The Drum Awards for Marketing - Extended Deadline

-d -h -min -sec

Neuroscience Science Agencies

‘Advertising is a big data problem’: what Known’s Rohan Ramesh learned in neuroscience

Author

By Sam Anderson, Network Editor

May 26, 2022 | 8 min read

Rohan Ramesh is vice-president of data science at integrated marketing agency Known. Known prides itself on merging data and creativity. Ramesh himself, with a neuroscience PhD from Harvard, manifests that mix by sitting across data science, social and search teams. We sat down with him to talk transitioning out of academia and advertising’s ‘big data problem.’

A floating brain

Harvard neuroscientist-turned agency data scientist at Known Rohan Ramesh on what he learned in academia’s most hallowed halls / Milad Fakurian via Unsplash

Hi, Rohan! Your current role is an interesting interaction between data and creativity...

Yes! I lead the search and social teams at Known. My responsibilities include making sure we deliver positive ROI for our clients, and applying data science to the social sphere and search, leveling up how we approach those platforms.

The hypothesis we have at Known is that advertising is a big data problem. Millions of impressions: how do you figure out who are the right people to target? How do you figure out how to measure success?

That’s an interesting team structure

Right. Our data scientists fit directly with social and search analysts. They’re not a set-aside team.

It’s a non-traditional data scientist role. I learned early on that having technical ability without knowledge of the thing you’re dealing with usually translates to useless results and insights. There’s often a disconnect between a data scientist’s recommendation and what’s practical for someone running a campaign. So all of our data scientists are required to have some hands-on experience with social and search campaigns. They can provide more real value because they understand the goals and what we’re trying to do.

Is it hard to find people who are both?

Mostly we find people with experience in the technical side of things, and an interest in wearing many different hats. It’s something I’m very open about in the hiring process: we want the people who want to learn things that are outside of a traditional data science responsibility.

You yourself didn’t start out in data, but neuroscience, right?

My interest in neuroscience started in high school. I worked for one summer with a professor at Johns Hopkins, where we were studying the walking patterns of people after a stroke. I just watched and picked up as much as I could, and that experience sparked a real interest in understanding how our brains work.

I went to Brown University, studying neuroscience; I walked away saying: “I want to be an academic, I want to be a professor, I want to have my own lab.”

I started working at a lab doing my PhD at Harvard. The thing that you learn first in grad school is how to wear many different hats and solve different types of problems. I think of a PhD as an advanced degree in problem-solving. For the first couple of years, I was equal part microscope technician, data analyst, computer and server fixer – whatever the lab needed me to do.

In a way that a fool like me can understand, what were you studying?

We studied how our visual system processes information differently depending on your motivational state. If you’re hungry and walking down the street, you’re way more likely to pay attention to the McDonald’s sign than if you’re not hungry; you’re way more likely to act on it by walking in and getting food.

We studied the part of the brain that processes, say, the golden arches of the McDonald’s sign. That area integrates information from all over the brain – the hypothalamus, which signals hunger; the amygdala, which tells you what’s important. I looked at how all that information comes together. How does the brain integrate all that information and represent it, and then how does that convince you to buy the burger?

What was the transition from neuroscience to data science like?

I started my journey into data science when we used a technique called two-photon microscopy – a way to videotape the activity of neurons in the brain. You have a mouse, remove a small piece of their skull, replace it with a piece of glass and look into their brain. We can record hundreds of thousands of neurons in their brains, collecting terabytes of data quickly, but then you have to try to extract data and signal from that. There was no one else in the lab who could code at that point, so I taught myself image analysis techniques; experimental design... By the end of my PhD, I liked doing analyses more than I liked doing experiments.

I loved that I was in a place where I was doing analyses and computational work for most of my day. In academia, the ability to do that primarily is not common. If I wanted to do a postdoc, I would be doing a little bit less of the thing that I really enjoyed.

In the industry, there was a lot of desire for people who are good at thinking on their feet, articulating and simplifying complex ideas, and doing data analysis. I loved talking about my work, being able to explain something complicated and have someone who maybe doesn’t have a background in neuroscience at all understand why it’s cool. There are a lot of parallels with data science.

Sounds like a lot of continuities between the two worlds...

In a small lab, you become like a manager. I really liked the ability in data science to touch lots of different projects. My role now is a bit like a professor. There are lots a different workflows, lots of different projects going on simultaneously. And I get to hop in, touch this project to help steer it.

In both worlds, you need to be technically competent, an agile thinker, a good problem solver and good at articulating yourself. My PhD taught me to take nothing for granted; question everything. The way we approach advertising is: we’re not going to assume that how agencies or the industry has treated it so far is the right way. We look for a way to do it better.

You have a little more freedom in academia, right? Does it ever kill you to leave a project behind?

There’s a difference in timelines. Sometimes we need to turn something around in two weeks, while in academia it’s something you could take two and a half months on. In the industry, timelines are not yours to set. So you have to get very good at scoping; saying, ‘this is the type of problem I need to solve, this is the type of answer I need to get to, I have this much time – what are the best ways to get from A to B?’

There’s a growth curve there. Your ideal solution starts to shift, because you start reading what clients want.

Is a PhD necessary in data science?

A PhD in a quantitative field gives you a skill set. But it’s not everything by itself. You also need to have the right mindset, the desire to learn and the ability to shift. Some people can be a little bit too academic, and they can’t shift to different timelines and requirements.

Would you take the same route if you had your time over?

Yes. I learned so much from the people who I worked with. The people you surround yourself with define what you learn and how much you enjoy it. I enjoyed every minute of my PhD, but my path led me from neuroscience to Known and advertising, and I’m thoroughly enjoying this as well.

Neuroscience Science Agencies

Content created with:

Known

Known is a modern marketing company that pairs Ph.D. data scientists with award-winning creatives, expert research teams and strategists who leverage machine learning,...

Find out more

More from Neuroscience

View all

Trending

Industry insights

View all
Add your own content +