How eBay’s Innovative Image Search Technology Will Change eCommerce Forever

Holly Cardew
5 min readSep 1, 2017

--

Photo by Josh Rose on Unsplash

During Labor Day Weekend of 1995, a coder sat in the living room of his dimly lit home and did what coders do… he wrote code — for hours on end — launching “AuctionWeb” a site that brought buyers and sellers together in a single marketplace. Shortly after AuctionWeb was up and running, the coder listed his very first product, a broken laser pointer, which quickly sold to a buyer thousands of miles away in Canada.

Today, that coder, Pierre Omidyar, is worth $9.4 billion. And, while he no longer runs the company, the website — now called eBay- is the 9th largest internet company in the world.

There are a couple of lessons that can be gleaned from this story:

  1. Coding can pay off big time, and
  2. Don’t purchase a laser pointer on a site that was put together over a weekend.

Anyway, one thing is certain: 23-year-old, the pioneer in eCommerce still has a few tricks up its sleeve — tricks that could potentially change the eCommerce world it created. Below, you will find two of the latest technologies that eBay has introduced on its platform — Image Search and Find It — tools that may change the way buyers search the web for products.

Can’t describe what you are searching for? Take a picture.

As you probably already realize, artificial intelligence and machine learning are spreading like wildfire. Before we know it, our toilets, toasters, and thermostats will have minds of their own. I don’t know about you, but the last thing the world needs is an AI toilet. That could get messy.

I am being facetious, but in all seriousness, both artificial intelligence and machine learning are giving the world tools to do some pretty amazing things.

And, as the reigning leader of eCommerce, eBay is leading the charge with two tools they recently launched: Image Search and Find It.

Have you ever gone shopping and stumbled upon a piece of furniture you really liked? One that kind of looked like a couch or a lay-z-boy or a recliner or was it an audubon?

Well, if you weren’t interested in paying the crazy price you were sure to find in a physical store, you probably decided to look online.

But, here is the problem… Try Google searching: “Couch/ lay-z-boy recliner/ audubon furniture piece”. The chances of you finding what you are looking for are about as good as a snowball’s chances in a microwave.

eBay is harnessing the power of AI to help customers more easily search for the products they can’t describe but really want.

How would an eCommerce titan go about accomplishing this? A lot of manpower, capital, artificial intelligence and… smartphone cameras.

So, if you want to prep your eCommerce operation for this new algorithm what steps you should be taking? And what kind of data do you need to start collecting?

eBay’s Image Search uses a deep learning model. The model’s output creates a representation of the shopper’s image that can be compared to images of live listings. That means that the listing images have to have the same type of representation or a set of attributes that the AI can match the search image against, similar to metadata.

It seems you have to feed all your images to the AI so it can study them, make notes and create a library of reference points. It’s a game changer so you better keep up.

See the Future: By 2020, 6.1 Billion People Could be Searching by Image

By 2020, nearly 6.1 billion people in the world will own a smartphone. This is a really important number because it means over 6 billion people will have access to a mobile camera on a daily basis.

I imagine it was this realization that sparked eBay to build Image Search, a tool that will allow customers to take a picture of something they see and then use it to search for a similar product on eBay.

According to VentureBeat, Mohan Patt, eBay’s vice president of buyer experience explained,

Moments of shopping inspiration can come at any time, whether you’re walking down the street or browsing your social media feed. At eBay, we’re focused on creating new complementary technology that helps our millions of shoppers easily find the things they love at the best value.”

Search for an Image Within an Image.

eBay will also be launching Find It, which will allow customers to click items on any website or social media and then search for them via URL on eBay. Find It will also allow online shoppers to zoom in on things they like within a larger image and search for that particular item on eBay.

So, imagine you are looking at a picture of a living room on Pinterest, and you notice a window curtain you really like. With Find It, you will be able to zoom in on that curtain and search eBay for similar products for sale.

Source: eBay

Bridge the Gap: Connecting the Physical and Digital Worlds

Back in 1911, Arthur Brisbane, a renowned newspaper editor and real estate investor, was giving a speech to the Syracuse Advertising Men’s Club. It was in this discussion that he coined the phrase, “Use a picture. It’s worth a thousand words.”

It is a cliche that has been used countless times throughout the past 100 years by artists, creatives and photographers… but it has quite a bit of truth to it. Sometimes, the things we want, simply can’t be described, and when this is the case, we turn to a picture.

eBay has discovered a way to successfully bridge the gap between physical and digital world through pictures — allowing them to target customers both online and in person.

If customers embrace these tools, we will see a complete 180 turn in the way in which they go about making purchases.

For example, let’s say a teenage girl is perusing Forever 21 or Urban Outfitters and finds a sweater that she loves. But, it is a little too pricey. Instead of walking away, she will take a picture, do a quick search on eBay… and within seconds, be given dozens of the similar options. And, if she is lucky, the same sweater will pop up for half the price or with a better shipping option.

This will have a tremendous impact on physical and digital retailers alike, forcing them to either offer products that can’t be found online or better prices. Either way, it will benefit the customer.

I am keen to hear if you have implemented any type of image search into your eCommerce platform or if this is something you are thinking of doing. Leave a comment below to let me know! 😊

--

--

Holly Cardew

Building solutions for the next generation of shopping