EnglishBack to Whitepapers
The Anatomy of Personalized Search
If you’re looking to unlock magical customer experiences that foster customer loyalty
online, the key lies with personalization and artificial intelligence (AI).
By Carl Bleich
05/05/2022
13 min read
Introduction
If you’re looking to unlock magical customer experiences that foster customer loyalty online, the key lies with personalization and artificial intelligence (AI).
But first, think back to the last in-store shopping experience you had that surprised or delighted you. Even if it’s been a while, we’ve all had at least one stellar shopping experience, and it most likely had to do with that extra bit of attention to detail or personalization.
The challenge, though, is how marketers and merchandisers can translate those same meaningful, in-person experiences to an online setting. When site search returns zero results or irrelevant products, that can lead to customer frustration and drop-off, especially with so much competition online in today’s market.
That’s where AI in marketing comes in. By leveraging customer data and machine learning, you can anticipate your customers’ next move and improve the customer experience.
Read on to learn how to use AI-driven tools and take advantage of personalized search features in order to enable your online shoppers to feel understood, delighted, and eager to return to your site.
DEFINITION
Semantic understanding is a product data search technique that aims to not only find keywords, but to determine the intent and meaning of the words a person is using for search.
EXPLANATION
A customer searching for “chocolate milk” is looking for something very different than one searching for “milk chocolate” — can your search bar tell the difference? To avoid customer frustration, your search bar must understand the subtle differences in search terms, find similar results, and make shoppers feel understood. A search bar with semantic understanding helps brands understand what customers are looking for and surface the most relevant search results — at scale.
DEFINITION
1:1 and segment-based personalization allows companies to rank the products shown on search results pages in a way that’s unique for each individual visitor.
EXPLANATION
This feature makes it easy to personalize search results based on a customer’s affinity for products having a particular attribute (such as gender, brand, or category). For example, a search request for a “winter coat” seems simple enough to service. But would you wear the same winter coat in California that you would in Vancouver? 1:1 and segment-based personalization adjusts how products appear and how products rank for each visitor, so they no longer need to sift through cluttered, irrelevant search results to find the product they had in mind. Understanding user preferences on an almost individual basis — thanks to the combined power of customer data and AI — allows companies to reorder products based on segments to offer an optimally personalized experience. This ensures your company isn’t returning search results for a light windbreaker to Canadians in January.
DEFINITION
Customize your SKU selection to make sure that the right product variant is identified and returned on search and filter queries
EXPLANATION
If your search results aren’t personalizing the thumbnail of the product image, that’s a missed opportunity. If a customer searches for “medium red yoga pants” on your site, they’re expecting to see an image that reflects the exact attributes of the product — in this case, that would be a picture of a model wearing, you guessed it, medium red yoga pants. If your search result shows a generic pair of black yoga pants, even though red is an available color variation, then you risk losing your customer’s interest immediately. By adjusting the preview image based on product attributes, you ensure that the product search experience is as personalized as possible.
Enabling Truly Personalized Search
To achieve truly personalized search, you need to be able to collect commerce data (the combination of collected customer data and product data), use that data to effectively show customers what they want, and properly balance the “in the moment” intentions of customers with previously collected data.
But your algorithm can’t personalize the same way every time. It must be able to pivot from personalizing based on a shopper’s perceived preferences to personalizing based on their in-session intentions. A customer may have been shopping for herself yesterday, and has come back today to find a gift for her father — your algorithm needs to recognize that and adjust in real time. Intelligent product search understands the difference.
Put Your Own Website to The Test!
What happens when your customers search for:
Organic bananas
Or organic anything, for that matter. It’s no secret that online grocery shoppers can be quite picky about their needs and wants. Is your search up to the task? Can your search bar take socio-economic factors into account and consider those factors when displaying products?
Dress shirt
What do you notice? More importantly, are you getting results for “dress shirt” rather than “shirt dress”? What type of dress shirt (men’s, women’s, long sleeve, short sleeve, etc.) results are surfacing? Are those results personalized to the shopper?
Gloves
How many varieties of gloves do you sell? Would different customer types expect to see different results for “gloves” (e.g., latex gloves for a hospital, food service gloves for a restaurant, electrical protection gloves for electricians) at the top of their results? Does your experience treat all customers the same?
Coastal photo frames
How about here? Does your search bar recognize the very specific nature of this search intent? Does it recognize that with the use of the term “photo,” the visitor is not looking for a coastal photo but rather a frame with a coastal design? When search results populate, are frames with beachy or coastal features the first to appear?
Screw machine drill bit
Do the results understand that “screw machine” drill bits are the same as “stub” or “short length” bits? Are results for “machine screws” appearing in results? Are there results for “screw machine” heavy equipment or parts/accessories?
Red joggers
The clothing color test. Can your search bar pass? Will it actually populate the correct color that the customer is looking for? Will it recognize the term “joggers” to mean active pants? Can your search bar pair with previous customer data and prebuilt segments to understand the intentions of the seeker and accurately populate results for “joggers for women”?
Ready for Some Site Search Inspo?
Check out our blog to see which brands are offering customers an incredible search experience and how.
