The AI Commerce Revolution
How personalization will replace navigation.
Navigation exists to solve a problem: a store has more products than any one customer wants to see, so we build menus, filters, and categories to help them get to the right slice of it. Every navigation system, no matter how well designed, is fundamentally a compromise — a single structure trying to serve every visitor’s different needs at once.
Ecommerce personalization doesn’t improve that compromise. It removes the need for it. Instead of one navigation system trying to work for everyone, the store itself rearranges around each visitor — surfacing what’s relevant to them specifically, before they’ve had to click a single filter. This is what makes the brand ecosystem from the last chapter actually functional at scale: personalization is the mechanism that lets an ecosystem respond to each person inside it. The best navigation, it turns out, might be the one a customer never has to use.
The Old Model: One Navigation, No Ecommerce Personalization
Traditional site navigation is built once and shown to everyone. It reflects the merchandiser’s best guess at how most people think about the catalog — organized by category, maybe by use case if the brand is more advanced — but it’s static. The 22-year-old training for their first 5K and the 45-year-old managing a chronic hydration issue see the exact same menu, the exact same homepage, the exact same “featured products” carousel.
That’s not a failure of design. It’s a structural limit of the model. A single navigation system, no matter how thoughtfully built, is optimized for an average visitor who doesn’t actually exist. Every real visitor is some distance away from that average, and the gap between “what the navigation assumes about you” and “what’s actually true about you” is friction the customer has to work through on their own.
The Shift: Ecommerce Personalization Replaces Navigation
Personalization at its best doesn’t just mean “recommended for you” widgets bolted onto an otherwise static site. It means the structure of the experience itself changes based on who’s looking at it — what’s featured, what’s prioritized, what path is suggested — without the customer having to declare their preferences through filters first.
Field Note — WHOOP
WHOOP’s site rebuild is a strong example of this shift in practice. Rather than a single, generic path through the catalog, the team built the experience around distinct audience segments — elite athletes, everyday performance-focused users, people newer to structured training — and mapped genuinely different journeys and content for each one.
They didn’t just adjust the messaging; they used mechanisms like progressive disclosure, revealing more depth as a visitor engaged further, so a newcomer and a serious athlete could both land on the same brand and get a meaningfully different, appropriately paced experience.
That’s the real shift ecommerce personalization enables at scale: instead of asking every visitor to navigate a single, static structure built for an imagined average customer, the structure adapts to who’s actually there. The customer doesn’t have to work harder to signal what they need. The store gets better at inferring it.
"The end state isn't a smarter filter. It's a store that stops asking the customer to describe themselves and starts recognizing them instead."
Why This Matters More Than It Looks Like It Does
It’s tempting to file this under “nice to have” — better UX, marginally higher conversion. That undersells it. According to McKinsey’s research on personalization, it can reduce customer acquisition costs by as much as 50% while lifting revenue by 5 to 15% — the kind of number that changes a P&L, not just a UX audit.
This also changes the relationship between merchandising and the customer. In a static model, the merchandiser has to guess who’s visiting and build for that guess. In a personalized model, the store learns who’s visiting and adjusts in response — which means the brand’s understanding of its customers gets more accurate over time, not just more elaborate.
There’s a trust dimension here too, and it cuts both ways. Done well, personalization feels like being understood. Done poorly or too aggressively, it feels invasive or, worse, wrong — recommending things that miss the mark in a way that erodes confidence rather than building it. That bar matters more in health and wellness than in almost any other category, because the stakes of a wrong recommendation feel higher.
What To Do Differently
This isn’t a call to rebuild your entire site architecture around AI overnight. It’s a call to start wherever the highest-friction generic moment currently lives — usually the homepage, or the first product category a new visitor lands on — and ask what a genuinely different experience would look like for two or three distinct types of customers you actually know you have.
For endurance and wellness brands, the segments are often obvious even without sophisticated modeling: the beginner versus the experienced athlete, the person managing a specific health condition versus the person optimizing general performance. Each of those groups is, in a meaningful sense, shopping at a different store even when they’re looking at the same catalog.
