Scrolling Re-Sale Sites Is About To Get More Fun - Thanks AI!
Vintage and pre-loved is hot right now, see the Cannes Red Carpet if you need proof. But we all know that searching and scrolling through vast marketplaces, with bajillions of SKUs takes patience, frankly too much patience for many of us.
Never fear, AI is here.
AI is going to revolutionize the luxury resale marketplaces - which is going to be a lot of fun for consumers looking for the thrill of the hunt, but need to increase their dopamine hits from successful finds.
AI is about processing the chaos, bringing order and structure to data that lacks it. What is more chaotic than scrolling eBay or The Real Real late night, searching for that one perfect item. Marketplaces are fun and hard because of the sheer number of unique items. And peer to peer marketplaces are even more of a wild west, with imagery standards that are all over the place.
But AI, in its various forms, has the power to revolutionize resale marketplaces (and frankly any marketplace where the sheer volume and diversity of SKUs for sale makes browsing potentially overwhelming). Most of these ideas are about serving you things you are more likely to like. I spent several years of my career at Taboola, a content recommendation platform popular on millions of websites. And what I learned there is that people love to click on that shiny next story/image/thing-to-buy. The trick is in serving them the right one. And so much of AIs advances, as seen below, will help with that.
Some ways resale-marketplaces will get more exciting for shoppers thanks to AI:
Making It Easier On Influencers- influencers are key to finding great things on marketplaces. How else would one wade through the thousand upon thousands of items. Image recognition and categorization improvements that come from deep learning tools will allow influencers to extend their scope and depth of recommendations, in a way that uniquely fits their style and still feels curated, without spending hours themselves trying to find the best beige ruffled blouses in silk on a resale platform.
Supercharging Search - again with image categorization and tagging, I’d expect to see search results improve markedly in the next 18 months, with consumer being able to use more complex phrases or “image-to-search” technology on all platforms, yielding search results closer to what they see “in their mind’s eye.” Poshmark recently noted the success they’re seeing with Image Search, proof in the pudding as they say.
Start-ups like Capsule will tighten the link between social media spots and searching for that item, allowing you to take what you see and quickly end up with search results without having to try to describe the item in question.
Better Categorization. While there’s a limit to the number of filters any UX can handle, certainly some additional orders of filtration could help customers, aided and abetted by the better categorization it will bring. E.g. if I’m in the mood for evening dress browsing, show me all the ones that look like something Zendaya would wear (a girl can dream?) or that don’t show too much cleavage, but also not too little. Getting more granular will give the customer more opportunity to search within their needs.
Next level categorization will also lead to curated feeds, much like people have discussed in relation to Twitter/X, where you subscribe to a feed created by a personality, and the algorithm helps to create that feed of clothing, of all stripes/shapes/sizes/seasons that align with the style of, say, the Olsen twins or #OldCeline.
Offsite Ads. (the ads run by say, Fashionphile on ad platforms, featuring a specific product or products) One of the delights of Instagram is how targeted the ads can feel, when my feed is dresses and shoes and swimsuits I might actually buy. But I LOATHE clicking on an ad only to see I’ve been lured by a hero product that’s not in stock or hasn’t been released yet. This issue is multiplied for re-commerce, where there may only be one of any item. Real time product insertion based on whatever degree of personalization is possible, but certainly based on available stock and basic demo/geo/etc data will surface the items most likely to convert to potential buyers, without hours of manual work to keep stock updated.
Better images. Many images on resale sites suck. Being able to manipulate the image to look more like a catalog while retaining the necessary veracity will increase the “curb” appeal of items. In addition, being able to scan the internet for original, high quality imagery, e.g. runway or ad-campaign imagery, will better highlight pieces and underscore the brand story behind them. @Vogue Runway I hope you’re thinking about how to monetize your archives for this use case.
Communicating with sellers. Sellers can often feel in the dark, hoping to get more money for their beloved goods. AI guidance and customer service will hold their hands at scale. Telling sellers why this image is better, this is why this sold so if you have more like that, do this, etc. will lead to higher NPS scores and ideally more quality inventory. In most industries, the best short term AI use cases require human involvement, a conversation if you will, and re-sale will be no different.
There are some examples I’ve left out that will significantly impact the business, but will be less evident to everyday users, like pricing automation, authentication advances (this was mentioned in several recent earnings calls) or that will affect so many business, like Co-pilot for developers or AI in customer service chatbots. Still a wild ride!
I recently listened to Carol Hilsum on the Future of Shopping podcast talk about how e-commerce was originally always trying to capture the fun of in-person shopping. Have we entered an age where the fun of digital shopping, with AR try-on, immersive stores, and ultimate curation will lead to a shopping experience that is significantly better than IRL? Only time will tell, but I’m hopeful.