Request a Demo

The Future of eCommerce Discovery is Here: Announcing the World’s First Augmented Site Search

Company Updates
Product Announcements
Ofer FrymanOfer Fryman | October 6, 2020

Over the last five years, eCommerce has undergone a tremendous evolution shaped by changing consumer expectations. However, despite the far-reaching progress, there has been a single lagging constant: on-site search.

Until now, eCommerce site search has uniformly relied on the faulty foundation of inconsistent existing product tags, creating a database that requires ongoing upkeep and, more often than not, produces limited, irrelevant results.

Today, I’m thrilled to announce Syte’s new Augmented Site Search. We engineered it with the express goal of overcoming one of the most significant barriers to purchasing online: inadequate and outdated text search architecture.

Upending traditional infrastructure, it is the only eCommerce search solution that uses visual AI to automatically enhance product tags with the most robust lexicon of fashion and home decor attributes in the world. Combined with NLP (natural language processing) layers that decipher human context and intent, augmented search unlocks true accuracy, providing hyper-relevant results instantly.

In short, it’s eCommerce search that truly understands the way your customers think: Their unique syntax, their intent, the look and feel of their style. For the first time ever, shoppers can search intuitively and get only the results they’re looking for. 

Augmented site search is Syte's advanced eCommerce search solution

Augmented search transcends traditional feed-based search to provide a flexible solution that can be customized to your brand’s specific business needs. 

It is the culmination of four years of research and development, combining two natural language processing layers with our advanced visual AI engine to pioneer an end-to-end intelligent text search solution for eCommerce. 

By enabling simplified navigation, augmented search clears the path to purchase for high-intent shoppers, connecting them with the products they want, dramatically improving customer experience, and increasing search-generated revenue in kind. Conversion rates in sessions with augmented site search are up to 48% higher than average.

eCommerce Search and the New Era of Customer Experience 

As commerce continues to evolve, consumer behavior has shifted in such a way that the linear customer journey has been rendered obsolete. Shoppers now interact with brands and retailers through dozens of touchpoints near-simultaneously. 

And at every point of contact, they expect — or demand — to be recognized as unique individuals and to be offered the experiences that align with their specific needs.

For every shopper that uploads a snapshot to find the product that captured their imagination via visual search, there are others who prefer to type out or voice what they’re looking for. 

For every text search completed, there’s a shopper adjusting their keywords in an effort to further narrow down results and find the cut they already know flatters them.  

Until now, eCommerce site search has fallen short — both for shoppers and brands. 

Why Traditional Site Search Doesn’t Measure Up 

Traditional eCommerce search engines work with a limited, basic set of tools that are not vertically optimized. They struggle to make sense of queries using weighting functions that fail to capture intent, disappointing shoppers and leaving revenue on the table time and time again. 

It’s important to keep in mind that this is not just about understanding human language and intent, which should be the bare minimum: For example, in an attempt to find formal shoes for a wedding, let’s say a shopper types “red dressy high heels” in the search bar on their favorite retailer’s website. While any person can identify the term “dressy” as the descriptor of the occasion, most search engines will conflate this term with the product category of “dresses.” Search engines are not built to truly understand the priority of words within a query nor the relationship between them. 

It’s also about simple queries such as “maxi dress v-neck red.” A query like this will produce search results that favor maxi dresses, even if they’re not red, whereas changing the order to “red dress v-neck maxi” will, in most cases, leave shoppers with red dresses of varying lengths. 

Syte's eCommerce search solution

What’s more, traditional search engines rely on matching queries with inflexible and incomplete product descriptions, so synonyms and inconsistencies (Party dress or evening dress? Blue couch or navy sofa?) mean relevant products –– sometimes all of them — get left out of the results.

This obviously translates to a sub-par customer experience in an era when nothing matters more. It also stems the flow of search-generated revenue by de facto shutting down a key avenue for high-intent shoppers. 

eCommerce site search

How Augmented Search Works 

Syte’s Augmented Search is built on a visual-AI-powered deep tagging layer. The AI engine automatically augments product information, assigning detailed, image-based tags to your inventory from our vertical-specific lexicon of more than 15,000 fashion and home decor attributes and their synonyms.

Since search engines are only as powerful as the metadata that feeds them, this added layer of product details ensures that relevant items are displayed in results even when customers use search terms that don’t align with your catalogue descriptions. For example, a shopper searching for “balloon-sleeve dress” will see all the dresses in that cut, even if they were initially tagged with generic terms, like “v-neck dress” or “black minidress” because the visual AI engine will recognize and tag the sleeve style automatically. Combining this augmented metadata with the NLP layers that translate language and context unlocks a uniquely powerful and accurate search experience.

The image-based tags also create automatic filters and navigation on the search results page, allowing shoppers to narrow down their search by price, style, length, color, occasion, and more. 

And, in addition to being served the results that most closely match their query with options for continued navigation, shoppers see items personalized for them based on behavioral data and/or in accordance with smart merchandising rules you set. 

Syte's eCommerce search solution

As one Syte customer explains it:

 “With other search providers, you have to send an index for suppliers, categories, and products. If I want a brand to be displayed, I need to type, for example ‘Levi’s sweater.’

“What’s powerful about Syte is that we can analyze all the sweaters and display all the brands related to that product, regardless of whether the brand name was entered as a search term — it’s great and a lot more efficient. The way synonyms are naturally detected is also much better than what is currently done.”

Despite the complexity of the technology that powers augmented search, we’ve designed it to fit into your eCommerce stack as a “plug and play” solution, so you can immediately reap the the key benefits it provides: 

  1. Smarter merchandising – By automatically augmenting your product feed with deep tags that go beyond your product descriptions, you not only save time on merchandising, but also capture critical details that ensure shoppers can use a range of terms to find the exact items they want. 
  2. Simplified navigation – Shoppers enjoy a better customer experience, where they can search intuitively, use icon-based navigation, and instantly see accurate results, despite variations in terminology and spelling.
  3. Improved UX –  The AI-generated product tags are automatically converted into filters for color, size, cut, and more on the search results page, enabling shoppers to further narrow down items and continue their discovery journey. 
  4. More search-generated revenue – Personalized results according to shopper behavior and history drive conversion while surfacing slow-moving inventory in addition to popular styles. Retailers can also add smart merchandising rules in accordance with promotions or inventory limitations.
  5. Access to critical data – Using data from search queries, you can uncover insights to inform merchandising and inventory decisions using intel directly from your customers. 

Keeping Pace With the Evolution of Customer Experience 

By offering Augmented Site Search, we aim to arm eCommerce brands with the advantage of moving as quickly as consumer behavior changes. Augmented search was developed to fit in seamlessly with the way shoppers live their day-to-day lives, so you can provide the intuitive and personalized discovery experiences they’re looking for. Understanding the context of human requests also lays the groundwork for conversational search in all languages, which is next on the roadmap. 

As eCommerce approaches a new frontier in light of the accelerated growth this year has brought with it, we are committed to providing you with the technology that will redefine your customer experience and win over the hearts of your shoppers. 


Ofer Fryman

Ofer Fryman

Ofer is the CEO and one of the co-founders of Syte. He brings in 22 years of expertise in machine learning and deep learning.

Latest posts

The A-Z of eCommerce Product Discovery
Sarah Hillel | October 26, 2020
READ MORE
Ofer Fryman | October 21, 2020
READ MORE
How Recommendation Engines Build Customer Relationships
Sarah Hillel | October 19, 2020
READ MORE
Request a Demo

© Syte 2020 All Rights Reserved.