Earned media platform Cision adds image tracking with ShareIQ purchase
ShareIQ’s technology is designed to track huge numbers of images across the web, even if they have been slightly modified.
Cision says its recently announced acquisition of Berlin-based visual content performance platform ShareIQ provides it with the technology to offer the first earned media platform that tracks visual as well as textual content.
Previously, Cision’s platform tracked image performance for brands, based on text associated with the image via social media hashtags or written metadata. It did not track the performance of visuals directly based on the image.
But ShareIQ’s patent-pending platform turns an entire image into a hash, then tracks it across the web even if there are small changes.
A hash is a unique code generated from a given file of data — in this case the configuration of bits in the image. With a cryptographic hash, a small change in the original file means a big change in the hash code, which is why a cryptographic hash is used to ensure that there has been no tampering with, say, a transmitted message.
But a perceptual hash is different. It accounts for small variations in the file, while still producing the same hash. This allows a kind of image fingerprint that continues to track the whole image, even if there are minor changes, such as optimization of file size or colors.
Because of this capability, a perceptual hash is often used to find out if a copyrighted file has been modified slightly and used without permission on the web. But it is difficult to scale perceptual hashes across the huge number of images that brands want to track.
Cision says there are over 3 billion images uploaded to the web every day, with 95 million images per day on Instagram alone. ShareIQ’s job is to track those images for brands, by downloading every image from every social network.
To accomplish this goal, it has come up with a proprietary way to apply perceptual hashes for images at scale, across zillions of images every day.
ShareIQ founder and CEO — and now Cision VP of Innovation and Audience Products — Brian Killen told me that this “massively optimized high scale perceptual hash” can track various image variations back to one image. He added that no other perceptual hash technology operates at this scale.
He pointed out that this is different from machine vision, which actually recognizes a logo, object or a face in a photo, instead of turning the entire image into a single tracking code. Here’s a screen shot employing ShareIQ’s tech, showing what images are shared from given URLs:
The ShareIQ tech means that Cision’s earned media platform can not only track, for instance, where a photo of a barbecue grill or other product has been shared on the web and the resulting engagement, but it can also target the users who shared it.
This is accomplished, Cision President of Data Solutions and Innovation Dave Barker told me, by associating the URL where the photo appears with a cookie or a mobile device ID, via an identity resolution provider like LiveRamp or an ad tech firm like MediaMath.
Since they track the online activities of huge numbers of users, they will know when User A’s browsing includes this particular URL, so they will then know User A’s cookie or mobile device ID.
When User A shows up at a site showing ads via an ad network that Cision works with, an ad for the same brand that generated the photo can be displayed, such as a sale on that barbecue grill.
Barker said that 40 to 60 percent of URLs where the image appeared can be successfully turned into targeting IDs. In addition to targeting online ad displays, this URL-to-ID can also be used for attribution, since LiveRamp or others might be able to determine when that user went out and bought that barbecue grill.
“We’re sure Google and other tech titans have [similar] sophisticated image tracking/identification technologies,” he said, “but none that offer a platform for marketing or communication professionals to utilize.”
Opinions expressed in this article are those of the guest author and not necessarily MarTech Today. Staff authors are listed here.