A social insights word problem: If a picture is worth a thousand words, how much is a human analyst worth?
If there is one major way in which social media has changed over the last decade, it is the dramatic rise in the use of images, rather than text, to communicate. Whether it is our use of GIFs to respond to tweets or pictures posted on Instagram or Facebook, or using emojis to respond to a text message, images no longer simply augment text—they are an integral part of the actual conversation. People respond to images more readily than text, images evoke more powerful emotions than text, and people tend to remember ideas and content from images more readily than they do from text.
Our reliance on images as a means of communicating has always been present, so it makes perfect sense that images are being used online for storytelling.
Visual Content Strategy
Using images in addition to or even in lieu of text has become so important and so prevalent that firms are now recommending that clients have a specific visual content strategy. While this is more applicable to some clients and projects than others—business to consumer brands are more likely to see ROI on a visual program than business to business brands—it’s useful for all PR and communications professionals to understand why a specific strategy surrounding visual content is important.
Social platforms know that visual content creates more engagement than text, so it is in their interest to encourage users to share images.
As the popularity of using imagery on social channels grew, more platforms added features that made it easier to share images. Sharing GIFs on Twitter and Facebook, creating stories around Snapchat, and the incorporation of video in the form of Facebook Live are all developments that collectively reinforce and promote the use of pictures. Pinterest and Instagram are all about photos.
This presents some challenges for communications professionals, some of which are easier to address than others. The first challenge for brands is finding the content in posts that might not mention a brand directly. Effective keyword or search phrase development, the ability to monitor hashtags, and logo identification software address this issue.
A more pressing challenge is the inability of automated monitoring programs to identify and apply sentiment to a series of images—thus far, there isn’t an available market solution that “reads” and understands a series of images.
The rise in the use of images as the dominant component of online content means that we can’t rely solely on automated analysis—we need humans who can visually process and understand what an image is conveying between the poster and the reader. For example, posts on BuzzFeed use images to tell a story. I went to the site and grabbed the first piece that caught my eye (in other words, I did not hunt around for this item). Read through the text. Without the images, what conclusions would you reach about the post? Was it positive or negative? What is the writer attempting to convey? Without considering the context provided by the images, was she successful in conveying her message?
This type of storytelling is effective only when the person reading the piece can incorporate the context provided by the images, which, for now at least, means a human being. Considering how widely used this type of post is (where the text and images contrast with one another, ultimately conveying a meaning that is the reverse of the text used), automated sentiment analysis would likely get this wrong.
Twitter responses now often include GIFs or video. Some tweets are simply pictures or GIFs, with no text. Again, analysis for these types of posts and responses requires human analysis.
Will automation come to image analysis?
It’s tempting to think that automation to analyze image-heavy content is just around the corner- and it might be. Getting too comfortable in that thought process is probably unwise, since the way in which social platforms have grown over the last decade has demonstrated we don’t necessarily know exactly how communications will change. If static images are analyzed by automation, what about live video? If live video can be analyzed by automation accurately, what about augmented reality—and will the analysis of either of these be able to determine the possible intent or meaning of the videographer’s filming choices?
Human communication can be frustratingly complex at times. Even written communication can present issues for automation, and imagery can be even more challenging for a computer—but a human can look at a GIF response to a tweet and immediately “get” what is being conveyed. For now, human analysis remains a vital component of social analytical work.