How Google nailed innovation in the mid-2000s | Vaughn Tan, University College London

Mana Labs Co-Founder Mimi Nguyen is in the host’s seat this week as she discusses product innovation with University College London’s Vaughn Tan.

https://apple.co/3oTzwQK

https://spoti.fi/2Lzv4rP

And Vaughn has a unique perspective on the subject. In 2005, he joined Google as a Product Marketing Manager and later worked on projects such as AdWords and Google Maps. Since then, he’s become a leading voice on innovation in the culinary sector, advising organisations like the Wellcome Collection and becoming an Assistant Professor at UCL.

Mimi and Vaughn discuss how Vaughn joined Google without a computer science degree, the leadership skills he learned at Mountain View, and why restaurants are great models for innovation. Plus, Vaughn tells us what companies can learn from messenger RNA and how to embrace uncertainty when hiring.

Check out Vaughn’s new book, The Uncertainty Mindset

 

Episode Highlights:

05:24: The innovative way Google approached product

10:25: How to lead without institutional authority

14:20: How efficiency can be the enemy of success

19:00: Spurring creative collaboration in the age of remote working

23:15: Restaurants as biomes for innovation

27:24: The lessons FinTech companies can learn from kitchens

33:23: How to hire with an uncertainty mindset

40:36: The one trait that has helped Vaughn succeed

42:49: What’s next for Vaughn

 

Transcript

Mimi Nguyen 

Hello, Vaughn.

Vaughn Tan 

Hello Mimi, how are you?

Mimi Nguyen 

Good. Good to have you here. Just for the introduction – Vaugh Tan is an author, lecturer and strategic consultant based in London. He currently teaches at University College London, where he teaches strategy and entrepreneurship. In his work, he specializes in designing organizations that are resilient to and benefit from uncertainty. After earning his PhD from Harvard in 2014, Vaughn has gone off onto work with such companies like Google, Hewlett Packard, the Wellcome Collection, and more. His latest book is The Uncertainty Mindset, which offers lessons on resilience and adaptability from some of the world’s best Michelin-starred restaurants. You have three degrees from Harvard so you’re obviously a high achiever. My first question for you is: were you always motivated to succeed in this way and what made you so committed?

Vaughn Tan 

No, I was definitely not always motivated to succeed at graduate school. I think grad school was kind of an accidental thing that I did for probably the wrong reasons. Around 2008 I was getting tired of my job at Google. I was trying to find another thing to do and in the absence of any clear direction forward, the easiest path forward is to go to grad school, and that’s what I decided to do. While I was there, I was very lucky, I ran into advisors and colleagues who were fun to hang out with and let me do whatever I wanted to do, which is why it was possible to get two more degrees from Harvard. I wouldn’t say that this was the plan from the very beginning, or that I was especially good at it, and neither would they. So, not necessarily means-ends oriented in that fashion.

Mimi Nguyen 

Let’s go back to the Google thing. After you finish your bachelor’s degree at Harvard, you went off to Google’s Mountain View campus for what I remember, for three years. How did that opportunity come about?

Vaughn Tan 

Again, it was completely random. I didn’t do computer science and usually if you are being interviewed for the Product Management position you have to be a computer science technical person. But they used to do this thing where they would randomly look at the piles of CVs that you send into the university careers office, and they would randomly pick some people, this is my understanding. I was one of, I think, six or eight people that they just randomly picked out, they called me up and they said ‘Do you want to come interview for a job?’. And I’m like ‘yeah’ because at that time I’d used Google, but didn’t really know very much about it. So I was like ‘Okay, sure, free trip out to California, I’ll go.’ Then I started doing more research and I was like ‘This is kind of cool. I want to get this job.’ So I went out there, I interviewed, the interview process was very long, which back then was famously long, it took a long time to do your interviews, and then get feedback and get a decision about whether or not you’re hired. But for no reason that I can tell I got the job. And then I was like ‘Great. I’ll definitely do it.’ It was also very much by chance, super lucky to get the opportunity to interview. I don’t think I would have applied. And if I had applied, I don’t think I would have been shortlisted for any kind of interview at all but because they just did it randomly, apparently works out.

Mimi Nguyen 

Why did you apply to product? I think back then, it was very unconventional, I don’t think it existed that much before. What was the product person doing? And as you say, people used to come to Google from computer science and you came from sociology. How did you learn about product roles and Google? I was barely using Google that time.

Vaughn Tan 

So again, I think with hindsight, it looks as though I did a lot more careful thinking about how I did it, but in reality, I didn’t really. So when they called me they said ‘The job that you’re eligible for is this product marketing position, which is part of the product team’ and I was like ‘Sure, I guess I know nothing about it.’ I’ve never even thought about becoming someone interested in product or marketing but I was like ‘If you think I’m eligible for it, you might as well put in my application.’ Then I sent in a little bit more stuff, I guess, I did the interview and it worked out. It was accidental. I think at the time, now that I have hindsight, the cool thing about working in product at Google is, like you said, the way people think about product in tech is different and Google was particularly strange. Because the product team at Google was, in my understanding anyway, particularly empowered. You saw Product Managers at other companies like Microsoft, which had a long history of having product, but at Google there was this weird sense that the Product Manager truly was, or at least they were told that they were, the leaders of a particular product. They would bring together the engineering, the sales engineering, the legal, everyone would be brought together by a Product Manager, who was the lead, who did not have true authority but had this tacit responsibility for making sure that things happened and that products were scoped correctly. It was a very unusual product role and it was also a very unusual place to have a product role because Google was very strange at the time. It was the year that they went public and it was a really exciting place to be. It was very strange.

Mimi Nguyen 

How many products did they have at that time? I recently spoke to someone. He said, when he came to Google, Google Cloud had eight products’ or six, and that now it’s eighty in the Google Cloud Platform. How many were there when you were there? What did you work on when you were at Google?

Vaughn Tan 

When I was there, I could not tell you even close to accurately how many products there were. This was actually comparatively early in Google’s lifetime. I think it was like less than a decade old. And so there were lots of products, and they were trying out lots and lots of things but it wasn’t the weird profusion of products and sub-products that you’ve got to do. So I couldn’t tell you, I don’t want to make something up. What did I work on while I was there? The first product that I was helping with, it was actually something that eventually never became its own product but became part of a lot of other products. We were trying to build an automated ad unit generator and ad keyword targeting system. If you know about the search advertising that Google makes a lot of its money from, the idea is that you design an ad unit that is shown in response to search. Then you also design the keyword terms that that ad unit is shown in response to so you search for a word, it shows an ad, you have to design the list of words as well as the ad. This way of thinking about advertising is actually quite counterintuitive for most people who are not born and bred search engine marketer people. So what we were trying to do was to build a system that would basically look at your website, figure out what you did, and build you an ad unit, as well as infer something about the keyword list that would prompt that ad unit, very challenging. You’ve got the advantage of Google, of having lots and lots of data about what people search for and what they see. But still, I think it was way too early for that. A lot of those things, you can see bits and pieces of them in the ad unit wizards and things that Google still has today. That was one product I worked on. Then I went to work with the geo team, which produces things like Earth and maps, and I worked on maps for a bit. I was part of the team that helped launch Street View. Then I also worked with a group called Google Earth Outreach, which is the nonprofit wing that gives Google Earth licenses for free to nonprofits that are using it, a lot of environmental justice nonprofits. After that, I was working on the Google Lunar X Prize, which is a prize that Google-funded in relation to the X PRIZE Foundation. Its goal was to create a prize driven incentive to put to design and then put an unmanned lunar lander on the Moon. Because the idea was that in doing so, you would create a lot of ancillary technologies that would be very useful like the first space race did. The thing I did before I left, the final thing, I worked on an internal project that was meant to have an external exposure. It was called Fusion Tables and it was meant to be a logical table structure that was arbitrarily long and arbitrarily wide so that you could put all sorts of structured data inside it no matter what it was, and you could find some interesting connections inside. That unstructured data is carried with me since then. It existed before Google and I’ve continued with it after that.

Mimi Nguyen 

You also mentioned that the product person is bringing together software team, tech teams and all people from different kinds of backgrounds. Do you think your degree in social studies helped you?

Vaughn Tan 

It’s a great question. One of the interesting things about working at Google was realizing that there is a lot tacit knowledge involved in managing teams when you have no authority. I was very young when I was there at Google, it was my first job out of university, and there was actually very little training in terms of how to lead without being granted authority. Like if you’ve got lots and lots of people from different parts of the organization that need to be working together on a thing. And you are both helping to define what thing they work on, you’re specifying the requirements of the product, as well as trying to incorporate their insights. Each one of these teams has their own specific expertise, which you need to incorporate into this, and you need to help them explain to the other groups. It’s actually a big social science problem of people with different vocabularies, who need to work together with each other. But it’s actually not very clear to everyone that they are using different words to mean the same thing, or using the same words to mean different things. We’ve actually talked about this particular problem many, many times in other contexts. I think what was interesting about being a social scientist, in that particular position, was realizing very concretely, what a lot of social scientists, especially sociologists, talk about the difference between power and responsibility, and where does power come from. A lot of these questions social sciences think about and then it becomes very clear to you that these are real questions when you are in a position where you have responsibility but you have no power at all. That was very interesting. And then the other thing that was interesting about coming at it from a weird social science background is two things. One of them was realizing that not knowing something is a real gift because a lot of the people who are in this context I was working with, they were supposed to be domain experts and so they could not admit that they had no idea what was going on. When somebody from a different domain was an expert there, was talking about their expertise. But for me, it was always okay to say, ‘you know what, I’m just a sociologist, I don’t even do numbers, you tell me what this really means in normal English’, so that was a huge help. And then the other thing that was really useful and interesting was this idea that if you are in different domains of professional expertise, where you spent a lot of time, your idea of what is valuable to do, and what’s important, may be very different from what someone in a different domain thinks is valuable and important. So if you are part of the Google legal team, your idea of what is valuable and important to do, what is essential to do, will look super different from if you are in the product team or if you’re in engineering, or if you’re in R&D, or if you’re in production engineering. Each one of those people will have a different idea of what it means to make a product successful. Because they’ve got a different idea of what it means for a product to be valuable. Because underneath all of that the values that they think are valuable, are very different from what somebody else in a different group coming from a different background thinks is valuable.

Mimi Nguyen 

Back to this different environment that you mentioned these multidisciplinary people, different jargons different value, also you mentioned the fixation that every discipline has. Is there any other organization structure that Google has, that it makes it so efficient and innovative?

Vaughn Tan 

I think even people at Google would maybe not say that they’re efficient. They might say that they have bits of the organization that are effective so their ad organization is very effective. So are there other things about Google that make it particularly good? When I was there, by the way, I was there so long ago that my take on what it’s like is pretty much no longer accurate. So when I was there, I think one of the interesting things about it, which probably isn’t still true, is that it used to give you a lot of space, because it was not efficient in the conventional sense. It used to give people who wanted it, a lot of space to just go find other weirdos to try weird things out. I think of this as this idea that if you have an organization that is at maximum optimization it is super-efficient. There is no room for slack, there is no wasted product, there is no wasted resource, everything is fully utilized and so there is no kind of strange place where you are able to do things where the outcome is not known because you just don’t have any time resource or energy for it. This is what happens in a fully optimized, highly efficient system and Google was not like that. Google was, actually, if you had to ask me at the time, it was very inefficient because it would let people like me spend a few hours every few days, just randomly talking to people that I shouldn’t really be talking to based on what my ostensible job is. The value of that is that then people who aren’t supposed to talk to each other, meet each other, for instance, and they realize, actually there is a connection. There is this person here who works on structured data standards, there is this other person here who is trying to build a logical table infrastructure for arbitrarily shaped structured data. And the two of them would never even meet or know of their existence if not, for instance, a massage therapist. They used to have massage therapists on-site, maybe they still do, and he happens to have done massage for both of them and know that both of them are interested in similar things, and he connects them. People like this guy who became a good friend, his name is Joel, he’s interesting in his own way. So people like him, bring together connections that would otherwise not be known about, they would be latent. And people like me at the time, I would make it because it was fun, I would go out and try and meet people in various contexts that I wasn’t supposed to be meeting them in professionally, and learn about what they were doing and that’s actually how the last project on Fusion Tables came about. It was just a bunch of weirdos from around the company, that all had an interest in some way in unstructured data. And we found an opportunity to actually build a product that would use a lot of the things that they wanted to build. But for a third party, a nonprofit, that was trying to sequence and barcode every living thing on the planet so that you would have a big database of genetic barcodes for entities. So it was a lot of weird, unexpected outcomes, which I think is what we should be thinking of if we think of innovation. One way that Google at the time used to be very good at innovating was because it was possible for weirdos to find other weirdos and come up with ideas that, the leaders would not necessarily have come up with on their own because it would not be a top-down kind of idea, it would be a bottom-up bubbling idea. And you can’t bottom-up in bubble unless you’ve got the ability to find other things that are down below and connect them together so that they can bubble together.

Mimi Nguyen 

…and the superconnector – the muscles.

Vaughn Tan 

For sure. Superconnector, I think it’s a great term for it. The one that I tried to get Google to use, which I always want companies to use, is this idea of messenger RNA. You’ve got this idea that inside an organization, you’ve got these different silos, and the silos all have lots of valuable information and they want to do things but usually, the big problem with silos is that they’re not connected to other silos and they’re not connected in a deep and fundamental enough way that they can really exchange ideas that allow them to connect in meaningful product outcome related ways. But if what you did is you had a bunch of people whose job 20%, 30% of the time was just to go around and find out what the other silos are doing and make connections, they would be like messenger RNA. They bring information from place to place in a way that the information sources would not be able to do on their own. It’s relatively cheap, you already hire them, they’re probably going to be goofing off a little bit of time anyway, you might as well direct the goofing off in a direction that makes sense for the organization and makes sense for the people. If they want to do something that is fun, which is to meet other people who are diverse and interesting, give them a way to do that in a way that makes the company more successful by allowing you to discover things about itself that you otherwise wouldn’t be able to discover.

Mimi Nguyen 

How can we keep this approach or this mindset, whilst working remote?

Vaughn Tan 

Great question. The answer is obviously contingent and complicated. If the work that you do happens to be the kind of work that requires physical presence, if you are cooking in a kitchen, like a team, so I study restaurant teams, and restaurant teams are interesting because if you are trying to be a restaurant team, it’s very hard to remote work a restaurant team. You can’t cook at home and then hand it off. It just doesn’t work. There is lots and lots of things like that. If you’re designing physical objects as part of a team, like you’re at Mercedes or BMW and you’re designing a car, you cannot really do that remotely because there’s a tactility of what you do. For those kinds of work, it’s not clear to me how you can do that remotely but if what you’re doing is in a sense of thought work, you’re writing things or you’re talking about things or you’re doing idea based stuff that can be done virtually, I think this kind of messenger RNA activity is actually easier because what it depends on is not highly structured interactions. What it depends on is really, really small micro-interactions and you can do the discovery that allows you to find the other people with whom you have these very small interactions with things like zoom or WhatsApp or whatever. I don’t know if you have experienced the same thing, but in this pandemic year, I now have more micro-interactions with more people than I’ve ever had before. Because now it comes through short emails from people who I get introduced to. Who now, because there’s no other option, they feel totally okay with sending me an email, or I get put into WhatsApp groups now. They send me an invite link. They’re like ‘hey, you might be interested in this thing, which is about some area that you might be interested in, do you want to join us?’ And this would never have happened before because we used to think that being physically present was somehow, I mean it is obviously somehow very valuable, but it was a barrier and if you didn’t meet in person first, it was harder to think about meeting at all. Now, this year, when you can’t meet at all, it’s like the default position. I might as well give it a shot, if they respond great if they don’t respond so well. So this year, lots and lots of micro-interactions, which at least for me, has been very valuable. And I think you can do this not just with people you don’t know outside of your organization but if you have an organization that you run, you should make it more possible. You should make it easier for people across your organization to find other people and just interact with them in a professional context but without a lot of expectation of long term heavy commitment. It should be ‘hey, go out and find people who might be interested in doing what you are doing or something related, and just have a ten-minute chat about what they’re doing and what you’re doing and if there’s no interest stop, and if there is an interest, set up another one.’ And then if there’s an interest after that set up another one, you know, it’s like, it’s a gradual step up, that allows you to discover whether or not something interesting is potentially going to happen. And if it looks like it might happen, you continue and you continue and continue. And it very naturally becomes either a deep and detailed interaction long term that results in something interesting that you wouldn’t have expected, or it doesn’t. And if it doesn’t, what have you lost? Nothing. If it does happen, what have you gained? Quite a lot, and something that you weren’t expecting in before.

Mimi Nguyen 

I have that experience. I think back then I would be stressed to go to conferences to meet people and nowadays I just email them and we have a call, which is fantastic because I don’t have to travel to States and we can easily just talk to the Tree People around quite easily. But you mentioned the restaurants that you work on recently and the book focuses on the culinary industry. What fascinates you about the sector?

Vaughn Tan 

Honestly, what fascinates me about the sector is that they produce delicious things. So the reason I normally give people when they ask why I study innovation and organization in these kinds of settings, there is a good reason, it’s that if you’re studying innovation, you need to look at innovation outputs. And if you study innovation in conventional places, like semiconductor fabrication, the cycle time before you get to an output is really long, it’s like a year and a half, two years, maybe longer before you go from an idea, for instance, of a new chip to an actual chip. In food, the cycle time can be several hours to several days, maybe several weeks, but you can see many, many iterations. If you’re interested in patterns of how teams work that lead them to be better, be able to innovate or come up with new things what you want to do really, is you want to see many, many, many cycles. You don’t just want to see one cycle. You want to see, like twenty-five, thirty-five, forty-five, seventy-five, a hundred and five cycles so that you can understand what things are patterns and what things are just things that happened. That’s my good, technically sound reason but the real reason, obviously, is fun. I used to work in tech, and I love tech people, but I don’t necessarily want to spend years of my life watching them. I don’t know code, because basically, what they do is they sit at a laptop and they code. So instead, if you study something where the work is embodied, like a lot of the thinking of cooking is not just in your head, but in your hands and not just in your hands but in the interaction between what people do with equipment, and objects with other people that also use equipment and objects. If you’ve ever looked in a restaurant kitchen, one thing that you will notice is that a restaurant kitchen operates by breaking up what people do. Every dish requires three or four or five sets of people to do things separately, either independently or interdependently and then it all gets recombined. For example, you might have one person who does a particular thing to a piece of meat, they prep it in some way, they slice it in some way, they bind it in some way, they hand it off to someone else who then cooks it, and while that person is cooking the meat, the person who’s on sauce is heating up the sauce or preparing a sauce. While that’s happening, someone who’s on garnish is preparing the vegetables that will be garnished and all three of those workstreams come together at the past, which is where everything gets plated, where it gets assembled in the correct order, and the correct configuration gets quality controlled and then gets sent up. All this happens hundreds of times every night in the kitchen or thousands of times in the kitchen. There’s something very interesting about understanding the innovation that happens before you even get to the production stage, which is in the kitchen, but also in how people think together. If you are in a restaurant kitchen, and you’re trying to understand how to make something that has to respond to unpredictable order flow. You never know, unless you are in a completely prefixed kind of situation, what people are going to order and when they’re going to order it. So if you’ve got a dining room of thirty people, and you give them the choice of an a la carte menu, you actually don’t know what they’re going to order over the entire course of the night. So you have to prepare extras of some things, you have to be ready for modifications and adaptations. You also don’t know the sequence in which you need to produce the food that they order, because if table one orders these things, and table two orders some other things, it looks different in terms of your production process from if table one orders a different set of things, and table two orders a different set of things. There’s a lot of in real-time organizational response to some kind of unpredictability, which I find very interesting. So the longer answer to why I choose to study these kinds of settings is because I think they’re really interesting for understanding how you build organizations that can deal with uncertainty of various different kinds, and still be innovative in ways that are delicious.

Mimi Nguyen 

What can these coders, or let’s say tech and finance companies learn from this innovation that happens in the kitchen?

Vaughn Tan 

It’s a great question. I obviously don’t think that companies outside of food can learn from exactly what it is that a kitchen team does to organize the R&D team that will come up with a new dish, or the way they organize the brigade that’s in the kitchen. But I think there is, essentially, one overarching main idea that people outside of food can take away and then there’s three sub-ideas that they can take away as well. I’ll start with the big idea. I think the big idea is that in the part of the fine-dining world that I looked at, which is what I call the cutting edge, or like high end cutting edge cuisine, one of the things that is almost taken for granted, actually, I would say it’s taken for granted now, what’s taken for granted, among the people who run these kinds of restaurants and these kinds of culinary R&D teams, is that the state of the art is changing all the time. What I mean by that is, the kinds of equipment that you use, the availability of ingredients, those are changing all the time, the next big thing is always something different, not something that you know about. Also, what customers want these restauranteurs, and these chefs also realize is changing all the time. So they don’t have this expectation, which I think a lot of businesses have, that in some way the future is predictable. This is the basis of the uncertainty mindset, which is that in the cutting edge of food, for various reasons, the people who work in the cutting edge of food, no longer expect to know with any great precision or any great certainty what they need to be doing in the future. They’re always ready for things to change because this is their lives. Every night, you show up and you don’t really know what you’re going to cook and whether there will be some crazy disaster that happens in the kitchen but you have to deal with it anyway. Also, if you’re developing new dishes, you never know whether the tools you use will always stay the same because you know that every single day, there is some new announcement of some new technique or some new piece of equipment so you expect that things will change in ways that you don’t expect. What that does, is it changes how you act and it changes how you act in these three other ways which are more concrete and smaller, which I think other companies can also learn from. One of them is how you hire. A second one is how you set goals. And the third one is how you motivate. In terms of how you hire, this is something which I think maybe we’ve talked about a lot because you are interested in hiring. One of the main inexplicabilities for me of conventional hiring is this idea that you fully develop your job description, even for jobs that you know you cannot fully define. So you’re going to hire someone who is like a very senior person to run your company in a rapidly changing industry, you have no idea what this person needs to do to make your company successful. And yet you somehow believe that you can write a complete job description that describes what this person will do for the rest of his or her career within the company. To me, this makes no sense. You are trying to hire someone to build your R&D team to, run your R&D team, to be an R&D person, you don’t know what they’re going to make and yet you’re able to write a complete job description that helps you find this person. Absurd, right? It’s brainless, literally. This is the conventional way of hiring where you say ‘we want this person, and these are all the characteristics we want, let’s hire that person.’ And then we hope that they’ll do a good job. In the teams that I was looking at, they don’t bother to do this. What they do is they may have some short list of things that they know this person will have to do, because obviously, they will have to do some things but a lot of the role is left open for mutual definition. And they bring these people in for varying lengths of time when they’re there kind of temporarily, they’re there, provisionally, I guess you would say. During that provisional time, they show that they can do the things that they know they need to do but the large part of their role that is left open-ended, they negotiate, they try something out, they try a small thing that they think might be useful. If that thing is useful for the group, they do more of it. If the thing is not useful for the group, they don’t do it anymore. Over time, if you keep trying these little, what I call role components, you build up a complete role that is partly comprised of things that the organization knows you need to do, as well as things that you have just tried, that you showed are useful to the organization and over time, your role keeps on changing. What this does, is it allows every person’s role to gradually evolve as the needs of the organization evolve, which is exactly what you would expect from an effective organization that is in a situation, which is constantly evolving. That’s the first thing – change the way you hire. Don’t just say ‘here’s my job description, I’m going to hire just for that.’ Instead, leave part of the job open and let people who come into your organization have time to try and fill up that part that’s left open with things that are mutually of interest and valuable to the organization.

Mimi Nguyen 

I think it’s also horrible in terms of these machines matching people’s CV. As you say, having a job spec nowadays, it’s quite tricky, because we can specify the job requirement or the future role that the person will be taking on but yet we still trying to put someone’s CV and with machine learning match word by word, to match the words like if someone has that skill set, which actually will change in the future anyway. I think nowadays, data scientists are using Python but back then they would have SQL in the CV. So, if we just match the keyword, it doesn’t really mix. What should we have in our job specs then? What should we look for when we’re hiring people with this mindset?

Vaughn Tan 

Let’s fully predefined a job description and then hire based on that. Obviously, I’m targeting with too broad a brush. There are lots and lots of especially small companies that look out for people who have kind of meta-level skills rather than concrete skills. So they look for people who have flexibility, who are interested in a lot of things. If I can reframe the question a bit, I would reframe in two directions. One direction is what can you concretely do in terms of a hiring process as a company to make it more likely that you hire these kinds of people? And we can talk about that later. And the other one is, what should you look for in a CV, so that you are more likely to find this kind of person? We’ll talk about that first. So I think looking for in a CV, the more conventional a CV looks, the less likely it is that the person who submitted the CV has some kind of unusualness to them that will make it likely that they have this uncertainty mindset. I think one of the problems is we often outsource CV review to people who do not really understand what it means to have open endedness in terms of what you’re trying to do, or what a job looks like because you outsource it to HR. Often you literally outsource it to people who are HR consultants, who are outside your company in the first place to look at your CVs and nowadays people, like you said, are outsourcing it to machine learning algorithms that will scan the text of a CV and then and then filter out the ones that don’t look very good. What that does, I guess, if you outsource it in either of those ways, you get CVs that look kind of all the same. They will all look good in ways that are well understood to be good looking ways and then you will lose all of the weird outliers. What I’m saying is when you are pre-screening CVs, you should be looking for outliers as well, because outliers are probably where the really interesting people are. They’re almost never going to be in the central core of the distribution so instead of looking for particular words, look for CVs that just look weird, then evaluate those separately and by a human. At the moment, we don’t have artificial intelligence of any real sort yet, we may have machine learning of some sort. But I don’t think machine learning yet is very good at finding signals where even humans can articulate what the signal is. So if you have humans who are probably not junior humans, senior humans who really understand what it means to have to do something, which is not fully programmed to have to respond in real-time, they can look at a CV, and maybe they can see that this person who would not look good on paper normally, maybe is worth having a longer conversation with. That’s about what you look for in a CV. I think another thing that companies can do, and now we’re shading into process, is companies can ask for different kinds of CVs. A lot of companies will just say send in your CV. But if I were hiring for a company, what I might ask is, I might describe my company in a lot more detail than usually companies describe. Then I might ask a person to explain why their CV makes particular sense in the context of what we’ve described to make them do a little bit of this negotiation before they’ve even started. That might be useful and interesting because it causes the CV to contain information that it would not normally contain, which is valuable. But I think the most important thing in terms of process is, we often think of the hiring process as being, you do the recruitment part where you get a lot of potential candidates, and then you interview and then you select, and then you hire full time. Once that person is hired full time to that role, you’ve done the hiring. I think a different way to think about it might be you still have to do the recruiting and all that. But instead of thinking of it as a binary transition, you hire this person in and then the job of hiring stops. Instead, one way to change how things in the company work so that you get people with this uncertainty mindset more is to say, okay, you get hired, and it can be hired full time, you don’t have to do it temporarily, but your role is provisional, for some number of months, three months, two months, six months. You get to choose based on your tolerance. And in those two, or four or six months 25%, 40% of your role, we’re not going to tell you what to do but you have to figure it out, don’t waste time. Go figure out what it is that you want to do that based on your understanding more and more about our company, you think we need to have done, that’s not being done. Just do that, it’s a very simple instruction. And if you’ve been choosing only boring people, you will get boring results. But if you choose people who are genuinely interested in creating a rule for themselves, you may get very, very unusual results that are refreshing, different and valuable. And then you may also push your company in directions that you don’t know.

Mimi Nguyen 

In some companies who can afford to do that I think that’s also a problem because you’re hiring, let’s say I was called an army of newcomers. And then they were all kind of changed. My first job I was hired to an analytics team because that was what my CV represents. That was my bachelor represent. And I think within two months, they moved me to strategic management. They just thought that maybe I fit in that team better.

Vaughn Tan 

One thing to say about that, though, is, it’s really good that in your company, they did that, because they saw that you had something unusual that was not a good fit for your current role, but was maybe a better fit for a different role. What I’m saying what you described happens in many places and that’s what happens when you’ve got a good manager. What I’m saying is to be very explicit, not if you’re lucky, and your manager sees that you’re good at x but not at y they move you from y to x. To say 60% of your role you do these things, please make sure you do them well. But 40% of your role, we are explicitly telling you, we’re not going to tell you what to do. You figure it out and you tell us. It’s the difference between if you’re lucky a manager sees and the manager and you the company and you haven’t an explicit agreement, that there is a part of your role that is not defined which you as the new employee help to define with the idea that you’re not just doing whatever you want. You should do whatever you want that based on you as the new employee understanding more about the company, as you work in it more, you as the new employee think the company needs. There is a lot of conditionality on top of it, that is important to know it’s not accidental. It’s explicit. It’s very explicit and bi-directional on both sides, the employee as well as the company. It’s not just as an employee, do whatever you want. It’s do what you think you’re good at, that you think the company needs that it doesn’t yet do. That’s super important.

Mimi Nguyen 

My last question about you is what’s your manner. What is the one trait that has guided you to success throughout your career?

Vaughn Tan 

It’s a good word. I like that word a lot. I guess the one trait is, it’s also a bad thing, but it’s been useful in some ways. It’s often the case that there is an obviously correct strategic path to take that will be clearly very correct and very valuable. I am very bad at doing that. I’ve got no discipline. I usually do whatever seems to be enjoyable and fun at the time, even if it seems to be strategically unwise so I’ve accidentally done a lot of weird things that have all been in some way nice. I think the trait of being unable to have the discipline to do what appears to be the strategically correct thing is actually itself quite valuable. For me, anyway, has been.

Mimi Nguyen 

I’m surprised. I always thought you’re a disciplined person. You had these three degrees from Harvard.

Vaughn Tan 

I’m disciplined in the sense that if I decide I’m going to do something, I will do my very best to get it done. But I think I’m talking about another kind of discipline, which is when you see that there is a path that everyone says is the correct path to take. After you finish university go become an investment banker or consultant. That kind of discipline where you say, I don’t want to do what I want to do. I got to do what someone else says I got to do. That kind of discipline I don’t have. I’ve never been able to do that, which is probably not the best thing long term if you want to be conventionally successful, which is why not conventionally successful.

Mimi Nguyen 

You didn’t stay in this one company for ten years. Great. Thank you so much. Oh, yeah,  you are researching winemakers right now. What are your future crazy things you’re doing?

Vaughn Tan 

I’m researching a bunch of things. I’ve had a research project on regenerative agriculture for a while, specifically interested in wheat, as well as in grapes because I love wheat, and I love wine. So I’ve got that. I’ve also got a separate set of research projects on the modern data economy. How businesses are now collecting and creating personal data, processing it and selling it in ways that often are, I mean, often we know that they’re doing it, but often what they’re doing and who they’re doing it for how they’re making money from it is very opaque. And the connections between all these different businesses also very opaque. I’m doing some work in that very provisionally. I have no idea what’s going to turn into yet but it’s quite interesting.

Mimi Nguyen 

That sounds exciting. Good luck with your crazy projects. Thank you so much for coming onto Searching for Mana.

Vaughn Tan 

Absolutely. Thanks for having me. It was great.


Also published on Medium.