AI-powered engine boosts retailers' personalization efforts

computer
The Engine uses machine learning to understand changes in customer data and behavior and reacts accordingly given the current circumstances and the customer's history.

The Monetate Intelligent Personalization Engine is a new platform with an open architecture that takes in customer data regardless of where it resides, then uses artificial intelligence (AI) to make decisions in real time to serve customers based on individual needs. The machine executes on the decision in milliseconds, delivering a personalized experience based on predictive power.

Handling nuances is where the Engine shines, according to Lucinda Duncalfe, Monetate's CEO. It uses machine learning to understand changes in customer data and behavior and reacts accordingly given the current circumstances and the customer's history. 

Monetate began testing the platform several months ago with 15 brands, including big names such as Club Monaco, JD Williams, Office Depot, Toms Shoes and Talbots. 

Now in full rollout, retailers can use the technology depending on their business goals. 

"The location of the experience matters a lot in your priority. Experiences that are highly exposed to a large number of your potential customers are critically important and will respond well to machine learning given the large amount of sessions/traffic," Duncalfe told FierceRetail. "In addition, there are many experiences that are, by their location and/or rules, targeted at a brand’s most valuable customers. Because of this, they may be good candidates for individualizing this content. Impacting your most valuable customers first is critical."

The platform can be used to serve up experiences that optimize a wide range of conversion or engagement metrics, such as revenue per session, email sign-ups and decreasing homepage bounce rates. 

For example, a recent Monetate client used content marketing to drive direct conversions from their blog. The blog did not command a lot of traffic, so the company displayed different blog content on their e-commerce homepage, dynamically, and the results were more than $140,000 in net revenue lift.

In another example, Duncalfe talks about the importance of retailer's Product Description Pages  (PDPs), which are often subject to a lot of content and copy that can help a visitor make a buying decision. However, PDPs can quickly become overwhelming when product details, pricing, discounts, reviews and other content are not ordered in a manner that is most relevant to the individual. Using The Engine, a Monetate client reorganized their PDPs by effectively determining when it was advantageous to emphasize customer reviews or related product recommendations. This content already existed on the PDP, but it was up to The Engine to determine which to display more prominently. Deployed across two of the client's sites, The Engine achieved 4.8% lift in revenue per session (RPS) on one site and 1.3% lift in RPS on the second site.

Duncalfe admits there are some challenges with integrating this program into an existing framework. She calls it the "personalization paradox," the perception that one-to-one isn't achievable because it's cost-prohibitive and impossible to scale. However, it's really not true because it's not creative for each person, but about providing the best experience for individuals based on everything the machines knows about them. 

"This approach is far better than optimization approaches that maximize the results for those that are pleased by the chosen experience, but at the same time, alienates those that are not responsive to that experience," Duncalfe said. "Getting over this mindset is an important first step."

Monetate set out do personalization nine years ago, but the market wasn't fully ready. But after initial testing, the company sees the market is now in place. As far as The Engine itself, the company has learned that the longer the algorithms run, the more it improves and provides useful insights to marketers. 

"We believe that this AI-driven approach is the future of marketing. It gives customers what they want: an experience that reflects their relationship with the brand; and it gives marketers what they want: better performance with less effort," Duncalfe said. "We also believe that the tried-and-true methods of A/B testing/optimization and segmentation/targeting will continue to be essential tools in the marketer’s toolbox. For example, if a retailer wants to test a concept for a future catalog on their website, then testing is the best bet."