Next Best Action versus Multi-Armed Bandit Experimentation

Traditional marketing automation from companies like Salesforce and Oracle, Artificial Intelligence Marketing (AIM) platforms from companies like Adgorithms and Amplero, and Next Best Action (NBA) decisioning from companies like Pega Systems are all trying to solve for the same business challenge: How do marketers leverage the volume and depth of customer data available to truly deliver personalized experiences at scale?

While legacy marketing automation works well within a static, pre-defined user journey or smaller data sets seen in some B2B or transactional retail environments, NBA and AIM attempts to address the ever-evolving nature of both market conditions and consumer behavior.

In recent years, new approaches focused on using algorithms to drive NBA have grown in popularity as seen in tools often bolting-on capabilities to existing business process management tools such as Pega Systems’ Customer Engagement product or Oracle’s Real-time Decision Platform.

While these approaches are helpful in using computers to solve narrow use cases (e.g., which offer to serve a customer at risk of churning), they have some pretty serious limitations.

For example, they are only capable of optimizing for short-term, direct response metrics such as opens, clicks, or immediate offer uptake. Furthermore, typical NBA approaches can only scale to leverage and optimize using a handful—typically 3 to 5—customer attributes, such as the customer’s current plan, their current CLV calculation, and their reason for churning

At Amplero, we refer to solutions such as NBA that are solving very narrow use cases as bolt-on AI. They are lightweight AI models (developed, deployed and often optimized by humans) to solve a narrow use case such as offer optimization, email send time optimization, or product recommendations.

Origins of Multi-Armed Bandit Experimentation

Companies such as Amplero are using Machine Learning + Multi-Armed Bandit Experimentation (MAB) to enabler broader modeling (auto-generated by machines not people) to optimize for more meaningful/impactful business KPI’s such as 30-day revenue, 90-day retention, and more.

The term multi-armed bandit (MAB), which traces its origins to the 1950s, describes a metaphorical casino in which a gambler faces a bank of slot machines (also known as “one-armed bandits” —hence the term MAB).

The machines have variable unknown payout characteristics (amounts and frequencies) and the gambler must methodically decide which arms to pull to maximize winnings while searching for the best machine in a sequence of fixed bets. More generally, it describes any situation where the value of exploration — trying a relatively unknown option — competes with the value of exploitation — choosing an option whose results you can count on — at any given stage in a repetitive decision-making process. This trade-off has wide applications in business.

Advantages of MAB for enterprise marketers

The advantages of using ML + MAB is that it allows your omni-channel marketing efforts to automatically and constantly be optimized for every interaction across the customer journey, focusing on driving KPI lift for the types of metrics that business report on to the street — revenue per customer, churn rates, customer lifetime value instead of being relegated to using vanity metrics such as offer up take or clicks/opens.

Currently, leading B2C enterprises are applying AI at the core, using ML + MAB to move away from running one-off, ad hoc campaigns to an always-on porfolio approach comprised of various strategies that the machine plays, optimally balancing exploitation with exploration to ensure mathematically optimized lift on core KPI’s.

Specifically, using Amplero, marketers in competitive, customer-obsessed industries like telecom, banking, gaming and consumer tech have seen measurable lift in key performance indicators, including a 1-3% incremental growth in customer topline revenue and 3-5x lift in retention rates, while delivering better customer experiences (fewer touches, higher relevance, earlier intervention).

Download the full case study

About Amplero

Headquartered in Seattle, Amplero is an Artificial Intelligence Marketing (AIM) company that enables business-to-consumer (B2C) marketers at global brands to optimize customer lifetime value at a scale that is not humanly possible.

Unlike traditional rules-based marketing automation systems, Amplero’s Artificial Intelligence Marketing Platform leverages machine learning and multi-armed bandit experimentation to dynamically test thousands of permutations to adaptively optimize every customer interaction and maximize customer lifetime value and loyalty.

With Amplero, marketers in competitive, customer-obsessed industries like telecom, banking, gaming and consumer tech are currently seeing measurable lift across key performance indicators—including 1-3% incremental growth in customer topline revenue and 3-5x lift in retention rates.

For more information, contact us today or follow us on Twitter.