Transformative Fashion AI: W Concept x YesPlz

W Concept's Digital Transformation

YesPlz.AI, October 2022

Fashion-forward, contemporary, and fashion powerhouse are just a few words that come to mind when describing W Concept, a leading online retailer for luxury fashion, with a worldwide presence across the U.S, Korea, China, and the U.K. 

 

With innovative products for women and men across categories, including clothing, shoes, handbags, and beauty products, W Concept has positioned itself as a leader in online retail, and is loved by fashionistas and influencers worldwide.

 

W Concept came to YesPlz with a unique challenge: as a growing online retailer for luxury fashion, how could they curate a robust and resourceful search experience, while still maintaining their innovative and fashion-forward image–all powered by fashion AI?

The Style Filter showing open, cropped, party vibe filter attributes

With fast growth, comes fast changes

 

As W Concept continues to grow, a number of challenges began to appear. YesPlz worked with W Concept to creatively solve each of these challenges.

 

Challenge 1: W Concept carries a large volume of products that need an innovative fashion AI solution.

 

With thousands of products available, W Concept’s text-based search could only go so far in enabling visual discovery. Shoppers weren’t receiving optimized search results based on their keyword inputs, resulting in missed opportunities for discovering new products. 

 

And, the sheer size of W Concept’s product volume is overwhelming: even the best text-based search system couldn’t keep up with the constant flow of new products. 

 

Challenge 2: Limited resources: across the team, members are stretched out to accommodate W Concept’s constantly increasing demands across the company. 

 

Innovative, intuitive visual discovery sounds nice in theory: but the day-to-day chaos of working at a fast-paced eCommerce can get in the way of planning and implementation. 

 

Without fashion AI tagging, W Concept struggled to manually keep up with incoming products, leading to overstretched resources.

 

Challenge 3: Lack of solution providers who can create customized, sophisticated solutions for W Concept’s unique needs.

 

Too slow. Not customized enough. Not easily integrated. Not accurate. Lack of sophistication. 

 

Those are just some of the problems that appeared for W Concept once they started the search for a visual discovery solution. Even for some search solutions using fashion AI, the technology simply wasn’t good enough for a top-tier retailer that needed a high-level of search accuracy and seamless integration into their existing systems.

 

Poorly trained fashion AI still requires extra work on behalf of the retailer like fixing mistagged products. And for the customer, it means they’re going to see poor quality search results.

 

The YesPlz approach for W Concept: agile fashion AI technology with the power to delight shoppers

 

YesPlz approached the challenge head-on, while taking a step back to analyze the overall user journey to generate a relevant, intuitive discovery experience for shoppers at every point. By thinking thoroughly about W Concept’s shoppers, website, and challenges, we curated the following solutions:

 

Intuitive, accurate fashion visual search: The Style Filter

A cell phone showing the Style filter with text about filtering with virtual mannequin using fashion AI

Two product filters on two cell phones, before and after fashion AI

The Style Filter, aka virtual mannequin search, lets shoppers quickly filter based on silhouette and preferred product attributes from pattern-type to jean wash to fit.

 

Through a mix of visual search and fashion AI, shoppers receive accurate, instant search results. 

 

Time is saved by behind-the-scenes fashion tagging, allowing for fast, accurate search results. 

 

The Style Filter is a new way to search that shoppers want to engage with–unlike other fashion visual search solutions, like camera search.

 

Complete retailer control over features:

And, W Concept remains in control of the features they want to publish and unpublish–leaving retailers with the power to craft their search experience as they want. 

 

Occasion filters crafted from real shopper feedback



Two filter options for finding festival vibes using fashion tagging for occasion

Thematic filters (widely known as occasion filters) are a best-practice recommendation for eCommerce, but are complex for many retailers to implement.

 

With a deep product catalog, W Concept faced a complex challenge: how could they offer occasion (aka thematic and vibe) filters that were accurate, specifically tailored to shoppers’ tastes? And could those occasion filters be scaled across the large volume of products?

 

Powerful fashion AI tagging + real shopper feedback = the true occasion filters shoppers want from W Concept

A step above the rest, YesPlz occasion filters use fashion image tagging to identify product attributes, laying the groundwork for creating occasion filters.

 

We also created a unique fashion survey to get real feedback from W Concept shoppers that trains the AI, creating a positive feedback loop between W Concept shopper needs and occasion filters 

 

The impact of YesPlz fashion AI on W Concept

 

Shoppers can take a breath of relief, as they’re now in control of their own shopping experience. 

 

No more getting lost in thousands of products. No more endless scrolling to try to find their preferred product attributes. No more feeling overwhelmed. 

 

Stay tuned for more details on the results, and check back as we update them.

 

Similar recommendations with a unique double filtering system


Different tops at W Concept using fashion AI product recommendations

Similar recommendations have the power to delight shoppers and leave a lasting positive impression.

 

How do YesPlz fashion AI recommendations work?

 

There are two steps we take to create accurate, robust similar recommendations:
1) Match the key design attributes to products
2) Match the recommendation, again, against vibe and occasion

 

The end result: AI-powered product recommendations that understand shoppers’ tastes. 

 

By using a unique double filtering system, YesPlz fashion AI recommendations go above and beyond other similar recommendations when it comes to accuracy and style.

 

Identifying the “why” behind shopper behavior

 

Similar recommendations in fashion eCommerce can fall short when they don’t consider the why behind the recommendation. Based on YesPlz research, shoppers value silhouette as a key product attribute. But, that’s not enough to create a powerful product recommendation.

 

The next layer of recommendation accuracy: identifying search intent and vibe. 

 

W Concept’s highly sophisticated shoppers required multi-faceted recommendations to continually meet their needs.

 

Collaborative filtering + similar recommendations = the power to solve common recommendation problems

 

In addition to similar recommendations, YesPlz built collaborative recommendations for W Concept–and later combined it with similar recommendations to create a robust recommendation experience. 

 

After running different collaborative recommendation models, YesPlz customized the model to be a perfect fit for W Concept shoppers’ experience.

 

And, W Concept didn’t have to make any trade-offs when deciding which recommendation system suited their needs best.  Each product recommendation type has its own merits, and comes together to enhance the overall quality of products recommended.

 

YesPlz machine learning capacity includes all data types, including image, text, numbers–you name it.

 

More than 1 product recommendation type can have massive impact:

 

A common problem with collaborative filtering is the “cold start” issue: it is difficult to use collaborative filtering to make recommendations when either a user or product is new, leading to hidden products and missed sales. 

 

By using “similar recommendations” in conjunction with collaborative filtering, we can:

 

  • Use each of the product recommendation types to enhance each other’s qualities
  • Create a product recommendation system that is fast, accurate, and shows off the depth of W Concept’s product catalog while catering to W Concept shopper tastes
  • With powerful machine learning on their side, W Concept now has the capability to curate shopping experiences that are intuitive, accurate, and delightful.

 

Out-of-this world text search empowered by fashion tagging



Fashion AI text search before and after for W Concept

Finally: text search that’s custom built to be more accurate than other search tools, with fashion AI tagging at its core. 

 

The final piece in completing W Concept’s visual discovery transformation: building text search from scratch.

 

The magic formula?

  • Keyword matching
  • Similar search: we trained the data to understand what other styles resemble a search term, to make search more accurate and robust
  • YesPlz fashion visual search: No product data provided by third-party vendors? No problem. With YesPlz fashion visual search, we can still pull the relevant information from a product and generate product information with laser-accurate fashion AI. 

 

Let’s say a shopper is searching for a “camisole,” and types the term into W Concept’s search bar.

 

Behind the scenes, search results are generated based on data that understands a camisole can include a number of different products (and those products may not be named “camisole” in the description). For example, what about a bustier with cami straps? Or a sleeveless shirt? 

 

Shoppers rarely have the exact keywords in mind that match a retailer’s product description–but with fashion visual search, YesPlz can train fashion AI to recognize product attributes in images, leading to flawless search results, even when shoppers get the terminology wrong or products aren’t precisely tagged. 

 

W Concept’s unforgettable digital transformation using fashion AI


Fashion AI solutions for W Concept


Now, with YesPlz’s fashion AI suite, W Concept shoppers can:

 

  • Find the exact styles they want, including silhouette and cut
  • Fast-track their searches by jumping to their favorite vibe or occasion 
  • Receive curated similar product recommendations, along with collaborative filtering, leading to perfect product discovery
  • Take a fashion quiz to help train fashion AI, and have an impact on the filters they see
  • Easily search by text and receive rich search suggestions


Digital transformation isn't just digitization--it's empowering both retailers and shoppers to take control of the product discovery experience. For W Concept, digital transformation also means using YesPlz fashion AI to accurately tag product attributes, and set the foundation for a curated shopping journey.

Fashion AI as a digital transformation tool has the potential to transform eCommerce brands like W Concept–it’s fast, powerful, and easy to integrate. Whether your eCommerce is looking for intuitive product discovery, laser-accurate recommendations, or fashion tagging, YesPlz can create a customized solution for your business.

 

Schedule a free 20-min demo 

by YesPlz.AI

We build the next gen visual search & recommendation for online fashion retailers

Next Readings

Schedule a demo today