Like artificial intelligence or blockchain, data science is a popular buzzword in the industry. Marketers are often confused around the distinction between data scientists — those who design and test experiments using statistics, calculus, linear algebra — and data analysts — those who use spreadsheets to implement strategy around data.

In the latest installment of our Confessions series, where we exchange anonymity for honesty, Digiday spoke with a data scientist inside the marketing department of a company who says marketers are still lost when it comes to the science and are wasting money on data scientists.

Do marketers understand data science?
Marketers don’t know what they’re asking for when they ask for a data scientist. It’s the Wild West, especially if a company has never dealt with data. Companies that aren’t ready for a data scientist, hire one and end up wasting $100,000 a year. Most small businesses don’t need a data scientist; they need someone to handle a spreadsheet or a data analyst. Marketers are often confused about whether they want data analysts or data scientists and interchange the terms. If you look at the variety of analysts jobs available on Linkedin, if a company asks for a person who can work in Excel, they’re looking for a data analyst, but if they are talking about R, Python or machine learning, they’re looking for a data scientist, regardless of what they call the job.

What confuses marketers?
Data science is such a vague, meaningless word in the industry. It’s a buzzword, and because it’s such a buzzword, it’s easy for anyone who touches data on a day-to-day basis to call themselves a data scientist. There’s a lot of people trying to hack their way in. It’s like bitcoin. Everyone wanted to buy it. Everyone wants to be a data scientist right now, but you have to know what you’re doing. Even the word “model.” Depending on whether you are talking to a data scientist or a marketer, you could be talking about very different things. It’s not just the industry’s fault.

What else is fueling the confusion?
Schools popping up around us, like coding bootcamps, are perpetuating the idea that you can do data science without knowing statistics. That is bullshit. Coding bootcamps will take people and put them through a class where they’ll learn code formulas and then call them data scientists, but they’ve never taken a statistics course or a linear algebra course so they don’t know what they’re actually doing. Anyone can type LM into an R terminal and get a linear model, but that doesn’t mean they actually understand what a linear model is doing or how predictive it actually is.

How does this hurt companies?
There’s a disconnect between the extra level of complexity done at the data science level, what large companies are really looking for, and the kind of analysts being pumped out by non-academic institutions. Whenever you find places that say how little math you have to do to become an analyst, they’re pretty much turning out pretty weak analysts. Anyone can look at any company’s performance for the past six months, take the average and say next that month, you’re going to do this. That’s not a real prediction; that’s about as weak as it gets and that’s the basic level of analyst they are getting. So if a company is missing its KPIs month over month, it’s probably because their data scientist didn’t know as much.

Is it strange being inside a marketing department? Would you rather be working at a tech company?
For me, I’m doing my job right if I can give my company and stakeholders the data they need to make decisions. It’s a great thing to be part of, no matter where you are at. You’re solving a puzzle every day. How can I predict future behavior? That’s kind of cool. Throughout the history of mankind, people have loved fortune tellers.

Are you satisfied with what you make? 
I get paid a little less than the median, but it’s because I have about five years less experience. Companies recognize how competitive it is, so it’s a well-paying career. I make $90,000 and the median salary is $120,000 and could go up to $300,000.

  • LinkedIn Icon