Applying the Netflix Customer Experience to Retail

When historians look to this century, there will be quite a few companies that stand out. Tesla is one, as are Uber and Airbnb. However, the three consumer platforms that will truly be seen as forward thinking and disrupting “business as usual” are Spotify, Netflix and Instagram. All three of these platforms have dramatically changed how we consume content and how we engage with the brands we love. The latter three companies are also ground-breaking trailblazers when it comes to their efficient browsing capabilities and consumer-friendly user interfaces powered by personalization. When Netflix matured beyond basic search by keywords, it adopted advanced personalization that uses data from what its customers liked or didn’t to find just the right shows based on what they (and others) watched previously.

It’s really not a stretch to see how retail brands can apply the personalization techniques of Netflix to tackle one of the key challenges they face: helping their customers easily find the content they want before they get frustrated and give up. Our own research on personalization, assured by PwC, found that personalization can deliver a 6 percent increase in revenue per visitor (RPV). A separate survey in the same report found that most consumers stay loyal to approximately five brands. The takeaway here is that unless your company can make a positive impact on customers the first time, they may not return.

Content ranking and evidence selection are two of the principles which effective personalization strategies are based upon. They help your brand create deeper engagement and customer loyalty by using customer context to make product search and shopping experiences far more relevant than what you would find on a site that's basically the digital version of a print catalog.

Content ranking and evidence selection also facilitate decision making and are effective tools that limit the cognitive load placed on a consumer wishing to make a choice. They are defined as follows:

• Content ranking is the order in which items, categories or images are listed on a page. We normally see products “content ranked” on category pages, within search results and in recommendations feeds. A marketer often has some control over this order by using filters or pre-set categories such as price, rating, location, relevance, etc.
• Evidence selection can be defined as endorsing a product with the most effective cue, message, tag or label to help shoppers choose a product with more confidence. Some familiar types of evidence selection include whether the item is new, a limited edition or award winner.

Simplifying Digital Decision Making

The inability to decide whether to buy or not to buy is a symptom of a world in which we’re faced with too many choices: latte or coffee, bagel or donut … you get the general idea. Even entertainment adds to the daily deluge of decision making. Looking at five similar e-commerce fashion brands from our customer portfolio over a 12-month period, we found 2,388 product categories with a total of more than 1.4 million products a consumer might choose from.

When it comes to structuring and organizing website content from an enormous list into something that's easily navigable and comprehensible for shoppers, merchandising teams have their work cut out for them. By automating or using artificial intelligence-powered content ranking and evidence selection techniques, merchandisers can more easily uncover the most effective message types and organize the products presented to shoppers, who in turn will have an easier time making a purchase decision.

Taking a Scientific Approach to Content Ranking

Content ranking does help manage an overabundance of choice, but all shoppers are inherently different, arriving with varied interests and via different channels. What’s relevant to each one depends on a host of factors, including whether the consumer has visited your site previously, if they’ve heard of your brand, and whether they’ve purchased from your brand before.

In fact, you may already be using evidence selection on your site right now, but how effectively? Are your product recommendations contextually aware? Context deals with a person’s need state at a specific time. Some of the signals observed by our AI model include the following:

• visit number;
• product categories they’ve visited;
• products they’ve viewed;
• products they’ve bought; and
• whether they’re on a mobile device or desktop computer.

Our analysis indicates that adding context to ranked content increases engagement and clickthrough by a further 6 percent to 10 percent. This has the additional benefit of helping inform which types of product and which types of context resonate with a customer. In turn, this creates a virtuous feedback loop that further enhances relevance and engagement, leading to an even more personalized experience.

Conclusion

One major benefit of personalization is being able to choose the most effective message for each type of visitor so that your site doesn’t become weighed down with too many “accessories.” Content ranking and evidence selection are two highly effective ways to streamline the experience. They can serve as essential tools to eliminate uncertainty in the decision-making process by helping consumers find the products they’re looking for, in less time and with a greater degree of confidence.

Simon Jaffery-Reed is the vice president of product at Qubit, a personalization software.