Image Recognition Analytics and the Future of Social Marketing

Image Recognition Analytics and the Future of Social Marketing

We exist in a very visual society. Thanks to social media, we are inundated by more and more images every day. Marketers have relied on their own two eyes to judge these images on social media — looking for patterns, moods, brands, and other clues to consumer behavior.

Last year, marketing agency Room 214 raised the question, Are Photos the Future of Social Listening? We think the answer to that question is a resounding yes, especially within the context of digital marketing. Social media data is a treasure trove of consumer insights, but social data from images, however, is still being explored. Here are a few features from Scraawl that offer image analytics and predictions for the future.

By the way, did we mention that signing up for Scraawl costs nothing and gives you access to the free personal account?

Media and Portrait Gallery

The Media Gallery Advanced Analytic in Scraawl gathers all media included in the collected social media posts and organizes them by popularity and frequency. This is useful for finding trending images and videos in social conversations or to judge the popularity of a marketing social campaign.

To find images that contain actual people, Scraawl offers Portrait Gallery, an Advanced Analytic that leverages face recognition technology. While face recognition technology has been used by the security and cyber-security world for a while, it is making further inroads into marketing. There’s a growing potential for truly personalized marketing and more innovative campaigns. Imagine a future where consumers can try on retail items without ever stepping in a brick-and-mortar store. Some see even greater potential for face recognition technology in email marketing.

In the context of Scraawl, the Portrait and Media Galleries have been used to get a sense of visual trends that may not be expressed in the body of social media text. It goes without saying, visuals are important in the age of social, but with analytics like the Media and Portrait Galleries, marketers can optimize their campaigns by picking images that truly appeal to the current social conversation.

Avatar Analytics (beta)

In Scraawl 3.0, we released a whole new feature called Avatar Analytics. Useful for finding individuals in the same affinity group that may not be explicitly interacting with one another but clearly care about the same topics and issues. For brands, this may be a worthwhile opportunity for connecting to these users and creating a stronger influence network.

In the example graphic above, we searched for #LWYMMDvideo during the height of pop singer Taylor Swift’s music video release, and took only a sample of 5,000 posts. The Avatar Analytic found a number of Twitter users with the same album cover as their profile pic. What’s interesting here is that within this sampling of Taylor Swift fans, there was also a smaller group that used a different image of Taylor, one with a shaggy haircut, winged eyeliner, and a black semi-turtleneck.

A social marketer would ask these questions, why did this group pick this image over the album cover? Is this an edgier representation of the pop star that seems to resonate with a certain demographic? What patterns exist here?

Despite the shared avatar image, these users did not necessarily interact. When we plugged two of these users’ handles into the Social Graph analytic, we looked for a common connection. Unsurprisingly, both users tweeted #LWYMMDvideo, but they had no connecting vertices indicating they had no direct engagement, nor had they replied to or mentioned each other. They were suspended in space, near to each by a shared interest yet unconnected.

The lack of interaction between two users that have so much in common could be seen as the end of the line for some, but for others, it’s an opportunity for brand to community (b2c) marketing. In this case, perhaps an influencer would be the best way to bridge these users, leveraging real relationships to connect users on a larger scale. By instigating a conversation between these users, a marketer could leverage both of their networks to then reach more people, creating stronger affinity groups.

Note: The feature is still in beta and currently available only to an exclusive group of Scraawl users, so get in touch if you would like to see this capability in action. Contact us.

5 Predictions for the Image Analytics

As the capabilities for image processing continue to grow, we’ll begin to see the marketing industry embrace the new technology, as they have with many other analytic tools. Here are five predictions for the future of image analytics and digital marketing:


Image analytics will play a bigger role in marketing, from market research to ad targeting. Image and entity recognition become more widely available thanks to services like IBM Watson and the Adobe Marketing cloud.


Digital marketers and marketing data analysts will have more overlap in their roles, especially in the context of analyzing images for social listening.


Affinity groups will no longer be identified by common phrase, hashtags, and keyword analysis. They will also be identified by common image or similar visual trends.


Images analytics will pave the way for better video analytics. Not only will brand logos and faces be recognized, consumers will be able to watch a video and pick out a cool product that captures their attention. There will also be better tracking of and greater transparency into video metrics, a current issue in sponsored video campaigns.


Image analytics will be integrated into geo-spatial, text, and network analytic technology empowering digital marketers to become social data experts for their brands.

Images are key to understanding consumer behavior, especially as we as a society rely more and more on images to convey our emotions and identities on social media, and with Big Data, there’s huge potential in what patterns can be discovered.

Subscribe to our email Newsletter

Sign up for Scraawl news, tips & tricks, and latest blogs. Don’t worry, we never ever spam our listserv and we only email once a month.

Related Posts---

Leave A Reply---

Back to top