Performance engineering / international leadership
Faster pages. More time for the customer.
A luxury e-commerce site had a post-click problem: paid campaigns attracted interested shoppers, but slow landing pages kept them waiting. I led a cross-border refactor to change that experience.
Reported reduction in landing-page load time
01 / The signal
Good ads. A slow landing page.
The company was attracting buyers through Meta campaigns for luxury custom cabinetry, sometimes involving $30,000–$40,000 projects. But campaign interest wasn't translating cleanly into engagement after the click.
My responsibilities crossed paid acquisition and technical quality, which let me compare advertising performance with what customers actually encountered on the site. Session recordings and heat maps showed many visitors sitting idle before they could meaningfully interact.
The original analysis recorded passive waiting behavior in 50–80% of observed sessions on key landing pages, with some page loads taking three to six seconds. That pointed to a performance problem worth fixing at the architecture level.
02 / Integrated diagnosis
Look beyond the click.
We investigated load time, PageSpeed results, mobile browsing, and the behavior of visitors arriving through Facebook and Instagram in-app browsers. Campaign data could tell us people were interested; session evidence showed what happened to them next.
The experience after the click mattered as much as the campaign that brought someone there.
With a business case tied to the customer journey, I set a sub-one-second performance target and coordinated a substantial front-end refactor rather than piecemeal changes.
03 / Performance trace
From waiting to browsing.
The original project records show landing-page load time dropping from about three seconds to 0.8 seconds. The illustration uses rounded values from that record, not a live production benchmark.
Faster access to the products customers came to explore.
04 / The refactor
Lead the work.
Let specialists build.
Reduce front-end overhead
The team refined Vue/Nuxt delivery, code splitting, caching, image loading, middleware, and the use of lighter native components.
Align distributed teams
Four developers in Bulgaria worked across a nine-hour time difference. I coordinated requirements, Jira and email documentation, stakeholder decisions, and staging reviews.
Verify the experience
We tested PageSpeed scores, staging environments, mobile devices, and in-app browsers. I maintained final quality sign-off responsibility; the development team implemented the code.
The intensive engineering effort took roughly a month, with additional coordination and validation time.
05 / The results
Performance people could feel.
Critical landing pages went from about 3.0 seconds to 0.8 seconds to load, and Google PageSpeed scores improved from the 70s into the 95+ range in the project record. Average session duration rose from approximately 1.5 to 2.3 minutes, a reported 53% increase.
The changes addressed a technical obstacle in a high-value purchase journey and improved the post-click experience for paid social traffic. They provided a stronger foundation for subsequent marketing and product work.
A customer journey is only as good as the experience that actually loads.
LOAD TIME CHANGE~3.0 → 0.8 SEC
REPORTED REDUCTION70%
SESSION DURATION+53%
PAGESPEED SCORE95+
STATUS: FRONT-END REFACTOR COMPLETEMore selected work
See the work
behind the result.
Explore the decisions and systems behind other projects.