How Stock Image Providers Can Provide Predictive Ad Targeting, Improve Search For Marketers
Marketers may soon look toward companies that license stock images and videos to integrate their third-party data into their predictive ad-targeting models.
Today, photographers contributing images can upload content and identify trending keywords, but the need for more visual, textual content could turn brands like Shutterstock into third-party analytics companies that provide predictive analytics based on analysis of trends. The company has begun rolling out an application programming interface (API) with Facebook to support agencies and brands looking for images to serve in ads on the social site.
Hundreds of signals create Shutterstock's search algorithm like clicks, downloads and hovers, but territories, language and more all play a role to create better search experiences. The company also brings in data from other sources, and has begun testing 4K videos aimed at supporting movie studios and television stations requesting the ability to run the clips on their sites.
Shutterstock, which analyzes three downloads per second from its site every 24 hours, already supports a suite of tools identifying trending topics and keywords based on consumer behavior, location, and data within videos and vectors. Some of the most common trending words include infographic, appetizing, BYOD (bring your own device), adorable, and responsive design. The trick is matching the phrases with products and services.
The data inside images helps Shutterstock make a correlation between objects and intent. Clicks on pages and content-based data, the ones with more or less focus, or those with specific colors can help to predict trends," said Wyatt Jenkins, Shutterstock head of product development.
Labs, a Shutterstock division, highlights experimental search prototypes being tested on the site. One allows visitors to find images based on color. The data produces color trends in search queries -- data that marketers could use for audience targeting across engines, as well as publisher and social sites, given the correct platform.
All searches and downloads on the company's Web site are tracked and analyzed. It does not
track the shares of images on publisher or social sites, but Jenkins said it's technologically feasible with consent from the site owner. Tying the share into the data could add to predictive
analytics to better target ads.
"Numbers" photo from Shutterstock.