About.
For over 25 years, Backcountry has been the go-to retailer for outdoor enthusiasts seeking premium gear and apparel. From mountaineers to trail runners, the brand built loyalty through a curated mix of top outdoor labels.
With over 200,000 products across categories like bike, climb, ski, camping, and trail running, Backcountry's catalog is a major competitive advantage. It's also a personalization problem at a scale no team could solve manually. To realize conversion gains, Backcountry needed a way to learn from every shopper interaction and act on it across the full assortment.
The challenge: experimentation at scale.
One of Backcountry's biggest strengths, its massive catalog, was also a source of unlocked potential. Personalizing experiences at this scale is impossible to do by hand, and the existing system was the bottleneck:
- Slow experimentation. Most tests required engineering lift, which throttled time to value and capped the volume of experiments the team could run.
- Generic landing experiences. Backcountry could test some experiences, but the majority of paid traffic landed on the same generic page no matter the ad or shopper. Paid media spend was deleveraged because conversion rates could be higher.
- Merchandising at scale. The existing framework offered strong consistency but limited the flexibility to tailor layouts by category, price point, or shopper intent. Expanding dynamic options for bundles and complementary gear would better align with how customers actually shop. Doing it across 200,000 products was a non-starter without automation.
A full site rebuild would have introduced unnecessary risk and long timelines. Backcountry needed AI agents that could iterate continuously across the catalog while preserving the strengths of the existing foundation.
How FERMÀT worked alongside the team.
FERMÀT is built on Anthropic's Claude AI models, powering the personalization and experimentation behind these results at scale. FERMÀT's AI agents embedded with Backcountry's marketing and merchandising teams as an experimentation and merchandising layer on top of the existing ecommerce stack. This is what AI native commerce looks like across a 200,000-product catalog: agents that learn from shopper behavior, run tests across every category and touchpoint, and scale what wins, with digital product teams in control.
The transformation was also present in the operating model. Digital product teams stopped queueing engineering tickets and started directing strategy. The agents handled the volume: page variations tailored by product type spun up automatically from catalog data, experiments running continuously across categories, layout adjustments shipped without engineering. Each test fed the next. Insights compounded across categories. The team moved from periodic experimentation to continuous iteration, at a scale no manual workflow could match.
That gave Backcountry:
- Direct control across teams
- AI-generated page variations tailored by product type, intent, and traffic source
- Continuous testing velocity across thousands of SKUs in parallel
- Behavioral learning that compounded across categories
Experimentation became an always-on capability instead of a one-off initiative.
"At our scale, growth doesn't come from plug-and-play tools; it comes from innovation partners willing to build alongside you. FERMÀT has been exactly that for Backcountry, helping us create AI-driven and dynamic landing experiences for Google Shopping that reflect how our customers actually browse and buy across a massive catalog." - Kevin Lenau, President, Backcountry
Gathering shopper behavior data and testing surfaced two distinct personas:
- Pro shoppers looking for fast access to specs and sizing nuance
- Casual shoppers needing simple explanations and helpful product context
Designing experiences for each behavior helped Backcountry match the page to both the product and the shopper. FERMÀT didn't just fix a page template. It changed how Backcountry worked. Backcountry described FERMÀT as an innovation partner, not a plug-and-play tool.
Results.
1,000+ SKUs with increased add-to-cart rates. Continuous experimentation across five key categories. Marketing and merchandising in direct control of the experience layer.
