In building out its “Complete the Look” personalization feature, which is designed to complete an full outfit based on customer style preferences and stock levels, global athleticwear and footwear giant Adidas knew it couldn’t scale a manual process.
According to Katharine Reinhold, director of innovations at Adidas, the brand’s merchandisers had to account for factors such as new product release dates, which varied in different markets, as well as whether they wanted to pair full-price items with product on sale. There was also the problem of a product being out of stock, Reinhold said in a Coresight Research webinar hosted by its founder Deborah Weinswig.
The original process took 27 steps and 20 minutes on average, resulting in fewer than 10 percent of products appearing in outfits. With that in mind, Reinhold and the Adidas innovation team turned to FindMine, an AI-based platform designed to automate the “Complete the Look” process. In the time after leveraging FindMine, Adidas merchandisers spent 95 percent less time working on the Complete the Look feature, increasing the number of items included in outfits by 960 percent, according the AI-based personalization platform’s case study.
“We saw increased rates in repeat purchase behavior, return visitors and then higher average order value, so consumers were really inspired by what they saw,” Reinhold said. “Visitors could have come back and said, ‘That was a cool outfit, I wonder what else is new at Adidas this week.’”
Related Stories
As a gridlocked apparel supply chain faces the threat of holiday out-of-stocks, FindMine tracks product sell-through, enabling brands to see how merchandise is performing relative to plan.
“Why put a bunch of stuff that’s already on fire into all your content?” FindMine CEO Michelle Bacharach said. “It’s going to sell out anyway. Use that opportunity for something that’s a middle performer or a low performer, all the while testing to make sure you’re not tanking your overall revenue by putting something that’s a low performer in everything.”
The Adidas case study illustrates one glaring e-commerce reality. Retailers are just scratching the surface when it comes to personalization, Bacharach said. Even Amazon’s personalized recommendation engine displays too many comparable products, such as multiple jackets after the customer already ordered one.
While the number of Adidas items Complete the Lock shows shoppers has skyrocketed, consumers have a narrower, and ideally more appropriate, selection to choose from.
“Some more of Adidas’ traditional partners in the recommendation space tried to use algorithms to help improve the Complete the Look program, but it would involve pairing three pairs of pants with a shirt,” Reinhold said. “People don’t wear three pairs of pants at a time.”
This happens because most merchants leverage personalization tactics such as sorting based on likeliness to buy, recommending products based on demographics and shopper intent or implementing predictive search, all of which steer the customer toward the right individual product.
Where they miss out is in ignoring storytelling’s potential to ultimately drive revenue and increase customer retention, Bacharach said. This includes investments in “how-to-use” guides for each product, which is essentially the Complete the Look function, as well as context-aware content that follows trending moments, weather patterns or inventory on hand.
“We had this idea of clicking through a Facebook post about the Met Gala and landing in a page that’s all the outfits inspired by the Met Gala. That’s personalized,” Bacharach said. “What’s happening a lot in the awareness and discovery world is that the customer is clicking through on a beautiful ad and landing in the generic homepage, or landing in a product page that has no information to tie back to the campaign that they came from. If you’re doing more of a storytelling moment, recognize that that customer came from a storytelling moment and continue the story wherever they landed.”
FindMine seeks to take brand storytelling to the next level by embedding content into product detail pages, new landing pages, the add-to-cart page and even in email marketing and post-purchase follow-ups. For example, consumers who buy a blazer online will receive an email with content providing style advice for that item, or showing additional products that complement their purchase.
Ahead of the holiday season, brands can leverage storytelling efforts to improve the gift-giving process as well.
“If you’re trying to personalize in a traditional sense all of your stuff is going to break down,” Bacharach said. “All the data you have on me doesn’t apply when I’m trying to buy for my husband or trying to buy for my brother-in-law who I drew in our Secret Santa. That’s where thinking about context is more important. If there’s a product that sparks an interest, such as a popular model of jewelry that everyone’s getting this year, there might be two or three pieces that can stack with that bracelet that you’re looking at.”