2 Main Types of Personalization: Which is Right for Your Brand?

As they look to better connect with and serve their customers, today’s retail marketers find that personalization is a strategic imperative. According to Infosys, 59 percent of customers say that personalization influences their shopping decisions, and Forrester found that 77 percent of consumers have chosen, recommended or paid more for a brand that provides a personalized experience. Retail personalization can range from simple campaigns like keeping the message and experience consistent from email to the website, to more advanced techniques such as individualized site experiences based on a shopper’s unique tastes.

Two different types of personalization drive these experiences.

• Rule-based personalization allows marketers to deliver experiences to specific segments of people based on the manual creation of business rules.
• Machine-learning personalization employs algorithms and predictive analytics to dynamically present the most relevant experience for each and every visitor.

Both are necessary. Let’s consider each in turn.

Segment-Based Communications

Retailers use rule-based personalization to target experiences to both broad and narrow segments of shoppers. Broad segments refer to wide groups, such as all the individuals in a retailer’s loyalty program, while narrow segments refer to much smaller groups, such as all high-value visitors with an affinity for a certain product category that arrived to the site via a particular ad campaign.

Once you have different segments in mind, you can set up rules to target them with different experiences. These rules have to be set up manually, but they can be incredibly powerful. For example, online retailer Shoeline.com typically saw clickthroughs on its generic homepage banners of approximately 1 percent, but it was able to generate clickthrough rates of up to 26 percent with rule-based personalization, improving discovery of some of its niche shoe categories. As these categories were relevant only to certain shoppers, Shoeline created rules to display customized homepage experiences to shoppers based on the category they had engaged with the most over time. So if a visitor shopped nursing shoes on the site, that person could easily find the category again on the homepage. Anyone else who wasn’t interested in nursing shoes wouldn't see that experience.

Typically, rules don't allow you to create individualized experiences, although you can create very narrow segments that apply to a limited number of people. As a marketer, you must decide whether the effort to design an experience for a small group is worthwhile. In most cases, when trying to personalize at the individual level, marketers find a machine-learning approach to be more scalable.

Individualized Experiences

Machine-learning personalization employs algorithms and provides a more scalable way to achieve unique experiences for each individual shopper, rather than segments of people. While algorithms are most often used for product or content recommendations, they can also be applied to dynamically modify site navigation, search results, list sorting, promotions, category-level recommendations, brand recommendations and more.

Zumiez uses machine-learning algorithms to uniquely engage individual shoppers throughout its site — from homepage to checkout. Zumiez responds to each shopper’s preferences and intent to guide which brands they see, the products they discover and the content they're served. For example, the retailer deploys product recommendations in the “Complete the Look,” “More to Check Out” and “Recently Viewed” areas on its product detail pages, as well as directly in its search bar, taking each shopper’s preferred brands and styles into account. Shoppers who clicked recommendations generated by machine learning converted 2.7 times more often than those who did not.

Rules and Algorithms Together

Rules and machine-learning algorithms both have a key role to play within a digital marketing strategy. Rule-based personalization will continue to be important for communicating to a particular audience such as a persona, all visitors at a particular buying stage, customers coming from a particular source, and even customers with a shared affinity or intent. However, machine learning-driven personalization is playing an increasingly important role as companies look to scale their efforts and deliver truly one-to-one experiences to their customers. Both methods — used separately or as part of a strategy together — help retailers better connect with visitors, drive more conversions, increase average order size and improve loyalty.

Karl Wirth is the CEO and co-founder of Evergage, which provides real-time personalization to more than 2 billion web visitors and application users.