Tech companies are building, buying and integrating visual tools in an effort to better understand visual content.
Tumblr announced earlier this week that it signed a deal with visual search firm Ditto Labs to analyze user photos for “clues” about brand affiliation.
David Rose, CEO of Ditto Labs, told Mashable that the company’s technology anatomizes an individual’s photos so thoroughly that it can establish “if you’re a sports guy or a foodie.”
In January, both Pinterest and Google bought companies specializing in such technology; the former acquiring VisualGraph, a firm that provides image-based search technology, and the latter buying DeepMind, an artificial intelligence company with the potential to improve image search.
So what’s the attraction? Online photo sharing at its current scale is unprecedented, and it represents a relatively untapped treasure trove of information about brand loyalty, influence, sentiment, product use and more—much of it valuable for the purposes of customer research and advertising.
Mashable notes several brands have already drawn important conclusions from photos:
“By parsing photos, Chobani was able to determine that many consumers were putting their yogurt containers in car cup-holders and engaging in some dashboard dining. Designer handbag maker Vera Bradley was also able to determine that its products were often given as a rite of passage, frequently to girls hitting their 16th birthday.”
Some of this information may also be helpful in determining how to better influence customers at the initial point of contact. Researchers at York University in England recently revealed that they have developed a model that can predict how people will react to certain facial characteristics in photos. As one of the scientists told BBC, the model is “obviously potentially very useful” to businesses.
All of these advances represent a welcome expansion of the commercial ecosystem surrounding user-generated photos: visual commerce. The fact that companies like Tumblr/Yahoo and Google are investing in that ecosystem is yet another validation of its promise.