Asking the Experts

The Rise of the Marketing Data Scientist

Tips for data scientists in the increasingly data-centric marketing world

Keith Loria
August 31 2019

Modern marketers want data-driven actions and results, and data scientists are in a unique position to show them the way.

Drawing on expertise in text mining and content analytics, data scientists can help marketers accurately classify customers’ voices through machine reading surveys and social media data, targeting the right offers to the right segments at the right time using machine learning models like collaborative filtering.

Not only will this increase marketing effectiveness and revenue, but it renders data scientists all but essential to team operations.

More from PostFunnel on the marketing team structure:
Grow Your Marketing Department: 5 New Roles to Watch
What is a Chief Digital Officer – And Does Your Company Need One?
The Role of Product Marketers in the Marketing Ecosystem

Yuval Ben-Itzhak, CEO of AI-powered social media marketing platform, Socialbakers, said that data scientists in particular are seeing big opportunities in digital marketing, as the scale of data from digital channels is far beyond any other source.

“As digital marketing data is created by people engaging with content, the wide variety of attributes and behaviors embedded within the data is an exciting treasure to discover,” he said. “Having this reality, a data scientist who wants to be successful in digital marketing should develop skills and experience in sourcing data from a wide range of systems, some of which are in real-time and at a very large scale.”

For data scientists, storing and retrieving data in a cost-effective manner are mandatory skills, as piling data indiscriminately can become costly very quickly. Therefore, leveraging data processing technologies like Spark and programming languages like Python are required for efficiently processing large amounts of data.

“On top of these skills, the data scientist should also understand how audiences engage with digital content and decode the behaviors embedded deep into data sets by leveraging modern machine learning and advanced statistical algorithms,” Ben-Itzhak said. “As marketers can no longer analyze digital channel data by using spreadsheets or analytics dashboards but rather must drive actionable insights out of the data, the data scientist’s role within marketing groups is becoming more and more important.”

Find Meaning

RedPeg’s senior director of brand strategy, Matt Sincaglia, said that strategic marketing means connecting a brand with what its audience finds meaningful. Understanding the audience means learning who they are beyond demographics and looking to passions, behaviors, motivators, and values for a bigger picture of who they are.

“You must understand who is consuming the marketing message. Data scientists must uncover whether the brand message is actually being consumed by the intended audience or run the risk of it failing to achieve goals,” he said. “It can be equally important in uncovering new audience types.”

He suggests that data scientists create hierarchical methods of defining performance, as marketing is all about optimization.

“The more a data scientist can set up an easily defined structure for results, the more valuable they will be in providing insights to the marketing team on what is effective and what is falling short of goal,” Sincaglia said.

Making a Connection

Zachary Jarvinen, head of AI technology strategy for OpenText, said that data scientists should feel at home jumping into marketing activities, as marketing today—at least those that are doing it right—is driven by data.

“Unfortunately, there are still many companies stuck in the past of ‘spray and pray’ marketing, and data scientists who find themselves in one of those roles should be prepared for a bumpy road,” he said. “A company that’s bringing in a data scientist shouldn’t be one of those. Or, at least be on the journey to evolve beyond that.”

Additionally, data scientists should remember to avoid speaking in ‘techy’ language.

“Many marketers’ eyes will gloss over if you begin to explain how Apache Spark works or show them Python code,” Jarvinen said. “Instead, talk about the outcomes you’re looking to achieve by adding data science to a specific project. Or, if you’re really good at communicating, try translating data science into the marketing language and jargon you hear the team around you using.”

Another helpful key to establishing a good relationship between data scientist and marketer is to turn data into visualizations.

“Whether it be a dashboard or infographic, know that most marketers are visual learners,” Sincaglia said. “On the whole, we don’t do great with Excel sheets. The more data takes a visual form, the easier it will be for decision makers to follow and react.”

Understanding the Medium

There is a core set of marketing concepts data scientists must understand to offer anything of value.

Vice president of product for Transform, Inc., Evan Dunn, said that perhaps most important is the distinction between inputs and outputs as it relates to marketing.

“I/O (input/output) models are common data science terminology referring to systems where influencing factors (inputs) go in and determine results (outputs). In marketing, inputs are budget allocations, ad creative, and audience targeting,” he said. “The output of the equation is conversions, purchases, net new customers, revenue. Results could also be mid-funnel KPIs like impressions and clicks.”

Understanding the lexicon gives data scientists the ability to develop quantitative attribution models (media mix models) that recommend shifts in budget, creative, or audience to improve marketing impact.

Dan Jenkins, vice president of data science at AdTheorent, said that not all data is created equal, and more data isn’t always better.

“There is an unprecedented amount of data available in digital advertising, but much of it is messy or low-quality,” he said. “Focusing only on the highest quality and most statistically significant data is a good way to create quick practical solutions. Additional data can always be added after careful cleaning and analysis.”

Kyle Ackerman, senior analyst at R/GA Austin, said that unlike financial analytics, when working with marketing data, it’s vital that data scientists consider the wider cultural and human context.

“As data scientists, we sometimes become too rigid in our thinking. We need to push ourselves to think more freely and creatively,” he said. “We’re dealing with emotional, cultural, and societal triggers that impact consideration and purchasing decisions. This type of larger thinking is the boon for our success in marketing.”

How to build your customer model

Keith Loria

A graduate of the University of Miami, Keith Loria is an award-winning journalist who has been writing for major newspapers and magazines for close to 20 years, on topics as diverse as sports, business and healthcare. You can view some of his recent writing at

Leave a comment:


Let's Get Acquainted

Chars: 0

Get your free print edition!

Fill out your complete details below

Chars: 0
Chars: 0
Chars: 0
Chars: 0
Chars: 0
Chars: 0
Chars: 0
Chars: 0