SDS 033: Building a personal brand in Data Science with senior Insights Manager Josh Coulson

Podcast Guest: Josh Coulson

March 9, 2017

Welcome to episode #033 of the SDS Podcast. Here we go!

Today’s guest is Senior Insights Manager at LinkedIn Josh Coulson
Are you using LinkedIn to find a new role within the space of data science? Or perhaps you are a hiring manager looking for more data science talent?
Josh Coulson leads the team that actually generates the reports that help employers understand the market and will share his top career tips on getting noticed by great employers or talent.
We also explore how the larger amount of data available is changing organisations and whether data scientists will continue to be around in the future.
Join us for this very useful and practical episode!
In this episode you will learn:
Data Use by LinkedIn Insights Team (12:18) 
Implications of Moving into the Gig Economy (22:29) 
Building a Personal Brand in Data Science (26:15) 
Data Democratization Across Organisations (37:11) 
Roadblocks to Organisations Optimizing Their Data (42:25) 
Data Science in the Future (55:06) 
Items mentioned in this podcast:
Follow Josh
Episode transcript

Podcast Transcript

Kirill: This is episode number 33 with Senior Insights Manager at LinkedIn Josh Coulson.

(background music plays)
Welcome to the SuperDataScience podcast. My name is Kirill Eremenko, data science coach and lifestyle entrepreneur. And each week we bring you inspiring people and ideas to help you build your successful career in data science. Thanks for being here today and now let’s make the complex simple.
(background music plays)
Hello and welcome everybody to the SuperDataScience podcast. Super excited about today’s guest. Today we have Josh Coulson, who is my friend back from the Deloitte days. So back in 2012 we met, and we did some work together, and then he moved on to a new role at LinkedIn, and he’s been there ever since, and he’s really grown in the team. He’s actually now a manager there and has a team of his own, and manages a huge part of the insights team of LinkedIn for Asia Pacific, where he helps employers find the right talent for their organisations and understand how to go about it.
So in this podcast, we’re going to be heavily focused on how data scientists, aspiring data scientists and analysts, or people who are already data scientists and analysts, can break into the space and find new jobs, find better jobs, improve their careers, and find opportunities that excite them. Because Josh obviously has access to all of this knowledge on how this hiring actually happens. He hires data scientists himself and he works with lots of companies who are looking for talented people, including data scientists and other professions as well.
So you will get a lot of very valuable information about that, and hopefully that will help you structure your career, or structure your approach to breaking into the space. Because as a lot of you guys shared in the survey that we ran recently, a lot of you are actually open to new opportunities and this was a good chance to ask the right questions because Josh has access to a lot of this information.
And also if you are an executive, or if you’re a manager, or if you’re an entrepreneur, or if you are a business owner, then you’ll find a lot of valuable information here on how to actually source the talent and how the top organisations in the world go about sourcing their talents and how they leverage LinkedIn data to do that.
So can’t wait for you guys to check this podcast out! Let’s get straight into it. Without further ado, I bring to you Josh Coulson, the Senior Insights Manager of LinkedIn in Asia Pacific.
(background music plays)
Hello everybody, welcome to the SuperDataScience podcast. Today I’ve got a very special guest, a good friend of mine, Josh Coulson on the line. Josh works at LinkedIn and he’s calling in from Sydney. Josh, welcome. How are you today?
Josh: Hi Kirill. Thanks for having me. I’m fantastic. It is Monday morning here in Sydney. I’ve already been up and went for a run this morning around the beautiful harbour, so I am ready and raring to go.
Kirill: Oh, fantastic. So you live somewhere nearby the harbour? Or do you run when you get to work?
Josh: It’s actually funny, I’ve been for the last 12 months or so, been trying a bit more to get fit and healthy, as you do when you get a little bit older and have kids, but I’ve been getting up in the morning and coming into the office, because we live about 10 or 12k out of the city, and then going for a run out in Sydney. It’s just probably the most magnificent city as you go for a run around. It’s just stunning.
Kirill: Fantastic. I can imagine. So many people envy you running near the city over at HAST.
Josh: Yup.
Kirill: That’s so cool. So Josh, we met for the first time at Deloitte back in 2012. Is that right?
Josh: 2012, that’s a long time ago. I guess it was 2012, man. I’m trying to rack my brain here. That sounds about right. I started at Deloitte in 2010, so you must have started at Deloitte in 2012, would that have been about right?
Kirill: Yeah, we had like just one month overlap, and I remember I did my very first project for you, and I was like super nervous. Do you remember what it was about? Like some data cleaning, or something like that. I was super nervous because it was my first project, and at Deloitte the managers rate your performance, and I was like, “Am I going to get a ‘meet expectation’? Or ‘exceeds expectation’?” And I think you gave me an EE or a CE or something like that. I was super excited about it.
Josh: Yeah, I vaguely remember that. I’m sure it was. Back at Deloitte we were doing a lot of audit and financial analytics, and there was a number of projects in there, but yeah, I distinctly remember a young and eager Russian analyst who was pretty damn good at his job!
Kirill: Yeah, it was a good time. And then, all of a sudden, not even a month passes and Josh heads off to—where did you head off to, Josh?
Josh: I headed off to LinkedIn. It was a really interesting story, actually. It’s funny you bring up Deloitte because that really was the starting point of so many things. But I remember distinctly, I was sitting there on a Tuesday afternoon in the Deloitte office up in Brisbane with that beautiful view we had over the storied bridge and the river. And I got an InMail, or what I now call an InMail but at the time it was an e-mail, from a recruiter at LinkedIn down in Sydney, and he mentioned a whole bunch of stuff and whatnot.
To be honest, at first glance I thought it was spam and completely ignored it for about 24 hours. When I came back to it on the Wednesday afternoon and had a look at it again and I shot him back an e-mail and said, “Look, this sounds really interesting. I’d be up for a conversation, you know, whatever, let me know.” I didn’t hear back from him for a couple of days. Looking to all the recruiters out there, get on those responses!
But when he finally did come back, we started a conversation, we started a process, we started a fairly lengthy journey of the recruiting process at LinkedIn. And about 10 or 12 interviews later, a trip to San Francisco, a trip to Sydney for interviews, and one of the hardest interview questions I’ve ever been asked, they offered me the job. I was blown away. To be honest, I wasn’t really ready to leave Deloitte. I felt like I had so much still to learn.
I remember walking into the partner’s office to let him know that I had accepted this new role and he’s like, “Oh, Josh, I think you’ve still got so much to go here, and so much to give. What’s it going to take to get you to stay?” I was like, “Look mate, to be honest, this tech has been something that’s really interested me for so long. There’s really not much you can say.” And he accepted that and that was that. Yeah, I’ve been at LinkedIn now for a little over 4 years, absolutely loving it. It’s just an amazing place to work. The calibre of the people here, the culture of the people, it just makes coming to work every day just incredibly fun.
Kirill: Fantastic! And the partner was Chris Noble, right?
Josh: That’s right. And Chris, as you know, is kind of a tough guy. He’s got a very big presence and I remember that conversation distinctly. It was not one of the easiest conversations I’ve ever had. Saying no to Chris is never something that you do lightly. So, that was hard.
Kirill: That’s true. Obviously, you’ve got us all hooked. What was the hardest interview question you were ever asked?
Josh: So, I remember sitting down with the Managing Director of APAC at the time, Asia Pacific at the time, in Sydney. And they’d set me up, the Head of Sales came in and interviewed me. That was for about ¾ of an hour or so. It was fairly congenial, selling LinkedIn kind of thing, and it was rewarding. And this Managing Director, a great guy, an American called Steve Barr came in and sat me down and grilled me, just absolutely grilled me for about an hour on my credentials, on my experience. It was a tough interview. And then towards the end of the interview he said to me, “Josh, look, I’m a sales guy so I don’t really know anything about what you’re going to be doing at LinkedIn. It’s going to be with data. So with data, tell me why should LinkedIn hire you?” And I kind of looked at him just blank. I was like, “What the hell is he trying to get at here?” I ended up bullshitting my way through it, so it was all good and I managed to get the job.
Kirill: Okay, gotcha. All right. Good to know. “What data evidence can you give us why we should hire you?” That’s a pretty cool question.
Josh: Yeah. (Laughs)
Kirill: All right. Cool. Tell us a bit more about what you do at LinkedIn now. What is your role and what are your tasks and goals on a day-to-day basis?
Josh: Sure. So, when I started in 2012, I was the first person in what we call the Insights Team. The Insights Team at LinkedIn is an analytics team that sits within the sales organization and we’re essentially tasked with creating the world’s best sales and marketing teams and data-driven insights. And the way that we do that is essentially building out tools and narratives and collateral for our salespeople to go out and essentially be the best at what they do and really empower them to be better salespeople. That’s on the one hand.
And then the other part of our role is being client-facing, so we’ll go out to clients and our largest corporate customers and help them understand their business really in a lens through LinkedIn data. So we’re able to apply that LinkedIn dataset to some really interesting business challenges, whether it might be something like entering a new market and understanding what the talent looks like there, or internally, answering questions around “What does my attrition look like?”
I’ve had clients come to me and say, “Look, we know that we’ve got something like 30% of people who join us leave within 12 months. It’s a massive cost to our hiring. I want to know why, I want to know how that compares to our competitors, I want to know in what pockets that’s actually happening. Is there anything in the data that you can tell me that may start to address that kind of issue?” So, that for me has just been incredibly interesting in the last 4 years.
As I said, I started in 2012 as the first person in that team here in Asia Pacific based in Sydney. And I was working with our customers all around the region, which was incredibly challenging. And since then, I have built a team of 13 people who are based here in Sydney and Singapore and Tokyo and Bangalore in India. It’s just incredibly fun. It’s a really gratifying thing to build a team around you and be able to deliver just so much more impact than what you’re able to do yourself. 
Kirill: Yeah, gotcha. So you’re managing a whole team that’s predominantly Asia Pacific?
Josh: Yes, that’s it.
Kirill: That’s cool. Okay, so you work with employers and recruiters, as I understand, to help them better assess the market conditions. I actually had a better look at your talent pool reports on LinkedIn, and anybody listening to this who is an entrepreneur or a business owner, they’re very useful. Can you tell us a bit more about them?
Like, back in 2013 I think you did the very first ones for Australia, and that kicked it off here. What are they all about and how do people use them?
Josh: Yeah, actually the talent pool reports are a really interesting story and it’s a really interesting illustration about how we think about innovation, both at LinkedIn at-large, but also within our Insights Team. So when I first started, one of the most common things that we would do is a deep analysis on a particular talent pool or for the non-HR talent acquisition people, essentially what does a particular segment of the talent market look like. So, what do the Java developers look like in Australia? Who’s employing them, where are they migrating from, who’s educating them, what kind of skills do they have, what kind of experience, all of these kinds of things. And then not just looking at the demographical type of stuff, but also looking at some of the drivers and the employee value propositions that drive them to potentially change jobs.
So, if I’m an up-and-coming Java developer, software house, what do I need to do to actually build a team? How do I compete with some of the big guys in Australia? The banks employ just about more technology people than anyone else. You know, Telstra, even up-and-coming guys like Atlassia. How do I compete with them at a scale where I can attract the best people? And that’s what the talent pool reports are all about, trying to help employers understand what the market looks like and how do I employ a really niche set of skills into my organization.
So when I started in 2012, these were very manual processes. It would probably take us 2 or 3 hours to define an audience, put this into practice, synthesize the data out of our systems, visualize all that kind of stuff, and put together a narrative.
Over the course of the next 2 years, we automated that process to a large degree where we now have a library of real-time talent pool reports that are an incredibly deep analysis customized to the company that we’re talking to, and there’s something like 2 or 3 million different variations that are available in that library. So we’ve gone from a manual process right through to a real-time productized version of what we’re looking at. And that report is something that organizations have come to rely on in terms of the type of analytics that’s available.
Kirill: Gotcha. And I was really surprised or, I guess intrigued, to see that you don’t only analyse supply, which would be the first thing that you think about when you think of LinkedIn, that there’s all these professionals putting up their profiles and you have all this data on how many people you have in a certain region, what their skills are, who’s endorsing them, what they’re interested in and so on. But in addition to supply, you also analyse demand. So you analyse how recruiters are behaving, what companies are putting out there, what offers, what employment conditions, and so on. Could you comment a bit on that? Like, how do you integrate the two, supply and demand, to come up with these insights?
Josh: Yeah, absolutely. Like any other market, the economics of the talent economy, if you like, are incredibly interesting and nuanced and dynamic. So, when you’re looking at a pool of talent that is fairly commoditized, so looking at something like accountants, there is a steady supply of accountants and the demand for accountants really hasn’t changed over a long period of time.
But when you’re looking at – especially in the technology sector, and it’s not just limited to the technology sector, but I’ll get back to that – when you’re looking at a niche set of skills that are incredibly high demand, and that’s in demand from a business point of view, you know, these are the kind of skills that are building businesses and that the businesses are succeeding because of. The demand on those individuals really does go through the roof compared to those more commoditized professionals.
And what that means is that the price companies need to pay goes through the roof and the scarcity increases as well. You know, down the track that means that more people enter the market, and economics takes over and price eventually comes down. Really, we’re able to start to help organizations understand what that demand looks like and really that’s a bit of a unique thing because of that LinkedIn dataset. Not only we have the job side of things, so we have that job advertisement, and that can be a good proxy for demand, but that doesn’t necessarily marry up the supply with the demand, which really gives you the price sensitivity analysis that you’re actually looking for.
So we’re able to, with that supply and demand data, help organizations understand and set expectations back with the business around how hard it is going to be to find that Java developer. And if you’re looking for a Java developer who’s got experience in the education sector, who speaks Mandarin and who was educated at the top universities, etc., it’s getting incredibly hard to find that person.
So what we’re able to do is use that data and that analysis to set expectations with organizations to look for criteria and a mix of criteria that is actually going to help them hire the best people with the right set of dynamics. And what I mean by that is the right cost of hire, so the right salary package, and the right time to hire so we can set expectations around, “Okay, it may take you 1-2 months to find that person, depending on how strong your employment brand is, and get that person into your organization.” So we’re able to help companies really understand what that dynamic talent market looks like, and it becomes just a whole lot more interesting when it has a look at some of those high demand skillsets.
And that’s not limited just to tech. You saw a really interesting phenomenon is the Australian economy over the last 5-7 years with the kind of boom and the bust of the mining sector. That’s starting to pick up again, but you saw truck drivers earning $200,000-$250,000 a year in fly-in-fly-out situations simply because of the lack of supply of talent who was willing to go in and do that kind of thing, even though I think traditionally the labour economy wouldn’t have necessarily seen those skills as high demand and then priced them accordingly.
That kind of dynamic is really impacting a lot of industries in a way that we haven’t seen before, especially when it comes to digital talent because digital talent is no longer just the realm of tech companies. Every single company wants to become a software company. They all want to build their own in-house capability, they all want to build their own in-house stack. And what that means is that that ever-shrinking pool of talent becomes ever-increasingly in demand.
Kirill: Yeah, I totally agree. I can attest to that. Actually, at SuperDataScience we are hiring developers right now because we want to bring that capability in-house. Yeah, we looked around and usually you hire people through connections or through your own network of people who are kind of already in the team. But we couldn’t find anybody so the next place we went to was LinkedIn, and we put up a job post and within a few days we got some really cool candidates who are now going through the entry process to see if they’ll fit. So thanks for that.
(Laughs)
Josh: My pleasure. And, you know, it’s really interesting when you can start to look at—Kirill, I don’t know where you were hiring those people, but the other thing we can start to help with is thinking about—for instance, a lot of tech companies are in the San Francisco Bay Area hiring talent, and software developers there can demand just an absolute premium for their services. And an even more relevant example is data scientist, right? Data scientists in the Bay Area earn an absolute shitload. Like, it’s crazy how much they earn, which is awesome, of course, and there’s more macroeconomic pressures around the cost of living and those kinds of things in the Bay Area.
But for a company that’s starting out, what we can start to help them understand is, if I was to shift my focus away from the Bay Area, there may be a shift in quality, but can I be looking at different markets to hire data scientists where it may actually be cheaper for me to hire them? Or can we help you make the decision around training up some data analysts who become data scientists internally, and what is the cost implication of that kind of a decision. So, that’s the kind of analysis that we’re really interested in the Insights Team, helping organizations.
Kirill: Yeah, totally. And that’s like the WordPress model, like a company that employs over a hundred people, completely decentralized even though they’re registered in the Bay Area. Some of them are in Ireland, some of them are in Latin America, some of them are in Asia and so on, like all over the place. And by opening up your mind to where you are looking for people to join your team, you are opening up your company to a global talent pool. And therefore the costs go down, and if you look well enough and in the right places, you still find quality or even higher quality and diversity of ideas and thoughts in the people you’re hiring.
Josh: Yeah, absolutely. And I think the other trend that’s really impacting that is the gig economy. You know, we’re seeing companies now that employ labour that doesn’t enter into a traditional relationship with them. They don’t have that employment relationship, yet they’re able to build an incredibly strong capability through the gig economy. Look, I think there’s risks with that as well. I think you’ve got risks around your intellectual property and all those kinds of things. But for the right situation, I think that the gig economy is incredibly interesting and I think it’ll go well beyond Uber or Airtasker to things that are much more interesting. LinkedIn is actually running a pilot right now in the U.S. with a service called ProFinder, which is kind of flying under the radar but doing incredibly well, and this is all about leveraging the LinkedIn network to really fuel the gig economy. And I’m really excited to see where that goes in the next couple of years as well.
Kirill: That’s very interesting. Could you expand on that a bit? I haven’t heard that term before, “gig economy.” What is that and what are the implications?
Josh: Well, the gig economy is all about—I think you’re a great example, Kirill. The work that you did on Udemy and whatnot, you didn’t necessarily have a relationship with Udemy before you put the courses on there. You recognized a hole in the market for high quality education and you were able to go ahead and create amazing content and do that without a formal employment relationship with Udemy. And I think we’re seeing more and more organizations popping up that are taking advantage of that kind of contracting.
Kirill: Gotcha. Okay, interesting. So when there isn’t a formal employment relationship but still there’s a mutual benefit for collaboration?
Josh: Exactly.
Kirill: All right, cool. That’s very interesting. And we talked a bit about the benefits and the different situations for employers, but what advice can you give to our listeners, for people who are aspiring data scientists or analysts who are maybe looking for jobs or maybe they already have a job but they’re open to new opportunities, they want to see what else is out there? What kind of tips or maybe things to look out for can you recommend for them?
Josh: Sure. Having built a team over the last few years—I mean, I’ve done a lot of hiring, and it is really interesting. I mean, I find hiring such a fascinating thing. You make such a huge decision on such a limited amount of information. I feel like I’ve learned a few things in that process. I still don’t think I’m the best hiring manager in the world; I’d like to think I’m okay. I think I’ve got a pretty decent track record. I’ve got an amazing team that sits around me and makes me look good.
Look, I think the number one thing when you’re looking for a new job, especially for people who are interested in getting into data science, it’s been fascinating seeing the number of people—again, going back to what I was talking about earlier, when a profession or a niche set of skills becomes in higher demand because of technology trends and the like, that we see all the new entrants into the market. It’s classic economics.
What that means is that you’re seeing new entrants coming into the market, into the data analytics and data science fields who have got incredible variety of experience and they’re coming from different backgrounds, and they’re bringing lots of experience. So, I think the number one thing for me, first of all for people who are trying to crack into data analytics, it’s all about proving yourselves. So, I am much more likely to take a risk on someone, to hire someone who has shown that they’ve gone out and have what I like to call the “growth mind-set,” and that’s the ability to independently show that their learning journey is consistent and constant and that they’re essentially learning the whole time, they’re looking for opportunities to learn, they’re always looking for feedback, they’re looking for ways that they can improve upon themselves, they’re not happy with the status quo, etc. That’s something that I really try to hire for. And that’s incredibly hard to identify quickly and to be right. That’s something that is demonstrative over a long period of time.
But I think there are things you can do to demonstrate that to hiring managers, especially if you are trying to crack into analytics, and that’s going and listening to the famous Kirill on Udemy, and taking the best courses you can, and really upskilling beforehand. And one thing that I would like to see more of in the analytics field is more portfolio based work, so the ability for you to—whether or not you’ve built a successful predictive model, or great visualization or whatever it is, look for public ways to actually build up that portfolio of work. I think not enough people in our field do that and that’s something I would love to see more of, whether or not that’s competing in public data science competitions on things like Kaggle, or publishing stuff on Tableau Public. I know Tableau run a whole bunch of different things trying to engage the community, and they do a fantastic job of that. And I’m sure there’s plenty of other opportunities out there to build a brand for yourself before you’re even starting to engage with employers. I think there is a huge amount of opportunity to prove that you have that growth mind-set. I think that’s probably the number one thing you can do, prove that you have the ability to learn.
Kirill: Yeah. I totally agree with that. I get that question a lot of the time from the students saying, “Hey. Look, I love your courses and so on, but I don’t have any experience. What can I do to demonstrate to a potential employer that I have the skills, or even that I’m willing to develop the skills, that I want to gain this?” A lot of people get put off by the fact that it seems so hard to break into the space of data science simply because in their mindset, employers require you to already have some real world hands-on experience before they’re even going to consider you.
Would you say that that is not the case, that even in your position you’d be open to hiring somebody who has this growth mind-set even though they don’t have any real world hands-on experience at a job before applying to your job?
Josh: Absolutely. I mean, first of all, for a lot of hiring managers and a lot of businesses out there—I think if you take a step back, I think a lot of businesses still hire purely on experience, even on education. You won’t even get through a recruiter unless you’ve got the right university on your resume, or your LinkedIn profile. I think that’s disappointing. I think that is becoming less and less of a requirement, which is amazing. And I think the educational model is changing rapidly. I think what we will start to see in our lifetime is university rates continue to decrease as more free form education and, especially in the technology fields, skill-based education become way more apparent, way more rampant. And employers will hire based on potential.
I mean, my hiring philosophy is always trying to hire someone better than myself and that’s in a number of different ways and that doesn’t necessarily need to mean overall. Looking for something, whether or not that’s enthusiasm or energy or that real kind of doggedness when it comes to learning, and the ability to pick something quickly, you need to be able to prove that. Especially if you’re coming from a low base of experience, you need to be able to prove to a hiring manager that that risk is worth it, and I think that is something that’s possible.
And I think something that in the analytics field gets overlooked a lot is building a personal brand around. Look for opportunities to publish content based on what you’ve done. I think LinkedIn platform medium and other places are just fantastic. This is terrible advice, because I don’t actually publish enough content. This is something that I’ve been wanting to do, and I need to find some more time to actually do it. But I think the ability to build a brand, especially if you’re looking for new opportunities, is incredibly important.
And then there’s some great features on linkedin.com. You can now show that you’re open to opportunities to recruiters, which is not shown on your profile, and you can define exactly what kind of opportunities you’re open for and which regions and all that kind of thing which is really, really interesting and getting a lot of traction. But I think it all comes back to lowering that risk for a hiring manager. And if she can see that you have a big presence in the analytics community and you’ve contributed in a meaningful way and you’ve gone about learning and produce some amazing content, I think that is a huge benefit to you. 
Kirill: Yeah, totally. I’m just going to recap some of those media so that people listening to this can either write this down or just have it in the back of their mind. So, like, publish content, get out there—for instance, go do a Kaggle competition, publish visualization on Tableau public and then write up an article about what you think about data science or what you’re learning and publish it on LinkedIn. There’s lots of other blogs that are open to publishing. If you’re interested in R programming, submit it to r-bloggers.com. Any other ones you can think of, Josh, off the top of your head?
Josh: There’s so many out there. Something that the more creative professions do a lot better than us in the analytics field is creating their own websites and showcasing their talents and what they’ve created. And I think that’s something that’s so cheap and easy to do these days and that really is underused.
Kirill: Exactly. Like, my most recent job at a suprannuation company, that’s how I got a recruiter to contact me. I was just posting content onto LinkedIn. I actually got this tool called Hootsuite and I scheduled content. Some of it was my content, some of it was my comments on somebody else’s content. I scheduled it to go out like twice a week, and for a few months I was posting, I got more and more interest and all of a sudden these recruiters started contacting me.
So, you can create a passive stream of job opportunities coming your way like that as well. But I wanted to ask you as well, what would you say are the most important skills for data scientists right now? Because we’ve heard of lots of things – R, Python, Tableau, Power BI, SQL – there’s so many different tools, a whole plethora of them, and for somebody getting into this space, it’s really hard to figure out what to focus on. Do you have any tips on that?
Josh: Yeah. I mean, to be honest, I think the most valuable skills in data science and in analytics really aren’t the technical stuff. It’s communication, it’s business acumen, it’s the ability to translate what you’re doing back to impact to the business. I think there is almost too much focus on the technical skillset in data science and not enough on what that’s actually meaning back to the business.
I think you’re seeing that in intangible ways where you’re seeing recommendation engines that just recommend what you’ve purchased prior and not enough intelligence going into the “so what?” if you like, to predictive modelling and a lot in the analytics field. I really think that there needs to be more of an emphasis on the business impact and the customer experience and end user centricity rather than the actual technical skills. I think too often data scientists and analytics professionals get way too excited about technical things rather than the impact those technical things actually have on the bottom line. And I think the other one is some of those softer skills around communication, just being a high quality professional, time management, etc.
Kirill: Yeah, thanks a lot for that. I really appreciate you saying that, because a lot of the time I end up in the same situation that I need to explain to people that these soft skills are so important. Like, crunching numbers has always been a thing. It’s always been around, there are so many people that can crunch numbers. But the true data scientists, the ones that can stand out, and the ones that are famous, and the ones that earn the big salaries, and the ones that are in demand are those that can not just crunch the numbers, but who can actually translate the insights into human language and convey them and help executives take action on those insights. So thank you very much. I appreciate you making that comment.
Josh: Yeah, I think the really interesting trend for me is in data democratization. That’s the proliferation of analytical skills and data skills into the entire organization. And I think what we’re seeing is really that commodification of analytical skills. They are going to be less and less niche in the future. They’re going to be more and more mainstream and you’re going to have people in – well, you’re already seeing it, right? People in marketing with analytical skills, people in sales, people in all sorts of functions who do their job better because they’re data driven and they have that data driven mindset and they have that analytical skillset.
I think hard data sciences will always be a specialized centralized function, but I think some of those other things, and what that really means is data science practitioners are going to need to become way more closely tied to the business. They’re going to need to make sure that what they’re doing is producing real and tangible and measurable impact. And to be honest, that’s one of the challenges that we’ve always faced, is measuring that impact. I think that’s something that we need to start to challenge as a function.
Kirill: Yeah, I agree with that. And I like what you mentioned that more and more people that are not specifically in the space of data science will slowly start to need to develop these skills around data. Can you comment a bit more on that? How does somebody that has never programmed before or has never even considered data as part of their role, somebody in accounting or HR, or maybe somebody who owns a little bakery down the road, somebody who is passionate about what they do, but they haven’t ever considered data to be a thing for them, how do they start thinking about data? How do they start using it in their day-to-day decisions and their business or their careers or whatever they’re doing?
Josh: I think there’s a few ways that this is going to happen. For an elite set of companies, you will see an opening up of data skills to the entire organization. I know at Facebook, every single employee has the ability to get access to the SQL databases to start to hack on [inaudible 39:45] and improve what they’re doing through a data-driven way. We have a similar – we have a more restrictive access here at LinkedIn. So I think that’s one way of doing it, I think just giving that technical ability and saying, “We’re not centralizing data skills anymore.”
I think the other way that that’s going to happen is through software platforms and Analytics as a Service platforms that are Cloud based, that makes your mom and pop store incredibly data-driven simply because they invest in that platform that they can take their transactional data, they can take their customer data, and have a customer segmentation model built automatically without the need for a consultant to come in. They can understand their customers in a more granular way. And all of those kind of data-driven things that we had considered were incredibly advanced and incredibly necessary of a highly paid consultant a number of years ago, I don’t think that’s going to be the case anymore. I think in the future you will see software platforms come in that are incredibly easy to use, where visualization is built in.
I think we’ve already seen that journey in a couple of places in the organization and I think that’s really being kicked off in marketing, where Google Analytics is incredibly easy to use. It is allowing marketing organizations to be incredibly savvy about who they’re targeting. You look at even Facebook targeting and Facebook advertising. That’s allowing marketing people to just be—the amount of targeting opportunities in there is just ridiculous, and that’s because they’ve done a lot of that work in the first instance to really drive that.
Similarly, some of the things you can do here at LinkedIn as a customer are just fascinating to be a data-driven marketer. And I think you’re starting to see that level of data democratization come to other functions in the organization. I think analytics in the HR space is something that’s maturing right now in real time and will continue to. I think at the moment it’s still the vanguard of all organizations, but you’ll start to see that trickle down to every single organization.
I think something like workforce planning for a long time has been a dirty word because no one does it right and it’s kind of been a long-running joke that you kind of—it’s a “finger in the air” type of thing, but I think that will become more and more prevalent and more and more important for the organizations to do. And it will continue to other parts of the organization.
Kirill: Yeah, it’s very interesting how we’re capturing more and more data in the world all the time. Like, it’s always been there, but now we’re just able to capture more of it and therefore apply it in different areas of an organization all the time. And what would you say is the biggest challenge for organizations dealing with data right now? What is the biggest hurdle or roadblock? Because things are becoming easier, technologies are growing exponentially, but we’re still seeing that it takes time for organizations to adopt these technologies and it takes time to start working with data more efficiently. What do you think is preventing them from doing it faster?
Josh: Yeah, that’s interesting, Kirill. To be honest, I’m incredibly fortunate to be in an organization that data exists in our culture. The very reason LinkedIn is successful as a company and exists as a company is because of the data our members entrust with us. We’re able to monetize that because of that trust. LinkedIn’s culture starts with members first. And that means that the members’ experience comes first no matter what. But data is at the centre of our organization. And data is at the centre of our culture. I know that I have worked in organizations where that is not the case and data is at the far reaches of their culture.
We were talking before about an organization who wouldn’t give access to the data to the marketing department. And I think that’s not an outlier. That’s pretty standard. I still think that a lot of IT departments see data as their domain and you will wrestle that out of their cold dead hands. I think that will change. There will be I think more and more CEOs and CFOs that are seeing their data as a valuable asset, and they want to extract as much bottom line value out of that asset as possible. So, I think there’s number one, a lack of data-driven culture. Number two, it’s access to data across the organization and there’s probably also an education factor around a lot of organizations around being data-driven. I would suggest those two are probably the biggest reasons, or the biggest roadblocks.
Kirill: Yeah, I agree with you. And in that sense, I’ll share a statistic with everybody listening to this because we ran a survey just recently of all of our listeners, of a lot of students on SuperDataScience, and I’m actually just bringing it up now. Approximately 40%, I think it’s 43% of people listening to this podcast—yeah, 43.3% of people listening to this podcast right now are employed but actively looking for new opportunities. These are people who have a job, who are excited about what they do, but they’re open to new things and more so, they’re actually actively looking for new places to work at.
Maybe the reason is because there is that little access to data. Maybe it’s some other reason. But what I wanted to say is that when you’re actually going into an interview, it’s not just important to impress the recruiter or the hiring manager, or your future manager. It’s also important for them to share with you the details of your future role and the structure of the organization, for them to impress you, so that you don’t end up in a situation where you leave your current job to join a new job where you don’t have access to the data and you’re not having fun and you’re not doing the work that you were thinking you were going to be doing.
So Josh, what would you say is the most important question or questions for somebody applying for a job to ask in the interview to make sure that they’re setting themselves up for a successful and interesting career?
Josh: Yeah Kirill, it’s a really good question. I am blown away by the number of people who I get to an end of an interview with and I say, “What would you like to know about LinkedIn? What questions do you have for me?” and they go “No, I’m good.” “Seriously?! Like, as much as I am interviewing you, you are about to decide upon, and embark upon, a massive part of the journey of your career.” I don’t know, maybe it’s because I’m now in an organization that I just incredibly enjoy, especially around the data parts of it, I don’t think I could go work somewhere right now that has some of those challenges and enjoy it as much.
Yeah, absolutely, you’ve got to decide what’s important for you. You’ve got to decide what you want in your career. You’ve got to decide what you want in a job. And if data is important to you, if culture is important to you, you’ve really got to investigate that and investigate that hard and make sure that what it looks like on the inside is what you actually want because otherwise that statistic of 30% of people leaving within the first 12 months, that’s largely because the culture and what you’re doing day-to-day when you enter the organization is very different from the picture that you had or the assumptions that you made about that organization pre-hire. That, for me, is a failure of their employment brand. You’ve really got to investigate that. You’ve really got to try to find out as much as you can, you’ve got to find content.
There’s lots of different ways for you to be able to find content about what it’s like to work in an organization. But one of the most crucial is that interview process. I would be spending at least—in my position it’s probably a little bit different than most—40-50% of that interview questioning them because it’s as much about you as it is about them.
Kirill: Yeah, I totally agree with you. And thanks for sharing that. I wanted to ask you a couple of quick rapid-fire questions about your experience with data science. Are you ready for this?
Josh: Go for it.
Kirill: All right. What’s the biggest challenge that you’ve ever had as a data scientist?
Josh: The biggest challenge? I’d probably have to go back to Deloitte days.
Kirill: Deloitte days. Good times!
Josh: I mean, I think the biggest challenge I ever have is when a client or a manager says, “Show me something interesting.” You know, you start digging into the data and you’re like, “Well, that looks kind of cool.” And then two weeks go past and you’re like, “Oh, I’m stuck.” I think every time you start down a journey without an end goal in mind, or without a plan, that for me is just absolutely the most challenging and I hate that. I hate getting into a position where I don’t know where the end goal is.
Kirill: Yeah. I think the onus should be on the person asking the question that they ask the right question.
Josh: I think the onus is always on you, right? You can’t necessarily control what question they ask. If they ask a stupid question like, “Show me something interesting,” I think the onus is on you to really push back on them and say, “Why? What do you actually want to see? What’s important to you? What challenges are you facing?” If you can’t be a good enough consultant and a professional to plan, it’s really your own fault if you get into that position.
Kirill: Do you have a specific example of when somebody couldn’t identify the business problem and you had to actually sit down and identify the problem first before you could do the analytics?
Josh: Well, I’ll give you an example of where it didn’t happen, where I didn’t do enough.
Kirill: Okay, sounds good.
Josh: Again, going back to Deloitte, this is where I was working on a case for an airline. We got their frequent flyer data—that kind of narrows it down in Australia—and they really said, “Show me something interesting.” I think our team had soldered in pretty much like that. And it did end up the engagement over budget, the engagement over time, and a large part of that was because there was a sense of exploring to kind of find something to find in the data and clean it up. We ended up finding some really interesting stuff, but without a plan in mind, it’s always going to be more expensive and more expensive on time and more stressful.
Kirill: What would you have done differently?
Josh: I mean really, these days, I wouldn’t have accepted the engagement unless we had a very, very clear plan of exactly what we were going to deliver and what the objectives were. These days my team kind of gets sick of me when we kind of get lost in the conversation and they go, “Hang on. Let’s take a step back and try to understand what we’re actually trying to achieve here and what’s the actual objective and what we’re doing.” I get very frustrated when these conversations kind of get lost in the quagmire, you know, lost without objective.
Kirill: Gotcha. Okay, continuing the rapid-fire questions, what is a recent win that you can share with us that you’ve had in your role, something that you’re proud of?
Josh: The biggest thing that I’m proud of is the end of 2015. I had a team of four people across Asia-Pacific and then we went on a massive expansion into 2016 and we went up to a team of thirteen. Tripling the team last year, and tripling the team in a way that we did it where the team came together and decided what that was going to look like and decided what our strategy, vision and our culture or who we wanted to be and who we wanted to be known as, and how we wanted to impact the business. I was incredibly proud of that.
When you move more into leadership positions, it’s less about the technical stuff. If you prefer a technical answer, I recently used the Twitter API to pull down some data off Twitter, which was quite interesting and quite fun, getting my hands back into the dirt with Python.
Kirill: Yeah. Okay, cool. Thank you. That’s awesome, double answer to the question. And what is your one most favourite thing about being a data scientist? 
Josh: My favourite thing about the data—like, I get an incredible sense of achievement, like one of those incredible challenges that someone else has and helping them solve that problem, and helping them see that challenge in a new light using data. My biggest sense of accomplishment is when I can change a decision or change course on something that everybody saw was failing and using data to change that through data.
Kirill: Yeah, that’s awesome. I definitely can agree with that. That’s very inspiring for yourself when you can do those types of things. And next one is actually a question from one of our listeners, from Natig Aliyev, and he asked will there be a need for data scientists in the near future while very powerful machine learning tools are being developed? Basically, will data scientists become obsolete?
Josh: I don’t think data scientists will become obsolete in the near term. Just the same as software engineers won’t become obsolete in the near term. I think we’re still going to need very, very intelligent individuals who can tell machines what to do, whether or not that’s in the machine learning space or software engineering space.
In the longer term it gets a little bit murkier, and I guess that really depends on your time horizon on that question. I still think the current set of software, especially in the machine learning space, still requires so much human intelligence, human configuration and input, especially on the data cleansing side. I’ve been involved in many data science or analytics projects where data cleansing and whatnot is still a massive part of what we do. And I think until that human element is removed, I don’t know, I just don’t see that becoming an issue in the near term.
Kirill: Beautiful. Thank you for reassuring, so people in data science keep studying. And from where you sit, from all the things you see about data, you know, you have a team, you work at LinkedIn, you have so much access to data, you actually have this bird’s eye view of what’s going on in the world. Where do you think the field of data science is going, and what should our listeners prepare for coming in the future?
Josh: I think I mentioned it in an earlier answer, Kirill. Every organization in the world who sees their data as an important part of the value that they create is going to continue to see that data as more valuable to their bottom line. And I think the real challenge for data science in the near term is connecting with the business and really focusing on measuring their impact, making sure that the models that they’re putting in place, the algorithms that they’re driving, are driving real revenue, bottom line profit. And that’s it, really. And I think that’s the biggest trend that’s happening right now. And I think the data scientists that will be most valued are the ones that are able to build a brand for themselves because they have delivered success to businesses. I think in the same way that private equity and venture capitalists provide value by stripping companies and turning over companies and that kind of thing, I think data scientists will build personal brands based on the work that they can do with organizations to reinforce the top line and bottom line.
Kirill: Yeah, gotcha. Fantastic! Well, thank you very much for your answers and thank you for coming on the show and sharing all the insights.
Josh: It was a pleasure.
Kirill: How can our listeners contact you or follow you or find you so they can learn more about how your career progresses?
Josh: Well, funnily enough, I’m on LinkedIn, so find me on there. And also Twitter. I still very much enjoy conversations that are happening on Twitter, especially in the analytics community.
Kirill: Thank you. We’ll add those links to the show notes. And one final question for you: What is your one favourite book that you can recommend to our listeners to help them become better data scientists?
Josh: Kirill, you’ll enjoy this. So, back when that one month overlap happened in our careers at Deloitte, you recommended me an author who I looked into and derided at the time for being a bit of a douchebag – Gary Vaynerchuk. I have rediscovered Gary Vee only in the last couple of months, so I apologize for my derision years ago. And to be honest, I wish I hadn’t. Because maybe I’m in a different headspace, maybe he’s communicating in a different way but we totally connected on a very impressive level. I’ve been more motivated and inspired by the work of Gary Vee. And I am currently reading his book “#AskGaryVee,” which is absolutely awesome.
Kirill: Oh, that’s the most recent one, right? Like, book number three or four?
Josh: Yeah.
Kirill: That’s so cool. I haven’t read that one yet, but I can totally attest to that. Gary Vee has got quite a few books. I’ve read his very first book—the name escapes me right now, but—
Josh: “Crush It!”
Kirill: Yeah, “Crush It!” There’s actually an audio book on Audible, if anybody listens to that. Very, very good book. He’s very arrogant, you’re right. He does come off as a douchebag sometimes, but on the other hand, his methods work.
Josh: I think his energy is just infectious and I think that for whatever—I’m not an entrepreneur in the traditional sense of the word, and I may be one day, but I think that for anyone in any part of their career, a lot of the things that Gary talks about, I think he’s incredibly valuable in learning some of those lessons. Even in the last few months, he’s been incredibly valuable to me.
Kirill: Have you seen his YouTube show, “The #AskGaryVee Show?”
Josh: I have. Well, the book is kind of a—
Kirill: Yeah, summary.
Josh: Yeah, summary or the best bits of that. So, I’ve been watching that and the “Daily Gary Vee” or whatever it is on his Facebook channel. He’s like – any social media, take your pick, he’s there. So I highly recommend him.
Kirill: Thank you very much. Well, once again, thank you very much for coming on the show. It’s been a great pleasure to catch up and reminisce about the old days and also for you to share all these valuable insights. Thank you so much.
Josh: Sure.
(background music plays)
Kirill: So there you have it. I hope you enjoyed this conversation. Lots of valuable insights and a bit different from our usual conversations about tools and techniques and data science. This time we were focused on careers. We were talking about how you can enhance your career, how you can find better opportunities, and how you can build the career that you deserve, the career that you want for yourself.
And personally, my favourite takeaway from today’s show was when Josh mentioned how you can get out there, how you can become interesting to recruiters, how you can create an agenda for yourself even if you don’t have the data science experience. And plus, even if you do have the data science experience, it’s an additional bonus.
So remember to get out there, to look at things like Kaggle, at the Tableau Public website, building your own blogs, submitting articles to LinkedIn — just publishing articles on LinkedIn, you don’t even have to submit them. Just go ahead and publish one. Or submitting articles to our bloggers and other blogs out there on data science, and tweeting about data science, building this profile, becoming somewhat of an influencer. I can imagine that it must be hard to become an influencer at the level of people like Peter Diamandis or Elon Musk, very, very top level, but you can become an influencer in your own circle of people that are interested in things that you’re interested in, and then put that on your resume, put that in your LinkedIn profile, put that in places where recruiters will see that. And it’s a big testament to the fact that you are driven.
And as Josh said, it’s not just about hiring for the skills. Recruiters are looking for people who are interested to learn, people who are interested to grow, people who want to enhance their careers, who are hungry for knowledge, who have this insatiable thirst for new things to learn and new things to master and contribute value to the companies they work for. And we all know that you are one of those people because you are listening to this podcast.
You are likely taking courses on Udemy or on SuperDataScience, you are completing the case studies, you are doing everything you can to enhance yourself, so you are already by de facto that type of person. You just have to show it. You just have to do a little bit of work, do a little bit of hassle to get your passion out there, to illustrate your passion. And then the recruiters are going to start coming to you, the offers are going to start flooding your inbox and that’s what you want. You don’t want to be chasing them; you want them to be chasing you. You want to put yourself in that position. So go out there, create content, share information, and create a fame agenda for yourself so that people are interested in who you are and want to know more about you.
And on that note, I am super excited that you were able to join us for this podcast. If you enjoy these episodes, then do us a favour – go on iTunes and rate the show. If you could leave a review, that would be fantastic. This is quite a new podcast and any help and assistance we can get in order to spread the word and tell the people out there that this is a pretty cool show to listen to would be very much appreciated. Thank you so much for being here and I can’t wait to see you next time. Until then, happy analyzing.
Show All

Share on

Related Podcasts