Machine learning – a sub-application of Artificial Intelligence (AI) that allows computers to analyze, interpret, and otherwise “learn” from data, has transformed business, allowing them to analyze and act upon massive datasets in real time to improve operations, logistics, and business decisions. Now machine learning is set to transform loyalty marketing as profoundly as the advent of the internet transformed the discipline twenty years ago. How can marketers best incorporate these powerful platforms into their customer strategy? The experts weigh in.By Rick Ferguson
We begin with Customer Think contributor Bob Hayes, who matter-of-factly describes the potential benefits of machine learning to loyalty marketers:
“Iterative in nature, machine learning algorithms continually learn from data. The more data they ingest, the better they get. Based on math, statistics and probability, algorithms find connections among variables that help optimize important organizational outcomes, in this case, customer loyalty. Coupled with the processing capability of today, these algorithms can provide insight quickly to improve marketing, sales and service functions.”
Media Post contributor Peter Lim, meanwhile, is even more bullish on the power of machine learning to unlock the potential inherent in customer data:
“Machine learning won’t just look at data and point out obvious facts. Instead, it will uncover correlations between groups of data sets. In other words, machine learning will comprehend exactly what the data is saying. It’s precisely this insightful aspect of machine learning — its ability to see patterns and understand consumer behavior — that equips marketers with the information they need to better segment and target their customers.
“For example, the technology can be applied to the creation of a recommendation engine, which is based on a customer’s prior behavior instead of an artificial set of perimeters or rules. It can predict the preferences of shoppers, travelers and diners and develop actionable insights such as which customers get which types of offers.”
Drilling down, SAS executives Wilson Raj and Suzanne Clayton highlight the potential of AI to provide that most elusive of insights – cross-channel customer identification:
“When customers feel they are not being recognized across channels, it can lead to churn and decreased satisfaction. Customers are permanently connected, and they expect to be treated as unique. In the vast sea of messages and offers customers receive each day, they have only time and patience for content that is personalized and relevant. Social media has empowered customers to expect nothing less and has also given them a forum to broadcast a negative experience.
“The good news, though, is that with all that bouncing from channel to channel, consumers leave many data traces, and all that data adds up to become big data. By analyzing these big data traces (structured and unstructured, residing within the company walls and outside of the company), marketers can get to know their customers with an unprecedented level of detail. For example, by using natural language processing techniques like text analytics, [marketers] can better understand how customers feel and why.”
And MarTech Advisor contributor Andrea Wildt illustrates the potential benefit to reducing customer churn:
“While the volume of data captured increases from every new channel or input, the analysis still needs a human touch. With insight into these analytics, marketers can identify how a certain segment of the market typically behaves or predict future patterns. But analyzing this for millions of customers across a dozen (or more) touch points is more than any marketing team can handle. Digital growth company Urban Airship, for example, has developed a machine learning algorithm that analyzes mobile customer behavior to help app publishers identify the most loyal users and predict those that are likely to churn. Armed with this insight, marketers can take action across digital channels to deepen customer engagement or invest more in retaining specific customer segments.”
If you’re a traditional loyalty marketer operating a program designed to identify, understand, and influence best customers, you may wonder if machine learning might eventually render your job obsolete. After all, one of the most effective ways to identify customers across channels is to link interactions to the unique loyalty identifier of each member. If AI can now track customer interactions across channels, then why have a loyalty program?
Similarly, if machine learning can learn how individual customers feel about your brand and experience, and then craft personalized, relevant messages based on that learning, then why do we need a loyalty program to deliver reward and recognition to best customers? If an algorithm can learn what offer will best prevent a customer from leaving, then why include a mechanism as cumbersome and expensive as a loyalty program to deliver that offer?
Just as the internet eliminated the need for paper reward catalogs and mailed quarterly statements, there are certain aspects of legacy loyalty program that machines will no doubt render obsolete. The combination of mobile wallets and customer identification facilitated by machine learning may end the need for a distinct loyalty identifier – although unless you’re already incorporating visual identification in your retail stores, you may still need the identifier to connect in-store to digital activity. The ability to deliver real-time offer personalization based on ever-learning algorithms may also render traditional loyalty campaign management superfluous. AI is disrupting legacy business functions in every industry, and there’s every reason to believe that loyalty marketing will undergo the same disruption.
There will remain, however, the need to segment customers by value, or much of that investment in machine learning will accrue to customers who aren’t interested in spending more with you or telling their friends about you; they’ll merely take up your offer and then skip merrily over to your competitor. As reciprocity is critical to any long-term relationship, there will also remain the need to reward and recognize your most valuable customers beyond mere personalization. Much of this activity will remain consolidated under some sort of branded customer program – for without that loyalty brand, it’ll be difficult for best customers to understand that you’re treating them with special care. Amazon, for example, could simply incorporate Prime elements into their core offering – but then how could they convince you that those disparate elements are worth paying $99 per year to access?
With these considerations in mind, we may humbly suggest that machine learning won’t replace loyalty marketing; rather, loyalty marketing will evolve to incorporate the latest applications in artificial intelligence. Loyalty marketing isn’t a software platform, a reward catalog, or a points balance; loyalty marketing is an established marketing discipline designed to build long-term relationships with your most valuable customers. Machine learning is merely another tool in the loyalty marketer’s toolbox.
Besides, as Peter Lim reminds us, the human element in marketing remains essential:
“It’s imperative to note, however, that machine learning cannot replace human marketers. The industry cannot be fully automated. In fact, machine learning is a tool that only works when it’s used in combination with human expertise and analysis. Turns out, two brains – one human, one machine – really are better than one.”
Amen to that, and welcome to the machine.
Rick Ferguson is Editor in Chief of the Wise Marketer Group.