Image Recognition on Visual Web Fills In for Focus Groups
Crowdsourcing has replaced focus groups, an executive recently declared in a New York Times article titled “Crowdsourcing to Get Ideas, and Perhaps Save Money.” The executive happens to have a vested interest in proclaiming the death of the focus group. His company, UberTesting, offers access to consumers who are willing to be recorded remotely as they try out products and navigate through websites. However, he has a point.
Visually-driven market research of the crowdsourced variety is definitely having its moment. The focus-group model of coaxing a dozen or so strangers into a conference room to critique a product is still a huge business, but more and more brands are looking for market-research solutions that are less costly, more immediate and more authentic. Remotely observing consumers in the actual marketplace surely offers real-world information and nuances that traditional focus groups simply aren’t designed to elicit.
The truth is, a lot of marketers are overlooking the wealth of product-centric visual information that’s within immediate reach. Consumers are increasingly flooding social media channels with feedback about products. They are doing this not only through text-based expressions, but by sharing images of products they use every day with no prompting or incentive. In other words, authentic, real-world, visually-driven consumer feedback about your products is out there already. You just have to find it and make sense of it.
Case in point: A Twitter user recently posted a photo of himself holding a package of Dove Men+Care soap labeled as having “purifying grains” along with caption: “So are @Dove ‘purifying grains’ just as bad as #microbeads or do they dissolve?” That’s perhaps not surprising given all the recent bad press about microbeads—plastic spheres that have been used as exfoliating agents in skincare products for years, but are now known to be harmful to the environment. In this case, the coupling of the microbeads question with a photo of the product literally in hand essentially serves as a critique of the product packaging. While Dove clearly regards the “purifying grains” language as an expression of a product benefit, this consumer regards it as a point of uncertainty.
Capturing Brand Images
Thanks to image-recognition technology, consumer product shots like these can now be identified by brands regardless of whether or not a consumer name-checks the brand. Logo-detection algorithms can be programmed to pick up on any logo on any product package.
For example, a recent Instagram post from PETA labeled “Vegan Snacks at 7-Eleven!” includes such products as Cliff Bars and Sabra Hummus. It was “liked” more than 1,500 times. Again, an image like this can be surfaced through image-recognition technology that recognizes distinctive product logos, even though the Instagram post didn’t mention any of the products shown in the photograph.
Should the makers of these products do a better job of pointing out their vegan friendliness? Those are product-positioning and packaging-design judgment calls—that Instagram images like this one bring to the fore.
In one recent case, the makers of Hidden Valley Ranch, long marketed as a salad dressing, used image-recognition technology to find social media images of its product in use. What the company discovered was that more and more consumers were using it as a dip for bar foods such as chicken wings. Using that real-world intelligence, Hidden Valley Ranch revamped its product packaging to position it as a “Topping and Dressing.” The version of Hidden Valley Ranch that comes in a squeeze bottle even includes a serving-suggestion photo showing it as a dip for veggies and wings.
Just a few years ago, it was impossible to “listen” to uncaptioned social-media images, much less get in-the-field market intelligence from such a broad and numerous group of people. Now, thanks to massive advances in image-recognition technology, the millions of consumers “talking” via social media images about product benefits, product shortcomings, use cases and more can finally be heard. Smart companies are beginning to use visually crowdsourced intelligence to optimize the positioning of their products.