I remote-chat this afternoon with Stefan van Tulder, founder of Talent Data Labs and one of the best brains I know in the field of Behavioural Science, Quantitative Psychology and Psychometrics. He’s born in Amsterdam and is currently home-bound in his lovely apartment in the historic museum quarter.
Hello, Stefan!
Hi Rachel. How are you?
I’m good, we are in a partial lockdown and most shops except groceries, hairdressers and supermarkets are closed till June but we are able to exercise and walk around outside. How’s the home-bound life for you?
Happy to hear you’re coping well. All well and healthy here. For us it’s quite similar, most sport activities are closed and all forms of group entertainment have been shut down until at least June. For someone who is used to travelling all over the place the world suddenly feels very small.
Yeah. How’s your team doing?
They are coping well-ish. It’s hard to really tell over digital interactions only. I think the ones stuck in a small house with young children are most affected but everyone seems equally bored. How about yours?
(=^_^=) I think everyone realises that working from home is harder than it seems! My team is trialling an app which promises therapy help. Are you into any of that?
Oh! That’s great and super interesting. We have recently had discussions about this too. As I said before, it’s very hard to assess the mental state of people over video conference (and we often barely show our faces as we’re screen sharing). How are you going about and doing that fully digitized or human in the process?
It seems completely digitized with mini ”rescue sessions”, to help in areas like : ’to feel connected’.
Good! In my experience, people like the option of having a “human option”.
I was thinking that therapy can’t be productized.
Are you basing it on any major psychological framework?
They use Big Five, and we have the option of choosing a learning pathway. I’m not sure if this is a major framework?
I think we live in a world where we are increasing the degree of self-assessment and self-help through well-being apps at the fastest pace ever. I think therapy would certainly be a great angle to introduce. As far as I’m concerned, the Big Five framework is the most solid psychological framework that I have seen.
Yeah, you were working on self-assessment as a service, right?
Yes, not sure if SAaaS is a very sticky acronym but we’re going for it!
Okay, but how would I know what the score represents and how does it help me ‘stay connected’?
Well, the way I see it, there’s two types of players. Those who make an app that does something really well and then stop. Or those that build a large framework of integrations and behavior around it. We don’t just stop at self-assessment, we add recommendations to third party solutions, or give concrete advice that people can follow. I’d say that’s key to making a useful self-assessment app.
I see. I was trying out a few ‘be your best self’ apps involving meditation and fitness. I found that one of them collects my recorded data on steps, sleep, and sex...
Oh really? How do they collect sex data?
I’m guessing from the Apple Watch? It could be presumed from the heart rate tracker or bodily movements, right?
That could be from playing video games in bed. So you’d wear the watch, during... ?
I wouldn’t, I don’t like to track anything (Plus I don’t ever wear a watch.)
It is interesting data, I think sex is such a primal urge that if you don’t get your share you might need help. We see that a lot now in these isolated times.
Maybe it is related to psychosis.
Everything essentially is tied to evolutionary behavior. If it made your gene pool survive, you’d build systematic chemicals for it in your body.
So, mainly, it’s testosterone (or the lack of) sending us off into chemical (ie hormone) imbalances? I don’t know how this kind of data collection would help.
I don’t think we know enough about the impact of chemical imbalances on behavior to justifiably say that it leads to depression but we do know a lot about what kinds of personalities and behaviours are exhibited by mentally healthy people versus the more depressed. A good sample of that is the use of alcohol and drugs.
I know we agree on how (work) culture also contributes to depression.
A concept that I find very appealing is growth. The gray area between doing things that are too easy and too hard. This also counts for environments and cultures. So looking at work culture can be a great aid at finding drivers for disengagement or passivity. They are often signs related to depression. You have dabbled in the use of cultural fit mapping before. Did you find it could justify why some individuals would be able to function better than others?
Yes, most individuals working in startups often prefer an environment that would give themselves the chance to grow faster in a shorter amount of time compared to conventional workplaces, and that environment should provide tools for them to thrive - strong, sound mentorship for development, autonomy to lead projects, sense of ownership. A bad decision for such individuals could be simply putting someone above him whom he could outgrow in 6 months, and would ultimately lead to less job satisfaction. But often those analyses are done in hindsight...
Well, that’s because most players have chosen a passive approach to these types of analytics. The key to self-assessment is to have the data first and the behavior later. The sample you provided is very relatable and also very measurable (with the right tools). Intellect plays a big part but so does culture and personality. So let’s say in your sample you would be able to predict (with relatively high probability) that the individual will outsmart their superior within a matter of months. You would have a few options such as thinning the lines between leadership and guidance/mentorship or enabling horizontal movement at an equal pace.
Yes and this also leads to right-sizing of the company which is imperative in these times, well, most times, actually.
Now, most companies do not sample their full workforce and can never really provide accurate predictions at such a granular level. That’s mostly because the workforce itself is poorly incentivized to use such an app. The key here is to incentivize the user. Typically, you want to decouple from the app and its usage entirely and focus on insight to the user, first. Our app would literally recommend an employee to leave their employer if the chance that they will ever be happy in that culture is too low. That’s not a place where most app providers want to be, as it’s hard to sell something like that.
That’s interesting, and I agree, most people would rather make their own discovery on this matter (and a whole lot of other things), than relying on what an app tells them to do.
We’re not in the business of making money, we’re Scientists with a company and want to make good predictions. Those two don’t always tango nicely.
I wanted to ask about how a Data Scientist like yourself employs the typical SaaS framework into self-assessment. You also work with Psychologists, and Neural Scientists?
Yes, but the Psychologist isn’t part of the app per se. The way it works is best explained through a case. Let’s say you are a lawyer in Japan for a prestigious firm but you aren’t very emotionally stable and lack the mental energy to do repetitive tasks structurally and precisely. This profile could most likely wear out in such an environment in a very short time. The software will assess all those features and benchmarks and prompt you to find some help. This is where you can directly ship your profile to one of the Psychologists for either a chat, a video call, or - when restrictions reduce, a session. We have a dedicated group of Psychologists connected right now but later want to create a larger infrastructure for it. This leaves all the power and all the data with the user until the very end. That’s also where we would make some money, we charge a very small fee for a user recommendation to any third party.
And no one else is doing anything like this?
By default I assume there’s 3-5 other parties doing something similar at any point in time. But not at a scale that’s relevant, I presume. Have you seen any? We are not aiming to become a “feature”, we are aiming to become the biggest database with monitored but anonymous behavior and psychometrics, so that our recommendation system is infallible. This is an extremely expensive and intense journey. Asking some of the brightest Neuropsychologists and Data Scientists to work on some of the most secure and safe software. This is a very long play, and we’re intrinsically motivated to play it, at scale. That’s probably why we don’t see many competitors.
From a Data Scientist standpoint it should be a totally useful and interesting project, (and I think what you are building has a really relevant approach for the future of work and life), but from a CEO perspective, is it too advanced for us normal humanfolk to grasp, much less make a decision on actioning on our behaviors?
You touch one of the biggest pain-points. How do we translate extremely complex and mostly statistical issues into digestible (and hopefully fun) advice without alienating users. To tackle it, we work with all our psychometric assessments in isolation and have user feedback embedded in our platform. We also use reinforcement learning to guarantee results (basically looping the outcomes back into the statistics/recommendations). Luckily we also have a large community of enthusiasts who love learning more about their profile and will tell us whether something is hard to understand, out of place, or just poorly designed in general. From a behavioural perspective it’s no easy feat, either. Getting people to actually do something is infinitely harder than to read about doing something. But this is talking about human laziness, very little can be done to get us off the couch and to stop watching television.
I do sleep more when I’m at home but I attribute it to the cosy feels of my space. It seems you have your work cut out for you, then.
I’m personally obsessed with self-assessment. In the early days of Facebook they would have those quizzes like ‘what type of fruit are you’, or ‘which Disney princess would you be?’, and I would always ship all my data just to find out. A lot of websites have quick assessment tools these days or political party picker apps. The phenomenal popularity of them is inspiring. In contrast, asking people on the street what they vote rarely matches the final outcome of the elections. So truthfulness to other people seems unlikely, whereas to apps, is becoming more mainstream.
You would be Belle. Lol!
For the last 10 years we have, en masse, offloaded all our data on social media apps and publicly created a legacy of information. I think most of us have given up on protecting themselves from the internet entirely. I gotta move, this was an interesting talk!
Thanks! Lots to think about! Hopefully we will see each other again before the year is over!
Yes, I certainly hope so too. It’s been a great chat. Thank you Rachel.