What if you could double your conversions and significantly reduce exit rates with just one optimization strategy? That’s exactly what Ray-Ban achieved: a 101% increase in mobile conversions, a 156% boost on desktop, and a 13% drop in exit rates.

The secret? Speculative preloading powered by the Speculation Rules API. By predicting users' next clicks and preloading pages ahead of time, it delivers faster transitions, smoother browsing, and friction-free shopping experiences.

We’re thrilled to see Google publishing this case study, helping to highlight speculative preloading as a major breakthrough in web performance optimization. However, implementing this strategy isn’t as simple as copying Ray-Ban’s approach. Every site has its own structure and user flows, which means speculative preloading needs to be customized for optimal results.

In this case study, you'll discover:

• Ray-Ban’s tailored strategy and achievements

• The key factors behind speculative preloading success

• How Navigation AI provides scalable, data-driven solutions tailored to your business needs

Ray-Ban’s results: Breaking through the performance plateau

Ray-Ban has been shaping eyewear trends for generations. But in today’s competitive market, speed and experience matter just as much as product design. Their ecommerce platform serves 6.6 million unique visitors monthly, making performance improvements a top priority.

Despite a recent redesign that improved performance, delays persisted due to their multi-page architecture (MPA). Each time a user clicked a link, the browser sent a full request to the server, disrupting key moments like browsing Product Detail Pages (PDPs) or completing purchases.

To overcome these challenges, Ray-Ban implemented a custom speculative preloading strategy, resulting in impressive improvements across key metrics:

| Metric               | Mobile               | Desktop               |
|--------------------------|----------------------------|---------------------------|
| Conversion Rate       | +101%                      | +156%                      |
| Pages Viewed per Session | +52%                       | +65%                       |
| Exit Rate Reduction  | -13%                       | -13%                       |
| Largest Contentful Paint (LCP) | 4.69s → 2.66s (43% faster) | 3.03s → 1.74s (43% faster)

These improvements transformed both user experience and business outcomes:

• Higher conversions: Faster load times reduced friction, resulting in smoother purchase flows and significant conversion gains.

• Improved engagement: Visitors explored more of the site, viewing 52% more pages per session on mobile and 65% more on desktop.

• Lower Exit rates: Optimized page transitions reduced drop-offs by 13% across both platforms.

• Enhanced Core Web Vitals: LCP improved by 43%, accelerating page load speeds and positively impacting SEO.

Speculative Preloading: the key to Ray-Ban’s success

To achieve these impressive results, Ray-Ban adopted speculative preloading.

Speculative preloading works by preloading content that users are likely to visit next, drastically reducing delays.

But setting this up manually isn’t simple. It requires an in-depth understanding of user behavior, page flows, and continuous optimization.

Ray-Ban’s tailored strategy

To get the most out of speculative preloading, Ray-Ban designed custom approaches for desktop and mobile users:

• Desktop: Prerendering was triggered when users hovered over product tiles on the Product Listing Page (PLP).

• Mobile: Immediate prerendering focused on the first four frequently clicked tiles to accommodate mobile’s lack of hover events.

Additionally, Ray-Ban optimized the back/forward cache (bfcache) by disabling blocking events and managing background connections, achieving a 73% cache hit rate. This led to further performance boosts: 30% faster LCP and an 83% improvement in CLS.

Key insights: What makes speculative preloading successful?

Ray-Ban’s strategy highlights several critical factors that determine the success of speculative preloading. Here’s how these insights apply to your business.

Customization is essential

One size doesn’t fit all when it comes to preloading strategies. The key question every business faces is: Which pages should be preloaded?

Browsers impose limits on how many links can be preloaded at once, which means you need to prioritize pages that are most likely to be visited next. However, this isn’t always easy, as user paths and behavior patterns can vary greatly not only across websites but also across device types.

In Ray-Ban’s case, their PLP-to-PDP journey (Product Listing Page to Product Detail Page) was a well-defined and highly trafficked path. Preloading decisions were straightforward: product exploration was a priority.

However, not all websites have such clear and predictable paths. For content-rich platforms, complex sites with many conversion paths, or marketplaces, user journeys can vary significantly by use case and device type. Visitors may browse differently on a desktop, where larger screen space allows for more exploratory behavior, versus on mobile, where navigation tends to be more focused but constrained by limited screen real estate.

Getting this prioritization wrong can result in wasted resources, with preloaded pages going unused while critical pages remain slow to load. Trial and error or static rules often lead to inefficiencies.

How Navigation AI helps:

With AI-driven automation, Navigation AI continuously adapts to your site’s changes. As new content or traffic patterns emerge, our solution adjusts preloading strategies in real-time. This ensures your optimizations remain effective without the need for constant rule maintenance. Businesses can scale performance improvements without dedicating large development resources to manual updates.

Long-term success requires constant tweaks

Manual preloading strategies may work temporarily, but they fall apart over time as sites grow and change.

Your site isn’t static: new pages are added, product catalogs expand, promotions come and go, and user behavior shifts with each update. Without automation, these changes require frequent manual intervention to keep preloading rules accurate and effective. Over time, this becomes a resource-intensive process that is difficult to sustain—particularly for businesses without a large technical team.

Ray-Ban tackled this by defining specific rules that targeted high-impact areas. However, maintaining these rules over time likely requires continuous adjustments as they launch new campaigns or update their content.

How Navigation AI helps:

With AI-driven automation, Navigation AI continuously adapts to your site’s changes. As new content or traffic patterns emerge, our solution adjusts preloading strategies in real-time. This ensures your optimizations remain effective without the need for constant rule maintenance. Businesses can scale performance improvements without dedicating large development resources to manual updates.

Evolving user behavior means static rules can’t keep up

One of the biggest pitfalls of manual optimization is its reliance on static assumptions about user flows. Even if you define rules based on today’s traffic patterns, users may behave differently tomorrow.

For example, seasonal changes, new products, or content updates can all shift user priorities. A product page that was crucial during one month may receive far less traffic the next. If your preloading strategy doesn’t account for these fluctuations, you’ll either preload the wrong content or fail to preload pages that have become high-priority.

How Navigation AI helps:

Our technology constantly monitors user activity to detect emerging trends. We retrain our models in real-time to adapt to shifts in behavior, ensuring that preloading decisions always align with how users are interacting with your site. This adaptability helps businesses stay ahead of performance bottlenecks as traffic patterns evolve.

Simplifying optimization with Navigation AI

Speculative preloading can revolutionize website performance, but as Ray-Ban’s experience shows, manual implementation comes with steep resources. This is where Navigation AI changes the game.

Navigation AI is an AI-powered solution that optimizes navigation by predicting and preloading the next pages users are likely to visit. By analyzing user behavior, it reduces wait times, boosts engagement, and increases conversions, all with minimal effort on your part.

Our solution makes speculative preloading accessible and scalable by automating the complex parts of optimization. Here’s how:

1. AI-Driven tailoring

Navigation AI dynamically creates and manages preloading rules based on both historical and real-time user behavior. You don’t have to worry about configuring eagerness settings, triggers, or resource balancing: we handle it for you, adapting to your site’s unique user flows.

2. Continuous Adaptation

Unlike static manual rules that quickly become outdated, Navigation AI continuously learns and evolves. It updates preloading rules to reflect new pages, design changes, and shifting traffic patterns, ensuring your site is always optimized.

3. No big development team needed

Implementing and maintaining manual preloading often requires a team of developers and performance specialists. With Navigation AI, you don’t need that. Our technology adapts in real-time, saving you the headache of continuous configuration and freeing your team to focus on other priorities.

Why this matters for your business

Ray-Ban’s success validates speculative preloading as a powerful performance strategy. But without automation and data-driven insights, many businesses struggle to maintain effective implementations. Navigation AI provides the solution by making preloading scalable, adaptable, and tailored to your unique site structure.

Ready to unlock seamless navigation and enhanced engagement for your users? Contact us today to learn how Navigation AI can transform your site’s performance.