As someone who has been working in Switzerland and Germany for many years, what changes have you seen to the employment market, particularly within Data and Analytics?

When I started, Data Science was very niche in companies. It was found in either companies that had connections with North America, where Data Science was already well established, or for very special projects, where someone would need a very specific and specialised model.

That’s very different from where we are today globally, and so I think Data Science is seen in Germany and Switzerland as more of a general requirement, and even further it now is part of the strategy of the company. You’re not just an isolated individual doing work - building a specialised model at a bank or building a requested model for recommendation in e-commerce. 

I think Switzerland and Germany have almost caught up to the rest of the world, even at smaller companies like start-ups. Everyone now knows they need strong engineering and data from the beginning - it's not something you think about later, but rather you need to build your fundamentals in terms of data - but that's also because most start-ups and business models are now based around data. 

Since your degree was in Physics, what would your second career choice have been, and why?

In the UK and North America, Physicists often end up doing something else - not everyone ends up being a professor or a researcher in a lab. It used to be that Physicists went into Finance - hedge funds and so on, whereas now, Data Science is a big draw.

I think one of the big advantages of being a Physicist is that we are trained quite well in problem-solving, abstraction and simplifying. We value the concept of insights from a scientific framework, rather than blind modelling, which I perceive as being the difference between scientists and someone who comes purely from a Computing background. 

In terms of alternative careers, I took some courses that overlapped with Electrical Engineering, and have some friends and colleagues that did that, and I think that Electrical Engineering is a good balance between what you get in Physics and what you get in Computing. Of course, this was in a time when these things were more distinct. These days, almost everyone's going to take some Computing courses, even if they're doing pure Physics. But if you think of Electrical Engineering, it combines aspects of Physics and Computing, which give you an equally broad education like Physics, but with a more Computing connection. 

In terms of wider career choices, I think that most of us had a plan of becoming research professors and when that doesn't work out, then it's really how do you adapt. I think the more interesting question is how do people who intend on becoming researchers then adapt to the real world and that's actually something that's well-structured now for Data Science. There are Data Science programs in the UK and North America, and now there’s also one in Zurich that takes people with recent PhDs and take them through an intensive course to become Data Scientists. 

So that's now a fairly straightforward course of action for people with PhDs in Science or Engineering to immediately become Data Scientists.

When I was younger, the world was very different - it’s changed so much now that you can really learn anything thanks to the internet and the range of online courses. There’s just a great exchange of information - blogs and videos - I think it's much wider. I think your career and education are much more loosely linked than they were in my time. 

What are your personal motivators?

For me, it’s about doing something interesting. As we discussed before, I wanted to do interesting Physics and research. Obviously, there’s interesting research in industry, but my career didn’t follow that path. I became more of a practical, applied person and a leader, but I always try to find something that’s technically interesting, whether from a Computing point of view, or from a Data Science point of view, or possibly from a Physics point of view. 

You might ask how I could have Physics in the work that I do, but I used to do research on complex systems, and come up with systems that involved networks and social interaction, so you know how something becomes viral, or how you can use a network point of view to solve problems. For me, it’s about something that’s an interesting problem in itself. It could be a business problem, could be a Data Science problem and there needs to be a technical aspect of it as well. I also need the social interaction as part of a team or group - and it should also be a fairly social company.

What would you say is the most rewarding part of your role? 

Now that I’m in a wholly leadership function, I would say finding business value in Data Science - that’s our goal, that's our job to do that.

But as I hire younger people, and the age between them and me increases, I really derive satisfaction from seeing young people grow in my team and when I receive the thanks that really means a lot to me in that sense. 

As we said, with the internet and the amount of information out there, people can almost do anything right now with their background, as long as they're educated in some regard. It's wonderful to see how people tackle their careers, relating back to the question before. I think career opportunities are so much broader than when I came out of university. There’s a vastly greater choice, in terms of topics. So these days I get a great deal of satisfaction from seeing people be successful in my team.

Looking back through your own career, what would you say was your personal highlight?

If I were to choose one highlight, then I would say that it was changing the view on marketing attribution at eBay, because it was not just at eBay, but it was also more globally. 

Prior to that, people had last click attribution and multi click attribution with static models, but we showed that all of these were, in a sense, completely wrong. In effect, no different than doing something random - we showed that you need to build dynamic models and that you needed some concept of propensity to convert at the basis of that. The fundamental notion of this was that channels have no meaning; looking at the customer is most important. Putting those together was how we solved it at eBay and then I spoke about it globally.

I think that was the most rewarding and truly a highlight. Changing something globally and having an impact on the whole industry. Related to that, I find it interesting that you can make that change, you can talk about it and you still find people doing it the wrong way, even today. 

I think it’s ironic that going back to the beginning of my career, the opposite of what I was doing as a researcher was where I have ended up - in marketing! However, marketing analytics does actually provide some of the most difficult problems in Data Science. People typically think of deep learning, computer vision and AI [Artificial Intelligence], but marketing is so ambiguous that it’s actually quite difficult to find a solution that you can implement and then derive a billion dollar budget based on it - that's high risk! 

So, what was the ‘lightbulb’ moment that led you to this discovery?

It started off with an experiment - we always have biases and we do have a tendency to go with what’s already been done, so you can only innovate and disrupt if you break out of your narrow human view with some experimentation or something. That's how science progresses: experimentation is most important for science. There are only rare cases, like Einstein, where you can sit there and think and solve it. Even in industry, there are certain rules and certain behaviours that happen that you need to discover. Sometimes you're so stuck in the status quo that you really need to experiment. 

So it started with an experiment and then it started with shifting the mentality, so in this case in the simple solution to attribution. Attribution is about marketing channels and so the solution prior to my team’s work was to divide things by marketing channels and that's what gave the wrong answer. It’s actually different customer segments that give you the insight and the solution to solving it in the simple aggregated view. 

We always say that ‘customer, customer, customer’ is important, but time and time again in Analytics, it proved to be also the case - not just business strategy, or growing your business or customer relations. It really is true that understanding customers drives solutions.

The second part when going to multi-touch attribution, people thought again about channels, so then they had static models that compared where the channel was in the past two conversions and that also proved to be quite wrong, because again channels don't have any meaning - again it’s the customer. 

The same customer coming through the same channel may have different intents in different journeys, so you can't just have a simple rule, which was what people thought: ‘OK, I'll move from last click to multiple clicks with some timing formulation or something important, such as first vs. last’, but unfortunately this was meaningless as well, because the same customer or different customers can have different intents when they come to your website, so really you have to look at the customer behaviour to drive marketing attribution -  so again, it comes back to customer.

In terms of people, what are the biggest recruitment challenges you’ve faced?

Now that Data Science is ‘hot’, which it’s been for the last five years or so, there’s no shortage of people wanting to be Data Scientists, or that are Data Scientists, but it reminds me of when I was looking for my first job, but it was in Software Development and I was a Physicist. In the first round, companies didn’t want to hire me, because they were looking for someone specifically with a degree in Computing or Software Engineering, but then six months later, I got a call back from many companies saying I think we made a mistake - we need someone who actually knows how to solve problems and can think beyond their coursework. 

I think you find that now in Data Science, because it's such a commodity, that finding talented Data Scientists becomes more difficult. People having the ability to understand business problems, solve problems and communicate is very important, as well as simplifying things from ambiguous situations. 

So it's not about the technical skills, it's actually can they do the job in a challenging business environment.  

In addition, because Data Science became very deep learning, hardcore machine learning-focused in terms of vision, natural language processing and so on, in terms of the customer or Marketing Analytics side, that still remains very important, which is what Data Science used to be, finding those people who are used to the ambiguity or used to building models that there isn't an easy ‘yes it works or not’ is the most challenging thing, especially in e-commerce companies. 

This is because the landscape has shifted more towards more practical machine learning and AI. 

The other one is very specific. In Germany, the talent is not as good as it is in North America or Asia, so if you want the best talent, especially in a lower compensation environment, then finding those good people becomes really difficult. You have to motivate them with other things - you also need to find them first! That’s a continuing challenge here.

A final question on a topic which we know is very important to you and for us here at EMEA Recruitment - Diversity in the workplace. Tell us about some of your observations and experiences. 

I think about Diversity in two ways - one is Inclusion, and the other is Diversity of talent and mentality. As Data Scientists, we come from different backgrounds: Physicists, Engineers, perhaps Economists, Mathematicians or the more traditional Computer Scientists. I try to hire people with a diverse range of backgrounds, because as I said, we’re solving ambiguous, complex problems and you want different viewpoints. 

Most very complex problems can’t be solved by a single genius - it’s very rare for an Einstein to solve a problem. Most problems require diverse inputs and a team to solve them. All of our educational backgrounds have a history which gives us different ways of viewing the same technical issue. For this reason, I’m very strong on hiring people with different technical backgrounds and different experiences. If you expand this point of the benefits of different viewpoints, then you also need different viewpoints in terms of gender, nationality and so forth, because we all have different backgrounds and viewpoints.

In business, if you have people in a company who don’t represent their customer, then how can you understand their issues? There are numerous examples of companies filled with one class of people, but they sell products to a different class of people and they have no clue about those people or the problems their customers face! So there's a practical need for Diversity including gender, orientation, nationality and so on, so that you know everything about your customers. 

Of course, there’s also the social issue as well. In many of the countries I’ve worked in, there has been a history of discrimination: gender, race and so on. So how do you proactively address that? If you consider it a societal issue, then you need to be proactively changing. You have to do that not by disadvantaging individuals, but by looking wider in your search and considering all types of applications.

Interestingly, something that I wasn't aware of until recently was the way I write job descriptions. A recruitment partner said that my job descriptions were very masculine and might dissuade a female candidate. I have very succinct, strong technical language, which was also a factor in my being successful, in the way that I can summarise topics easily, but if you write a job description that way as well, it can deter certain people. 

That was something new for me - I have been very aware of gender diversity - at eBay, I hired an almost entirely female team and then started another team. So there's always something new - we may think we're completely open, but there are always biases we are unaware of. I think the most important thing about Diversity is talking with other people and taking advice, so that you can learn your own biases. In this instance, this was entirely innocent, but could impact perhaps 40% of potential applicants. 

So I think there is intent around Diversity, but there's also the practical element, which is where I think we all can learn. We all come from our very specific trajectory in life and there's no way we can truly understand other people's trajectories. 

Thank you, those are great points. So, you recently started at Beat - what motivated you to join them?

Urban mobility is an industry that enables me to be engaged in many of my passions: cities, complexity science, transport, design and people. Cities are self-explanatory. Urban mobility is an excellent example of a complex system, with large numbers of different components acting independently but in unison. It’s a great challenge to understand this system and to optimise it towards both societal and business value. Transport is not attractive just as a boy who loves trains and cars, but as a society-enabling service that needs to evolve to become more sustainable. In terms of design, I don’t mean the app UX [user experience], but the whole framework of how ride-hailing as an experience will transform cities, and how technologies such as autonomous vehicles will evolve with it. And of course, people: from the interview stages to my current day-to-day, I love the Diversity and enabling culture of Beat. Plus, I report into a strong female leader, which will be a good experience for me.

The real reason, though, is to enjoy the food and people in places such as Lima and Mexico City. Post-COVID, of course.

Thank you to Suresh for speaking to us.

Views and opinions contained within our Executive Interviews are those of the interviewee and not views shared by EMEA Recruitment.