dog doing a handstand
Editorial

How Machine Learning is Upending Marketing

5 minute read
Mike Flannagan avatar

A marketing makeover is underway.

Machine learning is shaking up standard practices and enabling radical improvements around personalization, customer segmentation, and campaign performance.

For early adopters of machine learning, statistics show they are edging out their competitors and creating impossible-to-ignore customer engagements.

Machine Learning Gets Results

Predictive marketers who use machine learning are 2.9 times more likely to report revenue growth at rates higher than industry average, according to a Forrester Research study. They are also 2.1 times more likely to “occupy a commanding leadership position in the product/service markets they serve.”

Clearly, machine learning is giving these marketers an advantage. Customers are responding to the more personal offers that are in tune with their lifestyle. When companies offer relevant offers for the right product, at the right moment, at the right price, customers pay attention.

Poland’s largest online bank and fourth largest retail bank — mBank — fits squarely into this new breed of predictive marketers. Using predictive analytics with machine learning, the bank increased results for non-mortgage loans, insurance products and savings products from 200 percent to 400 percent.

A large portion of mBank’s 4.5 million customers pay by credit card, giving mBank a good understanding of their purchases.

Insights Into Customer Behaviors

By combining customer purchase history with a customer's recent activities, mBank learns its customer's preferences. That machine learning allows mBank’s marketers to predict which offers will be of most interest to each customer. The eye-opening results speak for themselves.

To widen appeal, mBank added personal offers outside its portfolio, bringing in bonus partners. The marketing team sends special deals from restaurants to customers who eat out often, home renovation offers to a recent home mortgage applicant and coupons for high-end consumer electronics are sent to known gadget lovers.

In this way, mBank is monetizing its customer data and creating future new business models. The personalized, highly relevant offers are raising the bank’s presence among customers and marking it as a leader in customer communications.

New Rules in the Machine Learning World

Machine learning provides the insight mBank marketers need to take action and drive better business outcomes.

When armed with machine learning that can predict customer actions, marketers are creating a new set of rules and best practices that deliver positive, personalized customer experiences. In fact, predictive, machine learning marketers are building customer-centered, one-to-one strategies around these new cornerstones.

Location, location, location is a relic. Knowing a customer’s physical location has been premium data for decades, and more recently geo-location has led to significantly better campaign performances. These campaigns, unfortunately, are one-dimensional.  

Location is merely one data point in the vast amount of customer knowledge available to marketers. Omnichannel marketing provides a complete picture of a customer. Marketers can apply machine learning to structured, unstructured, historic and real-time customer data to create a unique, highly personalized campaign for a sticky, long-lasting engagement.

Learning Opportunities

Introducing the new customer segment of one. Traditional campaigns cast a wide net targeting all women ages 30 to 35, living in Seattle, WA. With machine learning, marketers can create a campaign for a 35-year-old woman who attends Seattle Sonic basketball games, purchased a BMW five years ago and recently had an accident.

Predictive analytics can put a local dealer’s new car offer with a free home game ticket directly into that BMW-driving, basketball fan’s email, smartphone or mailbox.

High-performing creatives will come from machines, not humans. Marketers spend lots of money with design firms to create online, mobile, and print promotions. Machine learning could make those groups obsolete — or at least give them serious competition.

Marketers can apply machine learning to promos by taking all the raw material, stock images, copy, and call to action, and then predict which elements will be most appealing to different sets of customers. The software can even assign placement of all the elements — without human intervention.

Highly personalized incentives and promotions don’t require data scientists. Marketers have terabytes of customer data that neither their brains or spreadsheets can possibly process. Sophisticated, off-the-shelf machine learning software lets marketers get inside that wealth of data and manipulate it, visualize it, and ultimately act on the insights.

Marketers: The New Data Scientists

For the first time ever, marketers can pose queries, make comparisons and easily step into classification, regression, clustering and time-series forecasting models that had been reserved for data scientists.

Business-savvy marketers can create hyper-personalized incentives and promotions with machine learning and predictive analytics without help from IT. Logically, marketers know the customer better than IT does. Marketers know what questions to pose to produce the models that will lead to relevant, hyper-personalized customer offers.

Machine learning is sure to break other marketing rules. This is just the start of creative and curious marketing teams redefining how they engage with their customers.

With the advantage of getting smarter and smarter with every interaction, machine learning helps marketers stay in sync with their customers and ahead of their competition. 

About the author

Mike Flannagan

As Senior Vice President of Analytics at SAP, Mike Flannagan is responsible for Product Strategy and Product Management with a focus on building products that deliver business insights to anyone, anywhere, on any device with a world class user experience.