When we started reviewing photo optimization for an ecommerce site
The spark came from a practical question: why were some product pages struggling to stand out compared to category pages, despite solid content? We at Uneven Lab put photo optimization back at the center, because image behavior—affecting loading and perception—can flip organic performance.
We proposed treating images as a strategic element, not a finishing detail. The hypothesis was clear: a weighed-down rendering flow and inconsistent management of formats and visual variants were generating measurable slowdown and weak signals for snippets, with ripple effects on queries with more transactional intent.
How we identified the images that were slowing organic visibility and what surprised us
We started by comparing templates and markets: the same photos were being served with different dimensions, and in some views the thumbnail was actually a high-resolution resource resized via CSS. A simple clue, but decisive in explaining the drop in perceived speed in crucial page areas.
Weight analysis revealed another friction point: a preloaded zoom preview above the fold, plus a carousel with duplicate color variants. Here, photo optimization wasn’t just compression: it was eliminating superfluous requests and synchronizing loading order with what the user sees first.
Finally, textual signals: generic filenames and repeated alt text that didn’t help clarify the difference between similar variants. We weren’t looking for keyword stuffing, but a descriptive consistency capable of supporting search intent, especially where competition is based on attributes like material, finish, and use.
Small technical adjustments and content choices: the process that gave us back speed and clarity
We realigned breakpoints and actual dimensions, avoiding serving more pixels than necessary in initial views. Non-critical resources were moved outside the above the fold space and preloading only concerned the primary image, reducing friction on the actual rendering path.
On visual content we introduced a cleaner direction: consistent framing, crops that don’t alter the perception of proportions, priority to photos that answer the user’s first question. Photo optimization here means favoring readability, not spectacle, with a sequence that clarifies use, detail, and context.
In parallel, we made the textual component more useful: meaningful filenames and alt text designed to distinguish variants without unnecessary repetition. The result was a tangible improvement in perceived speed and a strengthening of relevance signals on queries where the product decides the zero or one of conversion.
The method behind the images: data, UX, and collaboration with the product team
Working on images, we combined three perspectives: what we measure, what the user actually sees, and what’s sustainable for the product team. In this framework, photo optimization becomes a choice of priorities: fewer assets, better orchestrated, with a clear impact on experience and SERPs.
Our method starts from a map of touchpoints: where the image supports intent, where it risks blocking loading, where it can tell a distinctive attribute. In this balance, sufficient quality is preferable to maximum quality that isn’t perceived and weighs more than the benefit it brings.
From numbers to visual decisions — how we balanced quality, SEO, and usability
We translated quantitative signals into practical decisions: the primary photo must be fast and readable, secondary ones enter only when needed. Format is chosen for use context, not habit, and every byte saved is reinvested in visual clarity or a truly useful variant.
For the semantic part we preferred labels and alt text that distinguish variants as a store consultant would: few words, but clear information. This way the page speaks the language of searches, without sacrificing aesthetics and visual consistency across devices and network conditions.
Collaboration with the product team established operational boundaries: what to automate in the pipeline and where editorial review is necessary. This combination gave us a robust baseline and fine-tuning spaces, keeping quality stable and reducing gaps between design, SEO, and user perception.
FAQ on photo optimization and how we work with clients
How much time do we need to evaluate and work on photo optimization for an ecommerce catalog?
It depends on scope, template variety, and asset status. We start from a representative sample and define impact priorities, aligning objectives and constraints with us at Uneven Lab and stakeholders. The intervention proceeds in consistent blocks, so each release brings measurable value.
What data tools do we use to understand the impact of images on traffic and ranking?
We cross-reference weight and dimension analysis, loading sequence, crawl signals, and query trends. We use comparable samples before/after on similar pages and related markets, to isolate the effect of images from other changes and estimate real contribution.
How do we coordinate changes with product teams and those handling image format?
We define shared guidelines and responsibilities: product governs flows and resources, design handles rendering, we oversee SEO choices and priorities. Image format is decided based on context, compatibility, and quality objectives, with clear and replicable criteria.
