A successful AI deployment is a must to stay competitive, but it goes beyond just data prep — it requires getting people, process, and principles all in place too. To learn how to avoid the pitfalls, get your strategy planned, and make sure you’ve got the infrastructure you need in place, don’t miss this VB Live event!
The designer lingerie and clothing company Natori is a design-driven brand, and a family business, explains the brand’s president, Ken Natori. The founder is still the CEO, and very involved in every creative decision. The brand’s creative has always always been controlled and calculated, from the assets themselves to how they’re distributed, but says Natori, AI threw the company a curveball.
“Obviously with AI you’re not making every granular decision,” he says. “You’re taking a bunch of assets that you approve, and some copy you approve, throwing them all into a machine, and then that machine is mixing and matching that copy with the assets and the audience and seeing what works and what doesn’t work.”
It was, he says, a bit out of character for the company to give up that kind of control over their creative output — and not something everyone was 100 percent comfortable or on board with initially. But as the brand has continued its evolution from a wholesale high-end sleepwear brand into direct-to-consumer with an expanded product line, marketing has evolved too, from the channels people shop in, to the ways consumers interact with brands and how they expect to be marketed to.
Their initial AI forays were via paid social ads. On Facebook, AI pushed out ads for both individual products and product carousels, alongside a number of campaigns catering more to the top of the funnel, specifically looking to reach out to people who may not have heard of the brand.
But the launch came with a warning. Before they kicked off their campaign with a partner, they were told that things would look a little grim at first.
“A couple months in we noticed people saying, why is your stuff always on sale?” Natori explains. “And it turned out that people like buying stuff on sale, and because that stuff was converting, the AI was probably feeding too many ads of sale products. It kind of looked like Natori was a discount site.”
It took some time to tweak the algorithm to match the brand perception. As their algorithms got fed more information, as the company provided more assets and the creative minds continued to do little course-correcting for the AI’s decision-making, things started to change.
Their return on ad spend (ROAS) was somewhere around 200 to 250 percent at their initial launch in June of 2017, but every month from then on, the cumulative ROAS continued to improve, Natori says. As of March or April of this year, it was up around almost 700 percent, and it continues to go up month after month.
“Our results continue to improve, we increased our spend based on that, and our results were through the roof earlier this year,” Natori says. “Based on the success of what we’ve done, we’re definitely considering other applications. We do like having some control over things. But on the margin, five years from now, or one year from now, will AI play a larger role in our technology stack than it is now? Yeah, it’s no question.”
To Read Our Daily News Updates, Please visit Inventiva or Subscribe Our Newsletter & Push.