The tortured idiom "a picture is worth a thousand words" is only true if you understand what the images are saying. When Pinterest and Instagram started getting really big a few years ago, it signaled a major shift in the way people share experiences on social media. One result is that it has become much harder for marketers to track mentions of their brands. Some 1.8 billion images are posted to social media each day, according to Mary Meeker's annual Internet Trends report, yet 80 percent of them lack identifying text.
Enter image recognition, which works by analyzing pixels in images to identify faces, objects, products, etc. This technology is transforming a wide variety of industries, from agriculture and home appliances to social media and native advertising. GumGum, an image recognition-based digital marketing company, highlighted recent advancements in an infographic, "Image Recognition: Year in Review."
For marketers specifically, this technology provides new ways to listen to and engage with consumers as well as glean competitive insights. Let's take a closer look at what this means for your brand.
How do marketers feel about the increasingly visual nature of the Web?
There's a lot to be excited about—new platforms, new ways of reaching consumers, new opportunities for creative. At the same time, visuals present a new set of challenges. In the early days of Facebook and Twitter, everything was text-based. For digital marketers, this made tracking your brand by name pretty easy. The visual Web dropped a cloak on social media. For example, if someone shares a photo of a brand but doesn't tag the brand with a hashtag or caption, there's nothing that tells the marketer, "Hey, I'm talking about your brand, so you should listen." This is where image recognition can help. Take the Wolfram Language Image Identification Project, for example, which can identify the content of any picture uploaded to the site. That's a pretty promising trend.
So how can marketers use image recognition in a social media context?
Image recognition makes it possible to find and categorize those untagged photos by analyzing their visual content for patterns specific to a brand (consumers holding cans of Coke, for example). This provides brands not only with useful content they couldn't find before without accompanying text, but also with better connections to their biggest fans. Real-time image recognition tool Visual Intelligence is used for "visual listening": It watches social media as new images appear, then grabs the ones that are relevant to the respective brand and displays them on a dashboard. It works similarly to many of the text-based social media listening tools that marketers use today, except it's using image recognition technology to scan the visual content in the picture instead of relying solely on words. Visual Intelligence also analyzes the demographics and sentiment associated with these images, so marketers can quickly get a sense of who's talking about their brand and what they're saying.
Why is this so useful?
Marketers can use tools like Visual Intelligence to identify the top influencers and engage with them through direct advertising or obtain rights to their photos for user-generated content. That's huge. The next time an influential consumer posts a selfie with their face smeared in chocolate from their favorite candy brand, it could turn into a yearlong advertising campaign. In that sense, "going viral" becomes more scientific and accessible for marketers, so they can ride the wave rather than chase after it. Brand managers can also use Visual Intelligence to benchmark their brand against competitors to get—err—the big picture of their industry on the visual Web.
Can image recognition and the Internet of Things work in tandem?
Yes—the smart home is a great example of this. At CES 2015, Panasonic demonstrated several image recognition-ready home appliances—adding cameras to everything from refrigerators and ovens to dishwashers and mirrors. Each appliance is internet-enabled, allowing consumers to track all of these appliances, including the products used with them, on their phone or tablet. The image recognition capability combined with the Internet of Things creates a stream of actionable data that benefits consumers and marketers alike. For example, Panasonic's smart mirror uses image recognition to identify and analyze your skin's condition, then makes beauty and grooming product recommendations that can be purchased immediately.
What other applications for image recognition are anticipated for marketers?
A few months ago, Pinterest launched a search feature that enables users to tap any object within a photo to match it to similar ones found in other pins—including where to purchase it. There's also a new software program out of Edith Cowan University in Australia that can determine the species of a plant based on a photo of a single leaf, kind of like a 'Shazam for plants.' For farmers, this could lead to better, faster ways to monitor crops. For food marketers, this could translate into new ways to connect consumers with the people growing their dinner—from farm to plate. Marketers are really just starting to understand the full potential of image recognition technology, and with each new image, there's a new opportunity.
Illustrations by Andrey Smirny