Black swan events, like the global pandemic, show the shortcomings of analyzing historical data to predict buying behavior. Perhaps the only thing for certain is that people facing a cataclysmic event will not behave normally.
What marketers need is a way to understand the customer beyond historical purchasing patterns. This isn’t to say history doesn’t matter. Think of it as knowing your best friend after years of interactions and achieving a level of understanding about how they will likely behave in unprecedented situations.
Do we have such a system? Can it learn about people to this degree quickly and at scale? Absolutely. It’s called artificial intelligence, or AI. Much has been written about AI, with the slogan “powered by AI” appearing on all sorts of marketing collateral. This has resulted in a lot of confusion. I’ve covered AI for years, and the best definition I’ve seen comes from Michael Wu, chief AI strategist at PROS, who writes:
#AI = a machine mimicry of human behavior with two characteristics: (1) the ability to #AUTOMATE human decision and subsequent actions, (2) the ability to #LEARN and improve its performance with usage.
Let’s consider this definition in a marketing context. Mimicking and automating human behavior is like having a marketer with the brain power of a superhero. The notion of learning, improving and adapting puts an emphasis on future behaviors, not historical models. What’s the net effect? Imagine having millions of friends and being able to predict with a high degree of accuracy their buying behavior in new purchasing scenarios.
This is the promise of AI and the start of a company’s ambitious journey to become a customer-intuitive enterprise. Isn’t this the future of marketing?
“AI is going to be a game changer in every aspect of marketing,” writes Raja Rajamannar, chief marketing and communications officer and president of healthcare business at Mastercard, in his new book Quantum Marketing. “If they don’t understand the way it works and the possibilities it can bring, they will be missing the boat. AI will never replace marketers. But marketers who resist it will be replaced by those who understand its power.”
AI has the potential to greatly impact campaign planning, media buying, targeting, personalization and other critical marketing functions. It might soon be able to create content, including advertisements, without the involvement of human marketers. Rajamannar says some AI solutions can already estimate ROI in advance of a campaign or promo — a step towards AI helping to solve marketing’s problem of business-outcome attribution.
Yet many marketers simply don’t have a good grasp of AI. They don’t know enough to separate fact from fiction, and this can get them into trouble.
Let’s say, for instance, a marketer relies on AI to analyze mountains of customer data and deliver actionable insights in the form of the best discount for a given customer or customer segment. Specifically, the “powered by AI” software analyzes past customer behavior, the customer’s current propensity to buy in the category, and buying behavior in other categories in regards to price and discounts. Then the sleek software churns out authentic-looking discounts that will be automatically offered to the customer in real-time.
But are they the right discounts? Maybe the discount is too high, thus leaving money on the table for the company. Maybe it’s too low and doesn’t convert enough customers. Or maybe the AI is racially biased, in which case the unfair discounts pose a serious risk to the brand.
“If you don’t understand what is actually happening inside AI’s black box, you will be in a dangerous position,” Rajamannar told the CIO Council. “Your marketing team needs to be able to decipher and understand it.”
How do you start? Rajamannar says marketers can begin with low-cost pilots with a small project and open-source AI. They can take online education programs from Harvard, MIT and the University of California, Berkeley. He suggests marketers collaborate closely with IT, which has the technical wherewithal to debunk “powered by AI” software.
Marketers need at least a basic knowledge of AI on their team, say, a marketing AI specialist. Much like a marketing statistician in the old days, this person should be technically savvy enough to consider many factors affecting output and be able to discern if AI is producing good insights. In other words, an AI specialist can help demystify the black box.
It’s time to take on any or all of these efforts, as AI becomes the engine of marketing.
“From learning about consumers deeply to enabling hyper-personalization to optimizing programs on the fly, AI can hugely enhance marketing effectiveness and efficiency,” Rajamannar writes in his book. “At this stage of the evolution of AI in the context of marketing, it is important for marketers to get their head around it.”