Syte Takes On Site Search: Announcing Enhanced Textual Search With Deep Tagging
By Idan Pinto • July 30, 2018
Syte is pleased to announce that our deep tagging tech can now be used to improve text search! A complement to our visual search feature, Syte’s AI will upgrade your text search to make it better than ever.
Teaching an old dog new tricks
The same technology that powers our visual search is now being applied to text search — it can enhance existing lackluster text searches or replace them altogether.
How it works is deceptively simple. Images are processed by our AI, which recognizes different elements in the image. These elements are then broken down into their attributes and tagged. This is a gold mine for search in e-commerce.
Take this photo of a model, like one you would find on any ecommerce clothing site:
When you run it through Syte’s deep-tagging technology, each element of her outfit gets recognized and is given a list of attributes, including color and material, style features (like the v neckline) and even jewelry shape:
These categories offer a superior granularity and consistency for breaking down an item with text search. Every silk item will get tagged as such, and with the identical language. Same for every green item, every curved-arm sofa or cold shoulder top. And the tags above are just a few of the tags that Syte can provide you — every image can have as many tags added as you need.
That means that categorization is more consistent and less reliant on the whims of a human tagger. Every item that is run through Syte’s deep-tagging tech can have any number of tags automatically added to it, so it’s customizable for a retailer’s needs.
For marketplace companies, who often rely on product images without full or consistent descriptions, deep tagging based on images streamlines product categorization.
For e-commerce retailers only dealing with their own merchandise, it can cut down the time it takes to make a product list and reduce human error when entering product data.
For retailers who deal in resale, it can take user error out of item descriptions for superior inventory and a smoother seller experience.
And for all retailers, it speeds the process of adding tags and descriptions to items manually.
Better UX in a flash
This deep tagging tech, when applied to text search, results directly in better UX for customers. Big frustrations with text search are:
Lack of detailed product tags leads to dead searches or poor results
Take this query for a shirt dress on J. Crew. Even with the filter “women,” the results were still dress shirts:
When the product we wanted was hiding in the inventory, nowhere to be found in the search:
When a shopper is struggling to find an item even though they feel like they’re describing or filtering their search well enough to bring up a match, they will bounce away from your site. Now look at how much more accurately Syte’s search deals with text terms:
Each specific feature shows up immediately in the results of the search, so users immediately see what they came for. Say goodbye to losing a conversion for someone who doesn’t want to wade through an entire product inventory to find what they want. With deep tagging for text search, Syte can match your customer with their dream items faster than ever before.
See what deep tagging can do for your eCommerce site