Balancing E-Commerce Personalization with Web Performance

We live in an era in which e-commerce and shopping online are integral to the “experience economy”. What this means is that shoppers shop in new ways, and online shopping is completely integrated in, and often driving, that experience. What makes e-shopping experiential, as compared to the highly experiential real-life, physical-world shopping experience, is the growth in personalized content and services, tailored individually for online shoppers. The personalized online shopping experience is expected by users, and retailers likewise aim to improve the user experience and fatten their bottom line by making personalized content lead to more sales, more engagement and greater customer loyalty. The better the algorithms work to tailor content for individual users, the better these results are.

And yes, the analysis and immediate application of user data creates these personalized experiences, which do work. Most e-commerce companies wouldn’t dream of skipping personalization. They recognize, underscored by data, such as survey results from Accenture, that for example, almost half of a site’s visitors will leave a site if it is poorly curated, and that 91 percent of consumers expect that brands will recognize and remember their preferences and serve them targeted, relevant recommendations and offers. 

The performance-personalization “cache 22”

Recognizing the power (and necessity) of personalization, internet retailers know it is essential, but they may not recognize the most damaging downside: personalization creates performance lags. Poor or slow website performance is one of the top reasons a consumer will leave a site (i.e. in as little as a quarter of a second), so finding the balance between tailored, personalized experiences and speed is the biggest challenge. 

Cashing in on the cache

Most websites cache their content, which is ideal for static or semi-dynamic or event-driven content. That is, content that never changes or only changes somewhat or sometimes. These kinds of caching are easily managed, and by caching, or storing cached copies of pages and elements, web page loading speed is sped up significantly, and a site protects its origin server from overload (avoiding downtime). A web cache is capable of serving tens of thousands of consecutive requests per second without missing a beat, ensuring smooth and fast end-user experiences. All good. Caching is an easy means to achieve high-performance content delivery.

But what happens when the content is not static or semi-dynamic and is truly individual?

Caching the truly uncacheable: Caching truly dynamic content

Until now, personalized content, ranging from tailored content or product recommendations to session IDs that allow users to add items to shopping carts - and more, have been completely uncacheable because these personalized items change from individual to individual. Because all of this individual content is unique to one user, it could not be cached/reused, meaning that the user experience would inevitably slow down because the content would have to be fetched and loaded each time the customer made a request. 

Over time, solutions have been introduced to try to address the performance problem, but most have required programming changes to underlying e-commerce platforms (potentially complicated and expensive).

Now, with Varnish E-Commerce Accelerator, the approach has changed: personalized content can be cached without making changes to the e-commerce platform, whether it’s well-known Magento or another e-commerce platform. We’ve written at some length about how this e-commerce acceleration works and how it works specifically for Magento. But essentially, personalized content is broken down into parts that can be reassembled where and as needed, so user information, session IDs, individual security protections and so on, can be cached and reused as users navigate across a site… and the personalization is retained along with performance.