For years, we've been hearing that big data is going to shape the future, and here we are. We recently unveiled Dash Hudson's very latest product, Vision, a software powered by machine learning technology and computer vision, to help our amazing customers do what they do best. But then do it even better. And now we can prove it.
Understanding the magnitude of connecting through photos and videos is vital in today's media entrepreneurship landscape. Brands have access to billions of visual consumer touchpoints, and the need for smarter, more effective messaging through these mediums is a pressing matter. Most contact between businesses and their customers now occur through images, making it increasingly dire for brands to be well versed in what audiences want to see, as well as what they respond positively to.
Thomas Rankin, our co-founder and CEO, states that "the main point of interaction between brands and their consumers is really happening through the photo." Visual storytelling is not only today's most important communication vehicle, but it's also critical to creating meaningful connections with audiences. If a business gets this wrong, its chances of keeping on the path of success reduces dramatically.
Social media is the most effective tool to capture the attention of consumers, where relevant, uninterrupted dialogues are being initiated through visual content-sharing. It's probably the utmost significant above-the-line marketing channel for brands to focus their efforts on. And it's deceptively complex: it demands both a highly creative and analytical skillset from its experts — two things that are often mutually exclusive.
The thing is, while more imagery is being created every day than at any other point in history, engagement finds itself on the decline. "People are just paying less attention to any individual photo because there's so much more out there. So they're getting a little bit tired of engaging with certain photos," explains Thomas. Audiences are fickle and computer vision applications are onto them. Dash Hudson's Vision is the answer to finding the right piece of content for the right audience at the right moment.
We built our image recognition search and machine learning software to empower consumer brands and digital publishing businesses to bring their content A game. We're helping companies reconcile what types of visuals resonate with their audiences through photo recognition, and improve their output through image search.
Thomas recently sat down with Kelsey Simmons, senior business intelligence analyst at Clique (home to Who What Wear, College Fashionista, and MyDomaine, to name a few), on the World Forum Disrupt DigiPublish stage to discuss how Dash Hudson's new technology drastically boosted the content performance of one of the company's brands.
One of Clique's main brands is online aesthetics destination Byrdie Beauty, for which Instagram is a major channel. And for good reason: the niche is intrinsically visual, and the vertical's beauty-obsessed audience lives on IG. In fact, the Clique group understood the power of Insta very early on. As Kelsey states: "When social media first came around, a lot of publishers thought about it as just a channel to distribute site content [...], but there's so much more about Instagram that's powerful for your brand."
HT to CMG for identifying that social was not just there to provide digital publishing solutions, but indeed an amazing opportunity to build a larger-than-life presence. But guessing what people will feel compelled to interact with is not an obvious task, and companies don't have the means to waste precious time on riddles. The thing is, reaching a following of over 287K can be tricky in an algorithmic world where content relies on engagement to be seen.
Because Clique is a leader, it was not interested in always reposting the popular Kylie Jenner shot. The company's approach is to choose more unique imagery that builds their brand, and identified Dash Hudson's image recognition technology to be the solution its creatives needed to make content decisions. Clique was seeking answers to two main problems: surfacing the most relevant user-generated content to repurpose, and identifying campaign images with the highest performance potential.
To get a better sense of how Vision could help, the Byrdie Beauty team elected to run a test. They uploaded thousands of images to their Vision dashboard to be analyzed, and for two straight weeks would only post images that were predicted to be high or top performers, as well as the photos that had already been scheduled prior to the experiment.
Over this period of time, the Vision-predicted posts saw a 65% increase in engagement rate over the ones that had already been planned, and a 40% increase over the previous two weeks.
How 'bout them apples?
![computer vision technology, machine learning platform, visual recognition, digital publishing platform for magazines, cmg, dash hudson, byrdie beauty ](/content/images/2017/11/computer-vision-technology--machine-learning-platform--visual-recognition--digital-publishing-platform-for-magazines--cmg--dash-hudson--byrdie-beauty-.jpg)
Kelsey and Thomas equate Dash Hudson Vision to a mini audience focus group. Except that this one evolves in tandem with it and is constantly learning and adjusting to its whims, desires, and tastes. "The nice thing about [Vision] is that it almost becomes a sentient thing about your brand; the more that you use it, the more that your audience responds to it, and it grows with your audience over time," states Thomas.
Kelsey anticipates that "image recognition is going to continue to grow," forecasting the technology's evolution to reach even deeper subjective abilities, enabling a more profound understanding of audiences and tastes. And then robots will take over the planet. Jk. But maybe they will a little.