Passion for CX Boosts Caleres Revenues With a 21% Increase From AI Search & Discovery Platform Coveo

An early adopter of the web, Caleres witnessed the trends and growth of ecommerce take shape. A few years ago, they identified that their search and discovery technology was not optimized to handle a complex catalog of more than 600,000 SKUs organized by color, size, width, and store availability.

They were unable to adjust, test, or access data regarding on-site search, product listing pages, and product recommendations. Finally, they relied on thousands of manual rules that were time-consuming and difficult for merchandisers to maintain. For the evolution of their next generation ecommerce experience, they chose Coveo’s AI platform to power their search and experience layer.

The Latin word calēre means to glow with passion or intensity — the very principle on which Caleres was founded. Caleres is one of the most successful shoe retailers you’ve never heard of. The $2.8 billion St. Louis-based parent of Famous Footwear traces its history to pre-Civil War, navigating more than 140 years of business by staying one step ahead of an ever-changing industry.

This meant beginning the transition to ecommerce early, even before the sharp shift of consumer habits during the pandemic. In addition to over 900 Famous Footwear stores, Caleres holds a distinguished portfolio of brands sold through wholesale and direct-to-consumer (DTC). These include Allen Edmonds, Blowfish Malibu, BZees, Circus by Sam Edelman, Dr. Scholl’s Shoes, Franco Sarto, LifeStride, Naturalizer, Ryka, Sam Edelman, and Zodiac

“We had a desire to deliver a more personalized, immersive shopping experience sitewide. We wanted to do more ‘storytelling’ with our site, in line with what other leading websites were doing,” states Dan Cornwell, Director of Ecommerce and Digital Experience at Caleres.

“It was important to invest in software and partners that shared our belief in the end customer experience and the empowerment of business users. We are big believers in the power of machine learning, especially with the volume of orders and visits on our websites.

We wanted smarter recommendations and less manual intervention so we could focus more on new functionality for the website, instead of fine-tuning the sort and hierarchy positions of the search results.

Transforming a $2.8B company at scale could not happen overnight so Caleres began a two-year project migrating 13 websites from their old platform to the new technology stack that included Coveo.

Once the first go-live was achieved, they rapidly launched 13 sites in an 18-month period — almost a site a month! (With some pauses for Black Friday and the holiday peaks.)

“The rollout schedule for the re-platformed sites was designed to minimize risk and disruption by starting with smaller, simpler brands and progressively adding more functionality and complexity,” explained Jessica Frame, Caleres’ ecommerce manager. “The Famous Footwear site was the most complex with the biggest volume of SKUs and physical store locations. Local inventory availability was an important element to factor. Rolling out all 13 sites in a relatively short window of time was a major project — considering how large some of the catalogs were — and all of the moving pieces involved.”

Shoppers like choice. And footwear offers plenty. Most women’s shoes, for example, come in 12 different sizes and two widths. Men’s shoes have up to 29 sizes and six widths. There are dozens of categories — shoes, boots, sandals, sneakers, slippers, etc. — and even more subcategories — heels and flats, cowboy boots and snow boots, flip flops and platform sandals, basketball shoes and running shoes, and more. Add brand, color, material, and price and multiply by several hundred thousand products, and the permutations grow exponentially.

“Coveo’s relevance engine is central to the functionality of our ecommerce sites,” explains Cornwell. “Not only for powering easy product discovery through the search box, but powering the product category pages, recommendations throughout the site, the sorting and ranking positions of products, as well as how the filters and facets work, appear, and are able to be used.”

Facets are the navigational tools that present attributes in logical groupings so that shoppers can narrow their results until they get to something they want to buy. With customers wanting Buy Online, Pick Up In Store (BOPIS) service, showing availability by location is now critical.

“It has been powerful that we can leverage machine learning (ML), specifically automatic relevance tuning, to power our search results pages,” states Jessica Frame. “We tested search pipelines (filters) that are fully AI-powered, requiring minimal manual tuning, against search pipelines that have a lot of manual rules associated with them.”

“With Coveo’s machine learning (ML), we’ve seen increases in all of our key metrics. For example, in the first few months of deployment, one of our brand sites saw a 25% improved conversion rate with search. “We’ve also seen improvements when it comes to the use of facets,” she explains. “About 30 to 50% of visitors engage with facets during a session on our site and those who use them convert at a rate that’s almost 2X higher than those who don’t.”

Another tactic Caleres used for their back-to-school special included a combination of merchandiser tailoring and ML capabilities. They created a simple rule for ‘featured product’ boosting to anchor any products that were part of the showcase banner. This strategy was designed to increase engagement by using ML to fill in product results around the boosted items. This kept the results personalized and relevant to each customer, leading to increased product views and purchases.

You don’t survive 140 years by standing still, and this transformation marks just the beginning of Caleres’ next iteration. Future development will be around headless architecture — offering a flexible user interface for multichannel. And around machine learning.

“We’ve recently migrated our product catalog to a new source model offered by Coveo as part of their ‘personalization as you go’ functionality,” explains Fels. “This means more product attributes (color and size) can be shared with the machine learning models to resolve issues related to both ‘cold-start shoppers’ and ‘cold-start products’ (new products with no interaction data).”

Finally, he adds, “we will continue our close collaboration with Coveo’s Customer Success team to test, experiment, and iterate based on the KPIs we achieve. They will inform our product recommendation strategy — especially the types of recommendations — and where they are placed — for maximum impact.”

Rest assured that Caleres’ future will be about intensely meeting the ever-changing customer expectations.