sub-teams within the department, they’re all going to be slightly different in the power structures, how they work together, overall levels, trade demonstration on certain … On the specific trades. With our tool, each person … Everyone owns their data. That’s one difference about us, is that when somebody completes the assessment and gets their behavioral resume, that’s theirs. Then if they’re applying to a job, they grant access to that company. If they work for the company, they grant access to that company.
Similarly to how when you log in using your Facebook login to Instagram or something else, you’re granting that application access to parts of your Facebook data. If you no longer want to use it, you revoke access. That’s our thinking about who should own the data, is the individual, and then if they do it multiple times, each of those instances of that data are there so we can see, “Okay, how did they demonstrate themselves when they entered this company? How about a year later? Two years later?” They can track changes, if there are any, to their self-described personality within that time frame, and then that’s how, a very long-winded way to say how we might be able to study the organism and how the behaviors change within a group.
John Sumser:
That’s great. It’s a fertile ground for conversation. You’ve talked a little bit about how jobFig is focused and who owns the data. What else makes the company different from its competitors?
Ravi Mikkelsen:
Yes. Well, one is that we’ve … Although we still have our TeamFit algorithm available and the [lulFit 00:21:28] stuff that we’re doing, we’re switching focus to being more of a research tool. One of the primary differences right from the get-go is that we left our assessment very vanilla in terms of the statements and the correlations to the traits. We took an open source multi-academic is how the Big 5 assessment called the IPIP, and it was very structured. Rather than say, “Okay, let’s do these statements, or if somebody has a collection of these traits we’ll call them a maverick or a creator or a warrior,” because now you’re creating a model of a model. Every time you take those averages, you get further from the truth.
We want to be as close to the truth as possible. As I said in the beginning, all models are wrong, but some are useful. How do we be as least wrong as possible? I think that’s our primary differentiator to our competitors, and now [inaudible 00:23:00] academic research market. That’s more of … Since most of them don’t have budget, that’s going to be free use for them, but for our … Jumping in to who we actually want to sell this to are those organizations that have already begun to do people analytics. They track how often you tag in to the cafeteria and into the company gym. Google and these other companies that track everything, a lot of them don’t have any sort of personality analytics. They’re not saying … They’re looking at the results of behavior. They’re looking at the what did you do, what did you accomplish, what was the task that you performed, but they’re not looking at the behavioral drivers of that.
I want to get in and sell our tool to them so that they can actually see the correlations between personality trait demonstration and employee actions. I think one result of this will be that we’ll stop hiring managers from doing this really stupid test, crumpling up a piece of paper and putting it on the ground or the corner of the table to see if the interview … In the job interview, to see if that interviewee will take up that crumpled up piece of paper and throw it in the recycling bin. They do this as a test of conscientiousness, and of the Big 5 traits, conscientiousness has the highest correlation to overall job performance across industries and roles. The only problem is they’re testing specifically for orderliness, and there’s a lot of brilliant people and great workers who have messy desks. If somebody’s comfortable with a little bit of disorder but they’re great, that’s going to create a false positive. “Oh, let’s not hire them because they’re not going to be a good worker.” No, they’re just comfortable with disorder.
Having this will allow them to see which of our interviewers … Another thing is self-selection. A lot of interviewers put forward people for hiring that look like themselves. They all think, “That person’s going to be awesome!” Well, because they’re looking in the mirror behaviorally. Yeah.
John Sumser:
What a rich world you inhabit. What an amazing and rich world you inhabit. You’ve kind of blown through the time, and I didn’t really give you enough opportunity to talk about your company. What do you want to-
Ravi Mikkelsen:
It’s such a huge, deep, rich, thick topic. We could talk for hours on this. The human experience, it’s so vast and rich to … As you said, to boil it down to 30 traits, it’s not really fair, and it’s one aspect-
John Sumser:
What do you want people to take away from this conversation? If you could have your wish and somebody’s listened to this all the way up to now, what are the two or three things you want them to remember?
Ravi Mikkelsen:
First, remember this quote. “All models are wrong, but some are useful.” If you’re using a predictive tool, it’s not the truth, but it could be close and it could be helpful to you. Use predictive analytic tools for hiring as an input. Don’t use it as a deciding factor. Three, everyone should have their own behavioral resume, in my opinion. Why not? It’s free, it’s fun, and you learn more about yourself. Then join in on the conversation at jobFig or at RaviMikkelsen, my name, on Twitter. Let’s use data to make this world a better place. That’s very qualitative, but yeah. A more humane, a more understanding place. Let’s not fire people for misunderstandings when we could switch them to a better fitting team or a better fitting role. Let’s get better movie recommendations. Let’s have a more personalized experience and understand that everybody’s experience is different from our own.
John Sumser:
That’s great. Take a moment and reintroduce yourself, and remind people how they might get in touch with you.
Ravi Mikkelsen:
All right. Again, I’m Ravi Mikkelsen, the founding CEO of jobFig. We are a behavioral data company. You can get in touch with me either Ravi@jobfig.com, or at RaviMikkelsen on Twitter, and at jobFig for our company. Always willing to talk and debate.
John Sumser:
Great. Thanks so much for taking the time to do this, Ravi.
Ravi Mikkelsen:
Thank you.
John Sumser:
It’s been a good, good conversation. You’ve been listening to HRExaminer Radio. I’m John Sumser, your host, and we’ve been talking with Ravi Mikkelsen, who is the Founding CEO of jobFig. JobFig, at jobFig.com is a place you can go and get something that they’re calling a behavioral resume, which is a sort of publicly-available version of the personality test that you take. It’s worth a look.
Thanks for tuning in. It’s been great to have you with us. Enjoy the weekend. I hope it’s bright and sunny where you are, and thanks again, Ravi. It’s been great.
Ravi Mikkelsen:
Thank you, John.
End transcript
