Everyone knows the number and it doesn’t lie: 1.6 Billion images uploaded daily. Out of the main platforms people use to communicate daily, the top 5 (Instagram, Snapchat, Whatsapp, Facebook, Pinterest) use almost exclusively photographs. It is estimated that 75% of the web’s real estate is also made of images. Yet marketers seem to completely ignore this vast pool of extremely valuable information, choosing to rely instead on inefficient hashtags or vague keywords.
The numbers tell the story:
Snapchat (8,796 photos shared per second), Whatsapp (8,102 Photos shared per second), Instagram (810 photos per second), Flickr (41 photos per second), Pinterest (1,388 per second) and Facebook (208,300 per second). Add Tumblr, Twitter and self-publishing platforms like WordPress or Medium and the numbers keep on rising. Photos are, by far, the most popular mode of communication, regardless of age, sex, race or nationality. They carry every single type of emotions human beings care to share, along with habits, likes and dislikes, intentions, desires, and opinions. With photos, people share what they like to eat, wear, listen to, play with, places they like to visit and the activities they enjoy doing. Far more than any survey or market analysis could ever reveal. Yet, marketers do not take advantage of this information.
Are they blind?
Partly. The main reason marketers ignore this data is that they do not have the proper tools to mine it. Only recently ( last week, in fact) has Facebook unleashed its content recognition engine into its platform. While allowing anyone to do searches that will consider the content of images (instead of just the caption), it is yet not clear how this will benefit marketers. It is assumed that in the future, Facebook will release some of the data acquired to allow advertisers to connect with users based on the content of the images they shared. It might even extend to letting marketers better understand how products are consumed via the analysis of millions of images. Facebook, for the time being, is silent on the topic. There is no doubt that this technology will be also unleashed onto Instagram and maybe WhatsApp. Regardless, this is only one solution, and a very limited one.
What is Martech doing about it?
Surprisingly enough, almost nothing. One would think that considering the vast landscape of adtech/martech start-ups, the majority would be all over solving this. Not at all. Only a tiny few have integrated content recognition into their offering. Which is baffling, considering how easy it is to implement. Companies like Imagga, Clarifai or Ditto Labs offer easy, performant and very affordable API’s that makes adding content recognition a breeze for any developer. No need to sell their souls to the Amazon, Google or Microsoft of the world. With solutions that range from auto-tagging (recognizing the content of images) to auto-categorization ( auto-classifying images in groups), color identification and even content moderation (identifying those NSFW pictures), these companies offer the full scope of A.I. powered content recognition for a fraction of the cost it would take to implement in-house.
A better world?
Imagine being able to accurately target an audience by the type of images they share and their content. Finding all those that are currently consuming your competitor’s products, where, when and how ( which they hardly voluntarily identify). In turn, being able to communicate with them with a precise, custom offering. Being able to appear in the images they look at, or even share, because it corresponds to their taste. Understand what dominant color they prefer and what repeatedly appears in their images. Or even just better understand a target group lifestyle, via an analysis of thousands of images, done in milliseconds. With visual content recognition, marketers could find out, for example, what is the number one preoccupation of teenage girls in the US right now or what style of cloth they prefer, just by analyzing the images they share. And for the consumers, that means a far better experience, away from interruption advertising that they increasingly ignore with adblockers.
When analyzing the content of images, the first reaction is always privacy concerns. Most photos shared are considered private and no one wants to see uninvited onlookers using them to spy. But like with big data in general, visual content recognition is identity blind, for a few reasons. It doesn’t need to know anything about the creator of the image nor its source to understand its content. Furthermore, it cannot generate accurate analysis without large data set. Analyzing one image is useless. It needs thousands, if not millions to extract a valuable trend. Finally, it does not store any image anywhere. Once the image is analyzed, it is trashed forever. Thus the privacy of the individual is fully preserved.
It is obvious, isn’t it? One or a few Martech companies will understand how to implement visual content recognition technology into their platform and make a huge leap forward, leaving their competition in the dust. As well, some tech savvy ad agency will develop in-house tools to take advantage of the technology and offer a level of service far beyond anyone else. Brands who will adopt image A.I. in their marketing approach will certainly see the efficiency of their campaign dramatically increase, as they will surf the image sharing wave with laser accuracy.
Photo by mrbill78636
Author: Paul Melcher
Paul Melcher is the founder of Kaptur. He is an entrepreneur, advisor, and consultant with a rich background in visual tech, content licensing, business strategy, and technology with more than 20 years experience in developing world-renowned photo based companies with already two successful exits.