How does it work?


1

User Generated

The user can provide feedback on an item through the StyleBot or through the online review page. The system will extract the keywords and update them as crowdsourced tags on the product detail page.


See more

 


2

See rankings

The product detail page shows a ranking of popular tags that customers use to describe the item after they purchased and received it, so that new customers can understand the fit and feel of the item before purchasing it.


See more

 


3

Easy Search

Crowdsourced description tags also make searching items easier and more intuitive.



Shoppers have a hard time understanding the descriptions and categories of clothing items online. We found that people use different words to describe items after seeing and feeling them in person than after looking at a picture. Inspired by customer reviews and tag systems, we explored crowdsourced tags to help people better understand what to expect, and to search for what they want in their own natural language.


If you were to implement it...

Get more users

Shoppers trust other shoppers to give descriptions of clothes they’ve already bought, felt, and tried on.

Rank by popularity

Most popular tags help shoppers predict online what clothes will fit and feel like when they get them in person.

Extract tags from reviews

Tags are easier for shoppers to skim than reading through full user reviews and convey different information than the online descriptions.

Let users search by tag

Shoppers can ask for what they want in natural language and the site can match items by tag.


Alternatives

2016-2017

Users manually generate tags when they leave a review. The system automatically extracts relevant words from reviews. The platform facilitates a product search by tags.

2017-2019

Allow voice dictated tag search.

2019-2022

Respond to a more complex request, e.g. “Pick the closest AEO out t that looks like Beyonce’s out t at last night’s concert.”