Google Bets on Startup’s ‘Photo-Based AdSense’

pixazza

Pixazza’s service in action
Source: Pixazza. Click to enlarge.

Google is investing in a new startup, Pixazza, that bills its core offering as the photo version of Google’s text-based advertising product for Web site publishers, AdSense.

Pixazza, whose $5.75 million financing round also included investments from August Capital and CMEA Capital, today formally launched its service, which overlays photos on Web pages with links that enable viewers to buy the products shown.


Like Google’s (NASDAQ: GOOG) AdSense, which bases the text ads it displays on the content of a Web publisher’s site, Pixazza’s tool turns the content of Web images into clickable e-commerce links.

“Pixazza hopes to do for images what Google’s AdSense did for Web pages,” said Pixazza CEO Bob Lisbonne.

Given the similarity to AdSense, there’s speculation that Google might employ the technology somehow at its site, or even eventually buy out Pixazza. But for now, Google’s remaining mum on any plans.


“Aside from saying that Google invested in Pixazza to support a promising and innovative new advertising technology, I don’t have much information I can share,” Google spokesperson Andrew Pederson told InternetNews.com.

Right now, the Pixazza technology is being used at fashion and apparel sites, but Lisbonne tells InternetNews.com that the company is planning to expand to the travel, sports, electronics and home furnishing categories.

“It makes sense to start with fashion, there’s a very passionate audience and photos are key parts of those Web sites, so it’s a good match,” Lisbonne said. “But our vision involves all the major e-commerce categories in time.”

He said installing the technology is “as simple as adding Google Analytics,” and involves copying and pasting a line of Java script on a Web page.

Crowdsourcing for commerce

Here’s how it works: Pixazza uses its proprietary platform and a group of subject-matter experts — at present, fashionistas — to identify taggable items in photos and connect them to similar products carried by its network of merchants.

On a site using the technology, visitors can mouse over the images to get additional information about products within a pop-up box. Clicking on the text takes them to an individual product page where they can get more information, or even purchase the item from one of Pixazza’s participating vendors.

The operation thus relies heavily on what industry insiders call “crowdsourcing,” enlisting the efforts of many people over the Internet, who each do a little work toward a common goal.

For the army of product experts comprising the crowdsource, money is an incentive: They get a slice of the revenue. When a consumer clicks on an item and buys through the Pixazza tool, it generates a commission that’s shared with the crowdsource expert and the publishing partner.

Lisbonne added that the approach sets his service apart from other online product recommendation technology because Pixazza is not solely relying on an algorithm to deliver results.

“Computers are great at some things, but awful at others, and one of those things is identifying products in a picture,” he said. “But our product experts have a passion for this, and this is how our technology approach is distinctive. It builds on our prior experience at LiveOps, which did crowdsourcing for virtual call centers.”

At the same time, the Web publisher housing the image also gets a cut of the commission from advertisers, as does Pixazza.

“It’s a recession-proof business model because everyone wins: advertisers get new customers, crowdsourcers get the opportunity to make money, the Web site publisher gets incremental revenue stream and Pixazza gets paid for enabling all of that,” he said.

While Pixazza is now opening its service to publishers, the service has been undergoing testing since November 2008 and is already live on multiple Web sites.

Get the Free Newsletter!

Subscribe to our newsletter.

Subscribe to Daily Tech Insider for top news, trends & analysis

News Around the Web