Jenson USA: performance, personalization and AI shopping on a large catalog.

The brief

Jenson USA runs a high-traffic store with a catalog of more than 30,000 products. It has to stay fast and stable through seasonal peaks while the marketing and merchandising teams keep testing, personalizing and launching campaigns. We worked as an embedded engineering team across every layer: the edge, the platform and the experience on top.

Multi-CDN delivery

We architected and operated a multi-CDN platform using Cloudflare, Amazon CloudFront and Azure’s edge network, with Amazon S3 for asset storage. Spreading delivery across providers gives a high-traffic store redundancy, fault tolerance and more room during peaks.

Core Web Vitals

Performance work covered Largest Contentful Paint, Cumulative Layout Shift and Interaction to Next Paint across product, listing and landing templates. The methods were script deferral, a critical CSS strategy, lazy loading, CDN tuning, markup refactoring, asset optimization and performance profiling.

Experimentation and personalization

We wrote JavaScript Monetate actions that inject UI components, swap promotional banners, change recommendation carousels and adjust checkout messaging by customer segment and real-time behavior. Every change loads after the critical rendering path, so experiments do not cause layout shift, and each one has a clean rollback. Custom analytics events feed results back to the merchandising team.

Alongside Monetate, we implemented Bloomreach and Exponea for search, recommendations and marketing automation: variant rendering, behavioral tracking, content targeting and dynamic content tied to campaigns.

AI-assisted shopping

Two AI tools are integrated into the storefront. Preezie adds conversational guided selling and product discovery. BikeMatrix checks compatibility and fitment across mountain bike parts, so a rider can confirm a component fits their bike. We built the front-end event hooks, analytics instrumentation and API integrations that connect both to the personalization stack.

Analytics, privacy and accessibility

• GA4 and custom events instrumented across the storefront.
• GDPR and CCPA handling built into how tags load.
• Semantic HTML, ARIA roles and keyboard navigation to WCAG standards.
• Figma design systems translated into reusable components and SCSS tokens.

Where this shows up in Creartful AI

The same habits run through our plans: Core Web Vitals checked before every launch, analytics you can trust before anyone reports on it, and experiments that never slow the page. If you want testing and personalization on your own store, see the Experimentation Setup package or the Summit and Expedition plans.