The brand.
Sigma Beauty is not a newcomer asking for a chance. The company behind the F80 Flat Kabuki — the brush that rewired how professional artists think about foundation application — is stocked at Sephora, Ulta, Nordstrom, and Neiman Marcus. It ships to more than 70 countries. Allure has been recognising it editorially for over 15 years. This is a brand that built a serious global following on the back of genuinely superior product engineering.
Which is exactly why a mobile PageSpeed score sitting in the 30s was such a costly mismatch. The product was world-class. The storefront's performance was not.
The problem.
Beauty e-commerce is one of the most image-intensive categories on the web. Every product needs multiple angles, swatch comparisons, model shots, and lifestyle photography. That creative investment is deliberate — it sells. But unmanaged, it compounds into a performance crisis that plays out quietly, one abandoned mobile session at a time.
Sigma's flagship product pages were carrying unoptimised, full-resolution images with no next-gen format strategy. JavaScript bundles had grown without disciplined pruning. Render-blocking resources were delaying the point at which a visitor could actually interact with the page. On mobile — where the majority of discovery traffic lands — the experience was measurably slow by any standard Google or a customer would apply.
The mobile PageSpeed score of 38 was not a surprise given the stack. It was the predictable result of a site that had scaled its visual ambition without scaling its performance discipline. The score itself mattered less than what it represented: slower LCP, higher bounce rates on product pages, and a compounding drag on organic search visibility at the exact point in the funnel where conversion intent peaks.
A PageSpeed score of 38 on mobile is not a minor optimisation problem. On a site shipping to 70+ countries, with a meaningful share of traffic arriving on mid-range devices over variable connections, it is a revenue leak hiding in plain sight.
What we did.
The scope was performance, not redesign. Sigma's visual identity is distinctive and earned — the goal was to make what they already had load significantly faster without changing a single pixel of the experience their team had built. That constraint actually sharpens the work. There is no hiding behind a visual refresh.
The image pipeline was the highest-leverage lever. Product photography is inherently heavy, but the difference between a PNG or unoptimised JPEG and a well-encoded AVIF served to a supporting browser is often 60–80% file size reduction with no perceptible quality loss at typical viewport sizes. Multiply that across a category page with 30 products and the impact is not incremental — it is transformational for load time.
Critical CSS extraction addressed the second major bottleneck. When a browser must download, parse, and apply a full stylesheet before rendering anything visible, every kilobyte of that stylesheet is felt by the user as blank screen time. Inlining only what is needed for the above-the-fold experience removes that wait entirely for the content that matters first.
JavaScript bundle reduction required an honest audit of what was actually being executed on page load versus what had accumulated as dead weight from integrations, A/B testing scripts, and third-party tags that were no longer earning their payload cost. We removed what was not pulling weight and deferred what could wait.
What we did not do.
We did not redesign anything. We did not change the information architecture, move modules, alter the typography, touch the brand colour palette, or suggest that Sigma's creative team had made wrong decisions. The visual experience a visitor sees after our work is identical to the one they saw before — it simply arrives far sooner.
This is a deliberate position. Performance optimisation and visual redesign are different disciplines with different risks. Conflating them slows down the performance work and introduces variables that make it harder to attribute results clearly. We do one thing at a time and we do it completely.
The result.
Mobile PageSpeed moved from 38 to 91. That is not a polish — it is a category change. Below 50, a site is actively slow by modern standards. Above 90, it is competing with purpose-built fast sites. The 53-point gain puts Sigma's storefront in a position where performance is no longer a drag on conversion or search ranking.
LCP improved by 2.1 seconds. Two seconds is a significant share of a mobile visitor's patience. Research across e-commerce consistently shows that load time improvements in this range correlate directly with reduced bounce and higher add-to-cart rates on the pages where the optimisation is most concentrated — in this case, the flagship product pages where the brand's highest-consideration items live.
Mobile bounce rates on product pages declined. The mechanism is straightforward: when a page loads fast enough that the visitor is not kept waiting, more of them stay to see the product. When they stay, the page's existing conversion mechanics — photography, copy, reviews, social proof — get to do their job.
The team described the result as "incredible." We will take that, but the more precise description is: expected. These are the results of disciplined performance engineering applied to a site with a clear bottleneck and a strong underlying product. The work surfaced what was already there.
Delivered in 4 weeks from engagement to launch. No scope creep, no redesign detour, no extended sign-off cycles. The brief was precise, the execution matched it.
PageSpeed Shopify optimisation at this level is not about running a single plugin or enabling image compression in the theme settings. It requires an audit of every resource the browser loads, in the order it loads them, and a deliberate decision about each one. That is the work we did.
What this means for your store.
If your mobile PageSpeed score is below 60 and your store carries significant image inventory — which describes most beauty, fashion, and lifestyle e-commerce sites at scale — the performance gap is costing you sales you cannot see. You are not watching visitors abandon; you are simply never seeing the session that would have converted if the page had loaded two seconds sooner.
The fix is not a redesign. It is not a platform migration. It is a structured performance audit, an image pipeline that serves the right format at the right size to the right device, and a disciplined approach to what the browser loads and when. That is a 4-week engagement, not a 6-month project.
Mobile speed e-commerce is one of the highest-return investments a growing direct-to-consumer brand can make. The infrastructure is already there. The products are already good. The traffic is already arriving. Performance engineering is what closes the gap between what your site could be converting and what it currently is.
Your store deserves a faster engine.
Tell us your current PageSpeed score and what platform you're on. We'll tell you what's blocking you and what a realistic improvement looks like — before any commitment.