Advertising & Marketing Fundamentals For Data Scientists

I am an advertising and marketing veteran who is currently transitioning towards data science. The purpose of this write-up is to give you some baseline understanding of marketing, grounded in my professional experience. I am hoping that my write-up will help you gain a bigger share of voice when working with advertising & marketing teams. Eventually, you might ask bigger questions and thus move beyond just optimizing their work.

I will expand this post into a series depending on your and other readers' feedback. Just let me know what you think in the comment section below.

My most important advice to you is:

Don't think like an ad or marketing guy
Think like a consumer

This has been the most rewarding exercise in my professional career. However, it's a very inconvenient one because customers don't want the same things as advertisers and marketers. According to Forbes, an average consumer is exposed to 4,000 - 10,000 advertisements per day. That said, your marketing departments effort to increase email-marketing opening rates, for example, stands in sharp contrast to your target audience's attempt to minimize the volume of incoming marketing mails (aka spam). The same goes with every other ad & marketing measure:

It's your try to get on a consumer's radar vs. a consumer's effort to get you off his radar

Taking a (potential) customer's point of view, you might ask questions such as:

  • "Why are we even sending out email blasts to our target audience?"
  • "Are there possibly alternative communication channels?"
  • "Is our content even relevant from a customer's point of view?"
  • "Do we understand the conflict of interests: What we want from our target group vs. what they truly want?"
  • "Are we willing to act upon these insights?"

These, in fact, are dangerous questions because they undermine the status quo and thus the authority of those in charge. What if your company has done things exactly that way for ages? What if you are stepping on someone's toes asking those questions? Welcome to the world of corporate politics! Make sure you test the waters cautiously and observe your team's reactions. You might stir up some waters, and that's good if your company is willing to change things and move beyond keeping their current modus operandi.

Let me give you some practical advice about how you can get better at thinking like a customer.

Acquire the skill of talking to customers and field staff

As a data scientist, you might get odd looks if you tell your superiors: “I am going to be out of office for a week, talking to customers, dealers, service and support.” Think of the ability to talk to customers and field staff as of an art because it goes beyond just checking off questionnaires. If you ask a customer directly: "What do you want?" He might come up with a bunch of ideas, literally out of thin air. But would he even buy a product that was configured due to that wish-list? Instead, ask questions that make a customer talk about the product experience, and then derive insights from that. Let me give you one example.

Customers behave in ways you sometimes would not expect

Back ten years ago, I worked for a POS (point of sale) design agency. One of our biggest customers was a family-operated retail business with approximately 300 duty-free airport outlets at that time. One thing that our agency wanted to remove were bargain bins which we found quite disturbing. It looked odd to us that Gucci, Prada, and Dolce & Gabbana bags worth north of 1,500 USD apiece were heartlessly dumped into those bins. “Wouldn’t it be better to have high-quality touchpoints for each of those brands?” we thought. Our customer’s executive team was supportive of that idea. But then, luckily, we hit the road and talked to salespeople working in our customer’s outlets. “Tell us about those bins! How do customers feel about them?” we asked. Many salespeople shared the same observations: “Oh, customers actually love those bargain bins. Some of them storm our shops during a stop-over and they spend 5,000 - 6,000 USD on two or three of those bags within minutes. They feel like hunters who stumbled upon a rare, overlooked treasure.”

As a result, we didn’t touch the bargain bins. Although the total volume of luxury handbags sold was microscopic, we realized that those bargain bins attracted a wealthy clientele. This had a positive impact on our customers’ brand, as the vast majority of customers with average incomes want to spend their money where the wealthy do.

Use products your customers are buying

A couple of years ago, I read an anectodical story about Ferdinand Piëch, the former chairman of the Volkswagen Group. He reportedly had a meeting with his truck division’s executives, and they jokingly talked about driving trucks, until Piëch produced a valid truck driver license from his pocket. Being known as someone who has gasoline in his veins, Piëch had little tolerance for executives who are industry-agnostic (aka don’t care what they sell, as long as the pay is good). One executive allegedly got fired on the spot because he felt that acquiring a truck driving license was a waste of time.

Audi Quattro was a pet project of Ferdinand Piëch. It was an important milestone for Audi to claim their "Vorsprung durch Technik" ("Lead by Technology") | Source: Wikipedia (photo by More Cars from Berlin, Germany) 

Airbnb is also known for hiring people who are fanatic about their service. This approach has a fundamental impact on the company’s culture: People who work at Airbnb care about Airbnb because they use Airbnb in their daily lives. As a result, their customer experience is superb.

Do data scientists at Netflix talk to customers and use their product?

I am a huge Netflix fan, and I have a deep admiration for the company. However, I noticed a couple of oddities as one of their millions of customers. Why does Netflix:

  • recommend series based on my watch history, even if I disliked them? I found “The Haunting of Hill House” boring, and I downvoted it, but Netflix kept recommending further, similar series because “you watched this”
  • suggest continuing watching “El Chapo” which I have already consumed in full length?
  • not provide any information whether stellar series such as “Mindhunters”, “Suburra” and “Ozark” will be continued?
  • change the covers of the series on my dashboard so that it’s nearly impossible to find them, even if they are on my watchlist?
  • not inform about upcoming series which soon are soon going to be released?
  • not provide trailers for every single series but instead just for some of them?

I assume that Netflix might be overly obsessed with optimizing things such as recommendation engines which sometimes are of questionable value from a customer's perspective. Instead, they might shift their focus on asking questions such as:

  • "How can we improve the overall customer experience?"
  • "How do I feel about Netflix when using our service?"
  • "How does my wife/husband feel about it?"
  • "What do our children like about it?"

Move beyond optimizing the things you do
Think like a customer
Become a customer

Act upon your insights

I work as a data literacy expert, and I cater to large companies in Europe and the US. You have questions I didn't answer in my write-up, or you want to share your experiences? Please leave a comment or reach out to me via email [email protected] or LinkedIn. Thank you!

Views: 2539

Tags: AI, Advertising, Data, FMCG, Marketing, Science, b2b, b2c


You need to be a member of Data Science Central to add comments!

Join Data Science Central

Comment by Ramesh Gopal on November 16, 2018 at 9:52pm

Thanks, Rafael, for articulating the folly of what business executives and data scientists think they know, versus what the reality is.

It is indeed a thought-provoking piece and thanks for sharing.

Comment by Arnuld on November 12, 2018 at 10:40pm

I am not a marketing guy, I am software developer turned data scientist, but I learned few things from your post, e.g. data science is basically a tool, questions are more important. Tools, software development, websites, data science, Artificial Intelligence etc, all are secondary to the questions. In the end,  "recommend series based on my watch history, even if I disliked them? " .. I have same experience on youtube :)

Comment by John Williams on November 8, 2018 at 2:01pm

Awesome advice! I teach Marketing at a business school, and I'm thinking of making this post required reading for my classes.

© 2021   TechTarget, Inc.   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service