In the early years of the commercial Internet at the tail end of the past millennium, Google only provided search functionality as a product and did not make any significant steps to monetize search as a service. Search terms were just that: words.
In October 2000, the company introduced Google AdWords as a way for brands to market their own products and services through search engines. The idea was simple: People were already entering words and phrases to search for products and services to Google, so why not sell the resulting keywords and search terms to marketers?
With AdWords, Google was able to directly use its search engine platform to advertise to its users, thus creating a whole new industry of search engine marketing (SEM), which has since grown to be worth more than 25 billion dollars.
Today, the Web is about more than just text, not to mention desktops; it has gone visual, and mobile. More than two billion images are shared online each day, and the growth of visual-first mobile apps such as Instagram and Snapchat -- not to mention Pinterest and the increasingly visual Facebook, Twitter, and WhatsApp -- continues on pace. And most of that growth is on smartphones and tablets.
With all those visual assets out there floating around, it’s no surprise that the next frontier of search, much less advertising, is figuring out how to monetize those pictures.
One of the key ways that marketers are approaching this opportunity is through image recognition technology, which uses deep-learning algorithms that can analyze pictures and correctly identify in them such aspects as objects, logos, and faces at a rate and volume much faster than ever before. Thanks to rapidly growing artificial intelligence (AI) technologies such as deep learning -- the same kind used in voice recognition, translation, and driverless cars -- image recognition algorithms become smarter and more accurate the more pictures they analyze (literally millions and millions of them).
Not surprisingly, Google is also a key player in this space. Image recognition is already being implemented in such consumer applications as Google Photos, which automatically analyzes users’ photo collections and organizes them by person, place, subject, activity, and other categories that previously were only classifiable by human-submitted tags. Google Photos is just the front-facing consumer application -- which, in part, also helps the company's image recognition capabilities improve (all those user photos). But this technology, which is so innovative and useful, is also being monetized, in some ways similarly to the way that search words have been monetized.
On Pinterest, for example, when pinners zero in their cursor on a specific object in a picture -- say, a table lamp -- image recognition is used to surface other pictures containing table lamps. It is also used to gather information at a more granular level about consumer interests. This insight can then be implemented in targeted advertising. Image recognition also powers in-image advertising, whereby the content of images not only helps with targeting where ads show up (in pictures of specific celebrities for a hair care product), but also with execution (blond hair-specific products for ads appearing in pictures of blond celebrities).
Even on less overtly advertising-focused initiatives, image recognition can help brands gain an edge. The smarter the image recognition algorithms become, the more insight they can provide to brands: How users share images of products on social media, for example -- whether or not there is any identifying text -- can inform marketing strategies down the line.
Google has been exceptional in terms of monetizing search, and there are lessons to be learned from its success. Consumers already willingly give away their search terms and keywords to Google, but they also give away their images to social media platforms. Billions of images shared online every day offer exceptional opportunities for monetization, especially if you get the right technology on it.