Algolia helps Sellpy narrow searches with Advanced Personalization and blazingly fast results

The challenge

Founded in Stockholm in 2014, Sellpy is an e-commerce platform designed to make it easy for people to buy and sell second-hand items.

The company aims to encourage people to live more circularly, to minimize emissions and to improve sustainability. Customers wanting to give their clothes or fashion accessories a second life order a Sellpy bag, which they fill and send to the company, where the company describes the product, photographs it, and puts it up for sale on the platform.

Now available in 24 international markets through its website and app, Sellpy has sold more than 25 million pre-owned items to date, eliminating nearly 150,000 tons of CO₂. 

With every item being relatively unique, helping customers narrow down the 5.3 million items currently on sale to a manageable number so they can see the products they want — and maybe didn’t even know they did — is a significant challenge. 

“There are millions of items, and we’re not showing them to everyone. You can find corners of our inventory that few customers have really seen, products that haven’t received any favors, which is our way of showing interest. Meanwhile, you can find a similar item with maybe 10 favors,” says Alies van Rhijn, e-Commerce Platform Product Owner at Sellpy. “How to correct that so those items become surfaced is my main question daily.”

She says there’s a delicate balancing act between wanting the buyer to find what they are looking for — which they may not even know they are — and making sure the seller’s items are given equal treatment. 

“It’s definitely good to have that balance top of mind, and it’s very focused on the customer journey. We try to be as customer-centric as possible because, at the end of the day, that’s who we serve.”

The solution

An Algolia early adopter

Originally, Sellpy was only a seller, taking items and auctioning them on various Swedish auction sites. 

The company soon decided to eliminate the middleman and build its own marketplace but recognized that it would need a search engine solution that could handle massive inventory and make updates to items frequently. The solution also needed to be exceptionally fast. Sellpy items are unique, so once they are placed in a basket, they must no longer be available for purchase.

The company selected Algolia since it met their mission-critical search criteria and “have never looked back.” 

“We have millions of items listed, and it’s just no problem for Algolia,” van Rhijn says. “Being able to handle our scale at speed is our main requirement, and most tools — not just search tools — generally can’t handle that. Everything else that Algolia can do is a bonus.”

In the decade since first adopting Algolia for Search, Sellpy has implemented a range of Algolia features including Rules, Virtual Indexes, AI Synonyms, and AI Personalization. 

“Up until AI Personalization, we had been using Algolia primarily as a technology tool, with our focus being on its ability to update fast,” says Mathias Østergaard, Head of Data & Analytics at Sellpy. “Now, we’ve found that when you have so many items for sale, being able to tweak search results can make a big difference. We’re looking at how we can optimize things, and Personalization is part of that.”

The results

Growing together with great results

Sellpy has been using Algolia since the launch of its marketplace, and implementing new Algolia features into its homegrown e-commerce platform has been quick and straightforward.

“We’ve always been happy with Algolia’s ease of use,” Østergaard says. “It’s always been a very good experience for the developers to work with.”

“In the past two years, as we’ve started to use Algolia more on the product side of the table and less as just a technology solution, it’s been nice that you can do so much straight from the dashboard,” van Rhijn adds. “But even while you can do everything from the dashboard, we can also reach everything through an API, so I can make small adjustments on my side without any problem. It’s also very easy for an engineer to step in and do things across the board.”

As Spellpy’s inventory has increased over the decade to more than 5 million listings, the need for a search solution that can handle its unique needs has become even more crucial and vendors that can supply it even harder to find, Østergaard suggests. Here, Algolia’s AI-powered capabilities have proven an asset, as has the company’s continuous innovation and improvement.

“It requires an intelligent infrastructure to be able to keep up with our growth, and we have that with Algolia.”

AI Personalization has been a significant boon for the company, helping to narrow down, for example, a customer search for a “black dress” to incorporate attributes unique to the customer such as size. Previous attempts at implementing personalization required too much configuration and had hard-to-set attributes, Østergaard says. 

“That’s what makes AI Personalization a game changer: we just feed it all the data we have, and it figures out what to do with each user.”

Since recently implementing AI Personalization, Østergaard says the company has seen more items added to carts — signifying that customers can find the fashions they seek — but also, importantly, that views of items are being spread more evenly to the benefit of sellers. 

After implementing the feature, Sellpy has seen a 2 percent conversion rate increase, plus a 6 percent jump in average conversions per user. But ultimately, it comes back to meeting the company’s customer-centric goals.

“So instead of always showing the same black dress in size medium, we can now show ones that address each individual customer, which means ones in their sizes or with different patterns or dots also get shown,” Østergaard says.

“The customers have this selection that’s tailored to them and hopefully they feel more understood.”