The Eyes Have It – the Rise of Image Recognition in Marketing
Image recognition and visual search on the cusp of transforming social media marketing as we know it. The momentum has reached a tipping point and expect to see major adoption in 2020.
Image recognition at scale is transformational
Only a few short years ago, in 2016, Instagram did not have native image recognition. Today, with the help of Facebook’s vast machine learning expertise, one billion people use Instagram every month and 500 million use the platform every day. (Compared to Facebook with over 2 billion monthly active users and YouTube with 1.9 billion monthly active users). While Instagram is only the 3rd largest social network its engagement rate is 10 times higher than that of Facebook, 54 times higher than Pinterest’s, and 84 times higher than Twitter’s.
What is even more stunning is that now Instagram processes hundreds of millions of images in real time every day. In determining what appears in your stream Instagram is moving to rely more on its own analysis of your images than the hashtags which you include with your posts.
That is a quite profound development as you could say that while hashtags can “lie” and are often used for newsjacking the images don’t lie. Perhaps in the near future, Instagram will start penalising images and hashtags which are incongruent e.g. penalising “fake news”.
Perhaps also, the streams of quote images pumped out by the Instagram “virtual assistants” hired by so many accounts will be penalised in favour of real content.
Attention to clear images with a clear message which is congruent with your brand will become an important task as image recognition at scale becomes routine.
Visual search moving to the mainstream
Pinterest processes over 1 billion visual searches per month. It is the leader of visual search for ecommerce, and has struck visual search alliances with firms such as Target, Walmart and Home Depot. Target dove into visual search headfirst by integrating Pinterest’s visual search tool, Pinterest Lens, into its own mobile app. From inside a Target store, customers can snap a picture of any product and the app will pull up similar items.
While retailers encourage customers to use the store brand app, so that a visual search brings up a selection of items from their own inventory, customers are rapidly moving to comparison apps or using Amazon’s visual search. Retailers are bracing for the onslaught of customers scanning their products for instant price comparisons (and then asking for a price match).
It’s extraordinary that the huge technical and intellectual challenge of super-fast and very accurate image processing has become something for which we expect an instant response on our phones.
Images are clearly only going to get more important in brand content. In fact, the 2019 Brand Content Study by Adobe found that 17% of consumers are annoyed by social media content without images or video.
Website design needs to adjust
While we generally think of visual search on our devices and when using our browsers for searching, just like chatbots visual search is going to impact website design.
Retail sites will need to offer a quick way to use images to search their sites and to make it easier to use visual search as a way to start a conversation with a brand’s chatbot.
Capturing visual insights via social listening
With the massive increase in the use of images and less dependence on text – due to the power of image recognition – social search is racing to keep up.
While image recognition as part of social media listening has been around for some time it has not yet hit the mainstream. That will change in 2020.
Brands now need to leverage the intersection of text and visual analytics to discover with greater clarity how customers feel about them, and immediately act on visual intelligence to power more meaningful, targeted connections.
For example, Sprinklr’s Visual Insights capability provides image recognition across all the main channels. In conjunction with the textual analysis the added visual analysis significantly enhances the ability to take action, the quality of market research insights, and the ability to protect the brand reputation.
But the true power comes from its ability to improve customer experiences in the moment. This means brands will be able to simultaneously analyse text and images to gain the deepest possible understanding of the content, and take action in response. For example, responding at live events and more rapidly utilising user-generated content.
The overriding message is to be prepared to embrace and to plan for image recognition at scale in how your content is consumed, and in how you analyse and gain insights from social media monitoring.
