Author: Akin Babu Joseph
Motivation – Solving a Marketing Problem
A large ecommerce company has allocated a hefty budget for marketing activities before its upcoming sale event—its biggest in the year. The marketing team needs to come up with the messaging to be delivered during the marketing campaign to attract maximum customers.
The big question is what should that messaging be?
Should it be, “Get the best discounts,” or should it be, “Get the best international brands,” or will “Get best quality at best rates” work better? Should all three messages be employed?
This is clearly a tricky situation, with millions on the line. The company, being smart, wishes to use data to answer these questions. They naturally reach out to the data scientist in the marketing team.
The data scientist proposes an intuitive and powerful solution:
· First, identify the major types of customers.
· Next, determine what each type cares about.
· Finally, choose a message that would appeal to each customer group.
Fast forward a few weeks. Using this approach, four major customer segments were identified. By studying their characteristics, what they cared about was inferred, and messages were tailored to each group. The company achieved their marketing targets, attracting all types of customers and making the sale event a big success. Such is the impact that a data scientist in marketing can create.
Modern Marketing Analytics
Marketing has been around for centuries in varied forms. In the 20th century, with the advent of radio and television that made reaching huge audiences easier than ever before, brands began to spend millions of dollars to gain customer attention. By the 70s, marketing analytics had evolved as a discipline, and many of the common models we use today (Four Ps and Market Mix Modeling being prominent examples) had been developed. Around 1990, wireless telephones and a more evolved mass media made it necessary for brands to take a scientific approach to marketing. Marketing analytics was flourishing.
With personal computers, the internet, and search engines, the landscape suddenly changed for marketing. Brands were now eager to build their presence and gain a share of voice in the digital space too. Digital channels provide far richer data and can help generate a much better
understanding of the customer. Marketing efforts are now Omnichannel—meaning, online and offline. This digital revolution has caused a paradigm shift in the way marketing is done and gave birth to Modern Marketing and Modern Marketing Analytics.
Not only are the customer touch points and marketing channels far more, adding complexity to measurement and attribution, but the volume of data generated is exponentially larger. Data earlier used to be mainly structured, but now, unstructured data in the form of tweets, blogs, and images are dominant. With huge sums being spent on affiliate marketing and SEO, effectiveness of these channels, the quality of traffic coming from these, and the conversions of the individual campaigns need to be assessed quickly. Customer segmentation needs to be done on enormous data—both structured and unstructured. Clearly, modern marketing analytics is far more complex and needs a modern data scientist.
The Modern Data Scientist in Marketing
The data scientist in marketing needs to be updated with not only the latest tools and techniques but should also have the mindset to create business impact in a dynamic and fast-paced environment. Along with having a good understanding of the domain, the modern data scientist needs to be:
· Conversant with modern tools of the trade
· Conversant with the techniques required
· Comfortable working with large volumes of data
· Comfortable working with messy data
· Application focused, to solve problems and create impact using data
The modern data scientist is a new breed of data scientist that doesn’t get intimidated by messy data, is conversant with the tools of the trade, and is eager to employ their skills to generate tremendous value from data, which makes them a valuable asset to any brand/company.
I want to work to become a marketing data scientist, but…
The following are some of the more common myths that deter professionals from becoming a data scientist in marketing.
1. I need to have a PhD in Computer Science or Statistics
No, you certainly don’t! While there could be some niche roles requiring mastery in a very specific skill, business cares about impact. If you can solve problems and bring value to the business using data, you will be tremendously valued.
2. I need to be an expert in Deep Learning and Artificial Intelligence
This myth is probably perpetuated by, ironically, the marketing done by self-declared champions of Deep Learning & AI (sigh!). Data science in general (and in marketing in particular) is far more than just building predictive models using Deep Learning. It is about solving a business
problem, which can sometimes be accomplished via smart visualization, or even the right pivot table view, or maybe assessing whether a marketing campaign worked or not.
3. I need to be an expert coder
Modern tools of the trade are now very evolved. You not only have excellent drag-and-drop tools for complete beginners, but also great programming languages like Python which are very mature now and allow you to do practically anything with a few lines of code. Predictive modeling, customer segmentation, data analysis… All of these require only a few lines of code. Anyone can get started.
4. Automation will take my job away; the career will be short-lived
Far from it! Instead, you can expect automation to take over the mundane tasks in the data science process, thereby allowing you to focus on the more interesting aspects: solutioning and problem-solving. Automation will help you be a more effective data scientist.
Fortunately, despite these myths, there are many young professionals who are exploring data science today and joining the industry in rewarding careers. Marketing analytics is one such space that is looking for dynamic professionals that can analyze the huge piles of data, derive actionable insights, and create value.
Long Story Short
Marketing has come a long way over the years; and marketing analytics as a science has evolved alongside it, providing tremendous benefit from several applications. The digital revolution in the 90s not only changed the way brands interact with their customers but has also created new complexities that need to be handled during analysis. Modern marketing analytics, therefore, needs to work on a higher volume, variety, and velocity of data. Accordingly, the modern data scientist needs to be comfortable working with these complexities and the tools of the trade. These skills are not esoteric, and despite the myths, data science in marketing is welcoming of individuals who can create value from data.
Marketing analytics is an exciting and evolving space. The modern data scientist will thoroughly enjoy being here and will learn and evolve along with the field, shaping for themselves a rewarding career.
Learn more about data science in marketing with the book Data Science for Marketing Analytics by Mirza Rahim Baig, Gururajan Govindan & Vishwesh Ravi Shrimali