Blog

AI-Based Personalization Surpasses its Rules-Based Counterpart

Posted on: March 18, 2020 | Posted by: Janet Jaiswal

Why marketing leaders rely upon AI-powered personalization

Customers now expect increasingly personalized shopping experiences that take into account their individual expectations, likes, dislikes, and intentions. Retailers are able to personalize interactions, messages, information and offers when dealing with a relatively small number of customers. However, to personalize experience at scale — for every customer, taking into account the volume of data streaming in from a variety of channels –they need to take advantage of technology. 

In particular, retailers that approach personalization from an omnichannel perspective struggle to scale without the right technology. To personalize in real-time, they need to understand customer in-the-moment intent. That requires tools capable of ingesting data from multiple channels, enriching and decisioning in real-time. Even then, retailers must deliver a contextually relevant personalized experience, whether that’s to a device with a user interface, or to enable interaction via the phone or with a live representative. 

Context for personalization is fragmented across multiple touchpoints

Businesses have used rules-based technology to render personalization to date. Now, with the advent of AI, the possibilities have increased manifold. Below, is a deep dive into the two technologies.

Understand the two key limits of rules-based personalization 

While brands have called upon rules-based personalization with some success, this approach falls short in two key ways according to Brendan Witcher VP, Principal Analyst at Forrester.

  1. A rules-based engine cannot improve over time
    Rules-based personalization relies on an impossible amount of human intervention and judgment. A developer would need to consider every data point and variable associated with each customer. They must then determine what precisely is causing a customer to behave a certain way – or which data points are simply correlated with customer behavior. The work doesn’t end there. A developer would then need to determine the impact of that correlation or causation in relation to every combination of factors. Imagine doing that across every customer every time a brand captured new data about that customer. No human is up to the task.
  2. Inability to update rules in real-time
    The first limitation helps explain the second limitation: rules-based personalization is determined and configured by a human. In other words, someone literally needs to set the personalization rules. It’s theoretically possible to do this on a daily basis but not on a one-to-one basis for more than a handful of customers. In fact, few brands ever change their personalization rules because it’s an onerous, endless task.

Harness the advantages of ML-based personalization

Fortunately, brands can now call upon machine learning (ML) and Artificial Intelligence (AI) to achieve newfound levels of real-time personalization at scale. In contrast to a rules-based engine, AI-based personalization can look across channels to arrive at a 360-degree view of the customer, along with the customer’s intent at that moment. By taking into consideration all critical variables – including customer behaviors and business factors – this technology assesses the relevancy, correlation, and likely causation of each in real-time. It weights these and then automatically predicts the best course of action to positively impact the customer’s experience.

For example: Say Susan is on a retailer’s website and looked at reviews for Tommy Hilfiger’s IONA boots. She goes back to the product page, selects a color of the boots she likes and selects the size, but does not make a purchase. A rules-based engine would later engage Susan with recommendations for similar boots. An AI-based personalization engine would take into account the availability of those particular boots in the color and size that Susan had browsed and interact with her about them.  In this way, the AI-based engine enables personalization in real-time.

Potential ROI of AI-based personalization

Customers expect and welcome in-the-moment personalization, such as highly relevant emails or valuable web interactions. However, they don’t immediately act in response to every instance of it. That’s why personalization leaders measure the impact of data-driven personalized omnichannel customer experiences using longer-term metrics. Per Witcher, these include:  

  • Lower attrition
  • Greater purchase frequency over time 
  • More frequent visits between purchases 
  • Longer online engagement 
  • Higher email open rates 
  • More app downloads 
  • Greater basket sizes (often driven by more items per transaction) 
  • More uses of different purchasing channels 
  • Fewer customer service calls 
The impact of AI-based personalization on conversions

Gathering customer feedback via surveys enables retailers to also measure for higher  Net Promoter Score (NPS) and customer satisfaction (CSAT) scores. 

Beyond metrics, retailers can justify delivering this type of hyper-personalized experience because customers remember who enables easy, convenient interactions. When they are ready to buy, they often turn to that retailer.

Enable AI-based personalization with ZineOne

Retailers ready to capitalize on the power of AI-based personalization can call upon the ZineOne Intelligent Customer Engagement platform. It combines and processes an unprecedented amount of past and present customer data points, and then leverages AI to react to customer triggers in real-time and on any channel. As a result, retailers can engage each customer at the right time, in the right place, with the right message.