Pascal Finette

Logo

blog.
RSS FeedLinkedIn ProfileFacebook ProfileXING ProfileTwitter UpdatesSkype Account

July 17th, 2008

Visual Search - The Next Frontier?

I guess I need to get a bit more creative (and consistent) with my blog titles. Anyway - let’s talk about visual search. Which might well be the next frontier. As mentioned yesterday I firmly believe that we see a change in e-commerce - we will see more and more rich image and media, which will make us want that item we are looking at right now so much more. You’ll see a range of companies like RichFX jump on the opportunity and solve this need.

What you’ll also see is the advent of visual search in product search - and this will be a tremendously powerful shift. Check out Like.com and Pixsta to get an idea of what is coming our way: You shop by visual reference and not by painstakingly describing the item you are looking for - go to Like.com, select a category and narrow down your search by clicking into the images, telling the Like.com engine which features (e.g. a specific color or style) you are interested in. The whole process is so intuitive and fun - which makes you realize how little fun it is to go to a ‘normal’ retailer website, click on the category, then on the subcategory, then on a brand, and so on and so on.

It’s interesting to see how similar but also different Like.com (a US-based company) and Pixsta (a UK-based company) are: Like.com is the offspring of Riya - a technology company which originally developed facial recognition technology to identify and tag your digital images by ‘understanding’ who is in the picture. Like.com was more or less a technology showcase - originally Like.com showed pictures of VIPs (models, actresses, etc.), allowed you to click on certain features (e.g. a handbag) and searched for products which looked similar (which in itself is a great concept - you see that damn hot but also damn expensive Prada handbag which Kate Moss recently wore; run the picture of Kate wearing it through a visual search engine and you’ll find similar looking bags from all the more affordable high street retailers). Interestingly enough Like.com gave up this feature and is now a destination site - a purely visual shopping engine which facilitates products from a broad range of retailers. Pixsta follows a slightly different path - instead of putting their money on one card (becoming a destination site), they bet on a distributed approach by licensing their technology to media sites (e.g. ELLE or the German Gala magazine). Both companies set out to create a new, much better shopping experience - and to a wide degree they already provide a better experience than most ‘traditional’ shopping experiences.

I believe that we are only at the very beginning of this trend - combine a visual search engine with rich media images and the results will get so much better. Combine a visual search engine with the semantic web and the results will become so much more meaningful. Build this as a whitelabel tool which you can easily integrate into an existing infrastructure (e.g. your aging Intershop online store) and you’ve got a great business opportunity at hand. Design a new user interface and mashup visual search with the data from price comparison sites and you’ve got a killer application at hand: Where do I find what I am looking for (also if it looks similar to the item I wanted to buy) - for the best price? Imagine a site could tell you that you could get a handback which looks very, very similar to the one you originally intended to buy for $100 less. How powerful an argument is that?

news.

twitter.