Everything You Need to Know
Jess Erdman, June 2023
The second-hand clothing market is one of the fastest-growing industries in retail, with second-hand clothing estimated to take up 20% of shoppers’ closets by 2029. Fashion resale platforms appeal to both environmentally conscious shoppers, and those looking for budget-friendly options.
As more small and medium-sized retailers consider opening up resell marketplaces on their own websites to tap into this new market, they will need to consider how to create a strategy for tagging incoming products.
Because when shoppers explore clothing options on a resale platform, they’re often left confused by inaccurate product filters and search.
The reason: poor eCommerce tagging.
eCommerce tagging is the process of labeling attributes and metadata for products, which are then used to build product filters and search.
When powered by AI, eCommerce tagging can be done automatically.
Here are the key benefits of automating eCommerce tagging for SMB second-hand marketplaces:
1) Save time and money in tagging product attributes (AI can tag thousands of products in a few hours, which would take the same human one week).
2) They can use tagging for better search and discovery such as filtering and site search.
3) Product tags can be used to automatically generate product descriptions.
Whether the second-hand platform is a managed inventory app (like the RealReal) or peer to peer resell app (such as Depop), every second-hand platform is tasked with processing product images to extract relevant attributes for search filters.
Fashion resale platforms have a ton of new products uploaded every day, leaving them with a huge catalog of products that need to be tagged.
And, user-uploaded images can have varying image quality, as well as busy backgrounds. That makes it difficult for image tagging software to accurately identify key product attributes.
Manually tagging the large volume of new products uploaded is also not the best solution, because manual tagging is time-consuming and humans are prone to making errors, inconsistently tagging products.
Without an accurate, fast eCommerce tagging strategy, sustainable fashion platforms can’t provide a seamless eCommerce experience for shoppers. Since tagging is at the root of providing a good eCommerce experience, platforms can’t provide filters, site search, or product recommendations that meet shoppers’ needs.
Ultimately, poor eCommerce tagging leads to a poor product search and discovery experience for shoppers.
Across fashion resale platforms, we found varying levels of accuracy for search because when photos are cut-off or low quality, eCommerce tagging can be challenging.
Depop is an example of a second-hand marketplace that has thousands of new user-uploaded products every day.
Depop is using product tags to create filters and similar recommendations, making it easier for shoppers to discover their vast product catalog.
As a best practice, secondhand marketplaces should include various filtering options like color, size, and brand.
But, what if we took it a step further?
With deep image tagging, retailers can get even more detailed product information from images, such as sleeve length, neckline, skirt length, pattern, and more–all attributes that shoppers value.
Let’s take another example of a second-hand marketplace–ThredUp. With over 50,000 women’s tops, without product filters and recommendations, it would be nearly impossible for shoppers to find what they want. And, filters and recommendations are powered by accurate metadata (aka eCommerce tagging).
Take a look at the product description for a pink sweater:
For sellers, it can be challenging to create a product description that will help their product stand out from others, and help their product appear in search results.
But, with YesPlz eCommerce tagging, combined with ChatGPT, sellers can create richer product descriptions that convert shoppers. For retailers, rich product descriptions could also be used for site search, which matches product titles and product descriptions as part of semantic search.
YesPlz built AI-powered eCommerce tagging that’s trained to identify product attributes, specifically for fashion, even in the most difficult lighting and low-quality images.
We can recognize category, silhouette, and design attributes despite the type of image, making eCommerce tagging more efficient for sustainable platforms and helping shoppers find what they want.
What makes our machine-learning model special?
We work with a small data set that is easily and quickly trained for each specific retailer. Our machine learning model can recognize unique types of product images, filtering out noise and filtering in the key product attributes.
Any second-hand marketplaces with a little bit of help with product tagging can easily replicate the success of Depop and Thredup with well-organized search and product discovery. YesPlz image tagging can help any size second-hand marketplace access fast and accurate product tagging that would normally take a week if done manually.
And, with accurate eCommerce product tags, any size of second-hand marketplaces can also tap into our powerful product discovery tools like product recommendations and smart filters, creating a seamless discovery experience.
In this article, we’ll explore an alternative to traditional filters: smart product filter & search. Retailers lack the resources to build sophisticated smart product filters from scratch, leaving shoppers with hard-to-understand and difficult-to-navigate filters. But, there's a better way.
AI is making waves in the fashion industry, with influencers testing out AI tools to see how they hold up to their fashion preferences. Can AI styling tools (or just ChatGPT) meet their high expectations?
TikTok Shop just went live. What does this mean for eCommerce businesses keeping up with the news? Are we entering a new era of product discovery? We’ll take a look at how TikTok Shop works, and compare it to other social shopping platforms to see how it works.