Work / Pulseheaven
Case study
Pulseheaven is a fashion marketplace where people buy and sell clothes and connect with each other. As it grew, I redesigned the whole thing and rebuilt the WordPress setup underneath it. I did both the design and the development.
01The challenge
Pulseheaven kept adding features like selling, profiles, messaging, favorites, and AI try-on. Each one worked on its own, but together they didn't feel like one product anymore.
As the platform expanded, pages picked up different styles, spacing, and components. The site looked less trustworthy than the product actually was.
Shopping, selling, and social actions sat side by side at similar weight, so new visitors had to figure out where to start.
Marketplace, account, and social features came from different plugins, each with its own look and behavior.
Layouts designed desktop-first made important information harder to find on a phone.
Every new feature added weight, and eventually visitors feel that as a slow page or an error.
02Goals and constraints
03Strategy and design direction
A few simple rules guided most of the decisions.
The homepage explains the marketplace first, then splits into clear paths for shoppers and sellers instead of presenting every action at once.
Shared cards, buttons, forms, and surfaces are applied across plugin-driven pages so marketplace, account, and social screens feel like one product.
Each core page was designed for a phone first, deciding what a visitor needs to see before they scroll, then expanded for desktop.
Soft colors, consistent spacing, and clean type, so the product photos get the attention.
If a feature slowed the site down, it had to be worth it. I fixed heavy scripts and repeated database work during the redesign instead of leaving them for later.
04Before and after
Same screen size and the same content wherever possible, so the only thing that changed is the design.
05Under the hood
A site also has to take payments, pay sellers the right amount, and stay fast when it's busy. Here's what that took. Open Technical details if you want the specifics.
Some customers couldn't complete checkout with Apple Pay. I tracked down the cause and fixed it, so Apple Pay just works now.
The failure came from a nonce mismatch buried in the payment flow's JavaScript. The fix corrected how the nonce was issued and passed through the checkout sequence.
Seller earnings have to be right every time. I rewrote the custom shipping plugin so commissions and payouts are calculated correctly and orders aren't processed twice.
The rewrite fixed commission errors, vendor earnings that were being deducted twice, and duplicate webhook processing.
When traffic picked up, the server was running out of room. I found the two requests causing it, cached them, and rebalanced the server so the site stays quick.
PHP worker exhaustion was traced to two slow, uncached REST endpoints. I added caching for both, rebalanced the PHP worker pools, and run a persistent object cache to cut repeated database work.
The tax setup was recalculating way more often than it needed to. I rebuilt it to only calculate when it matters, and the tax reporting stays accurate.
The Stripe Tax integration recalculated on every cart update. The rebuild removed redundant recalculation while keeping accurate nexus reporting intact.
Shoppers can see how an item might look on them before they buy. It's built right into the product page.
Buyers and sellers can follow each other, send messages, and save favorites, so it feels more like a community than a catalog. It all matches the rest of the site.
06The finished experience
07Results
Pulseheaven is still growing, so I'm not claiming a jump in sales or sign-ups here. What I can point to is a site that looks consistent on every page, works well on phones, and runs on a faster, more stable setup.
08Reflection
Plugins make it quick to add features, but each one comes with its own look and slows things down a little. The hardest calls were deciding when to restyle a plugin, when to rebuild a piece of it, and when a feature wasn't worth what it cost in speed.
Design problems and technical problems usually show up together. A confusing page and a slow page both make people leave, and fixing one usually means touching the other.
Now that the new version is live, I want to watch how real shoppers use the product and listing pages and see where they still get stuck.
But if your website looks outdated, feels difficult on mobile, or needs more than a template can do, I can help.