An AI Approach to Omnichannel
Marketers have access to more data on their customers than ever before. The challenge is getting rapid insights from all the channels used, in time to act effectively. One solution is an AI-powered customer retargeting platform for omni-channel marketing operations. Abhi Yadav, ZyloTech’s co-founder and CEO, spoke to me about the function of AI in omni-channel marketing.
ZyloTech, formerly DataXylo, launched in 2014 as an MIT spin off. Its advisory teams includes PhDs in AI, data scientists, and other marketing experts from the university. The platform uses machine learning to analyze all customer data continuously, and in near real time, for actionable insights on omnichannel marketing.
In the current environment, marketers really have “no way of knowing whether the individual” targeted by the ad “is a new, lost, inactive, loyal or brand-conscious customer,” says Yadav.
The problems is that marketers rely on “platforms like Google and Facebook” for their targeting, and “the underlying segmentation data and analytics are flawed.” Studies consistently show that even with the sophisticated big data technologies, and armies of data scientists and engineers, companies are leveraging just 10 to 15% of relevant customer data in weeks or months, Yadav explains. In a world of fast-changing customer behavior, this is too little, too late. Traditional analytic approaches are also a problem, because they are based on arbitrary segmentation rules. Customer behavior is far too complex for these simple approaches to be effective.
For marketing to be really effective, says Yadav, it has to fit what the customers want, and for that you need to know a lot more about them. In the Amazon and Netflix age, customers expect content to be tailored and timely.
“Given that the personalization bar is so high, brands must offer only very individualized offers and communications that may alter customer behavior, instead of tracking behavior in the rearview,” Yadav says, and AI makes this possible by sifting through the huge volumes of individual data to surface their “behavioral patterns” and point to “actionable insights.”
Best of all, the AI functions assures “freedom for marketers.” By “freedom,” Yadav means:
- Freedom from the chaos of big data
- Freedom from recruiting and hiring an army of data scientists. And freedom from chasing IT for perishable customer insights
- The ZyloTech platform works differently than the standard marketing campaign that Yadav says uses “ a top-down approach.” The set out the campaign and “gather the data with ad-hoc, manual approaches and data scientists” who can then apply analytics to direct the campaign. One of the problems with that approach is that the data in such cased tends to amount to “just 10-15 percent of easily accessible customer data, gathered within few weeks” rather than all the data on the customer.
That’s why ZyloTech flips this around for a bottom-up approach that makes the most of the customer data. That entails two layers of machine learning.
The “first machine learning layer – our dynamic data engine – gathers customer data from dozens or hundreds of sources and millions of data points.” All of it is then classified “into four categories: demographic, psychographic, behavioral and intent.” The next steps: curation and unification, occur in “near real time, without time-consuming Extract, Transform, Load (ETL) tools and numerous layers of processing.”
The second layer is the “embedded analytics engine,” which “analyzes this data continuously and in near real-time with five unique ‘lenses’: life stage, life style, value (RFV), potential spend and mission.” The result of that is “dynamic micro-segments” that the company refers to as “cohorts.” These produce a far more comprehensive and exact view of the customer for more accurate predictions of behavior and likelihood to purchase.
Yadav offers the example of marketing baby food to “two suburban moms.” While the “top-down approach would” assume the same approach would work for both, “a dynamic, bottom-up approach would reveal that one is a brand-conscious shopper, while the other would switch to a private brand if the price was right.”
Yadav sums this up as “we’re creating an AI-powered Customer 360 view with raw data points and metrics by each customer for more accurate, and more powerful remarketing.”