Ecommerce Merchandising: What Drives Revenue vs. What Fills Pages

Ecommerce merchandising is the practice of deciding which products to show, where to show them, and how to present them to maximise conversion and revenue. Done well, it is one of the highest-leverage activities in ecommerce. Done poorly, it is just a product catalogue with a checkout bolted on.

Most ecommerce teams underinvest in merchandising relative to paid media. They spend heavily on traffic acquisition and then leave money on the table through weak product presentation, poor category logic, and search results that surface the wrong items. The gap between a merchandised site and an unoptimised one is rarely visible in a single session. It compounds over months.

Ecommerce merchandising sits at the intersection of commercial strategy, customer psychology, and data. It is not a design exercise or a content exercise. It is a revenue exercise. If you want to understand how it fits into a broader growth framework, the Go-To-Market and Growth Strategy hub covers the wider commercial context that merchandising decisions need to operate within.

Why Most Ecommerce Teams Underestimate Merchandising

There is a pattern I have seen repeatedly across ecommerce clients. The marketing team is focused on acquisition. The technology team is focused on platform stability. The trading team is focused on margin and stock. And merchandising, which touches all three, ends up owned by no one in particular.

When I was running agency teams across retail and ecommerce accounts, we would often be brought in to improve conversion rates. The brief was usually framed around CRO: button colours, checkout flows, page speed. Occasionally we would dig into the merchandising layer and find that category pages were sorted by margin rather than demand, that search results were surfacing out-of-stock products in the top three positions, and that the homepage was promoting last season’s lines because no one had updated the rules. These were not technology failures. They were merchandising failures dressed up as conversion problems.

The reason this happens is partly structural and partly attitudinal. Merchandising requires cross-functional ownership. It sits between trading, marketing, and product. When no single team owns it clearly, it defaults to whatever the platform does out of the box, which is almost never optimised for revenue.

What Ecommerce Merchandising Actually Covers

Merchandising in ecommerce spans four main areas. Each one has a direct line to revenue, and each one is more nuanced than it first appears.

Category and Navigation Architecture

How you organise products into categories shapes how customers browse. More importantly, it shapes how search engines index your catalogue. Category architecture is a commercial decision, not a taxonomy exercise. The question is not “what are the logical groupings?” It is “how do customers think about these products when they are ready to buy?”

A customer searching for “running shoes for wide feet” does not think in the same hierarchy as a buying team organising by brand. If your category structure reflects internal logic rather than customer intent, you will lose traffic at the top of the funnel and lose browsers mid-session. Both are expensive problems.

Product Sequencing and Sort Logic

The default sort order on most ecommerce platforms is either “newest first” or “best sellers.” Neither is necessarily wrong, but neither is necessarily right. The question is what you are optimising for: revenue per session, margin, stock clearance, or new product discovery.

I have seen trading teams override sort logic manually to clear aged stock, which makes sense from an inventory perspective but often damages conversion because customers see products they do not want before products they do. The smarter approach is to build rules that account for multiple signals simultaneously: demand score, margin, availability, and recency. Most platforms support this. Few teams configure it properly.

On-Site Search

On-site search is one of the most consistently under-optimised revenue channels in ecommerce. Customers who use search are further along in their decision process than customers who browse. They know roughly what they want. The job of search merchandising is to give them the best possible match quickly, and to surface related or complementary products where relevant.

The common failure modes are: zero-results pages for queries that should return results, poor synonym handling so “sofa” and “couch” return different sets, and relevance algorithms that prioritise margin over match quality. Any one of these will suppress conversion from search users, who are often your highest-intent visitors.

There is a useful analogy here. Someone who walks into a clothes shop and tries something on is far more likely to buy than someone who just browses the rail. On-site search users are the equivalent of the customer who has already picked something up. The merchandising job is not to get in their way.

Product Detail Pages

The product detail page is where the conversion decision is made. Merchandising at this level covers the quality and completeness of product information, the logic behind “you may also like” and cross-sell modules, and the sequencing of images and variants.

Image sequencing matters more than most teams realise. The first image is the one that appears in category listings and search results. If it does not communicate the product clearly in a small format, you lose clicks before the customer even reaches the detail page. The subsequent images should answer the questions a customer would have if they were holding the product in a physical store: what does it look like from the side, how does it fit, what is the scale.

The Data Layer That Merchandising Depends On

Good merchandising decisions require good data. Not perfect data, but data that is current, consistent, and connected. The three data sources that matter most are: sales velocity by product and category, on-site behavioural data showing where customers drop off or change direction, and stock availability by variant.

The mistake I see consistently is teams investing in sophisticated merchandising tooling before they have clean underlying data. The tooling cannot fix bad inputs. A personalisation engine that recommends products based on stale purchase data or ignores stock availability will actively harm the customer experience.

Behavioural analytics tools can give you a reasonable picture of where customers are losing confidence or changing their minds mid-session. Hotjar and similar tools are useful for session-level insight, but they are a perspective on behaviour, not a definitive explanation of it. Use them to generate hypotheses, not to confirm them.

The more useful data discipline is connecting your merchandising decisions to actual revenue outcomes. If you change the sort order on a category page, you need to be able to measure what happened to revenue per session from that category. If you cannot measure it, you are operating on instinct. Instinct is not useless, but it is not a strategy.

Personalisation: What It Can and Cannot Do

Personalisation in ecommerce merchandising is real and valuable. It is also frequently oversold by platform vendors and misapplied by teams who have not done the foundational work first.

The foundational work is this: clean product data, consistent category taxonomy, and a reliable behavioural signal. Without these three things, personalisation algorithms have nothing useful to work with. They will surface irrelevant recommendations, which is worse than surfacing no recommendations at all because it erodes trust in the browsing experience.

Where personalisation genuinely adds value is in returning customer journeys. A customer who has bought from you twice has a signal history that can meaningfully inform what you surface next. A first-time visitor does not. For new visitors, good category merchandising and strong search results will outperform any personalisation layer because there is no signal to personalise from.

I have sat in enough vendor presentations to know that personalisation is often positioned as a solution to a conversion problem that is actually a merchandising fundamentals problem. Before you invest in a personalisation platform, audit your category logic, your search configuration, and your product data quality. In most cases, fixing those three things will move the needle more than any algorithmic layer on top of them.

Seasonal and Promotional Merchandising

Seasonal merchandising is where the gap between planned and reactive teams becomes most visible. A planned team has its promotional calendar mapped to its merchandising rules three to four weeks in advance. Category pages are pre-configured for the promotional period. Search synonyms are updated to reflect seasonal language. Homepage slots are scheduled, not manually updated on the day.

A reactive team is updating category pages on the morning of a sale event, manually pinning promoted products because the rules were not set up in advance, and discovering at 9am that the search results for the sale’s hero category are still showing full-price items.

I have seen both. The reactive version is not just stressful. It is commercially expensive. Peak trading periods are when your traffic is highest and your conversion opportunity is greatest. If your merchandising is not ready when the traffic arrives, you are leaving revenue on the table at exactly the moment it matters most.

Working with creator-led campaigns around key seasonal moments adds another layer of complexity. When a campaign drives a spike in traffic around a specific product or category, the merchandising layer needs to be ready to convert that interest. A well-executed creator campaign that drives traffic to a poorly merchandised category page is a waste of media spend. Later’s work on creator-led holiday campaigns touches on this alignment problem between campaign execution and on-site readiness.

The Relationship Between Merchandising and Paid Media

Earlier in my career I overvalued lower-funnel performance channels. I thought paid search was creating demand. It was mostly capturing it. The distinction matters because it changes how you allocate budget and how you measure success.

Merchandising sits in a similar conceptual space. Performance marketing drives customers to the door. Merchandising determines what happens once they are inside. If you are spending heavily on paid media to drive category traffic and your category pages are poorly sequenced, your ROAS figures are telling you about your ad creative and bidding strategy, not about the commercial efficiency of the full customer experience.

The teams that get this right treat merchandising as a multiplier on their media spend. They ask: if we improve the conversion rate of this category page by 15%, what does that do to the effective cost per acquisition from paid traffic to that category? The answer is usually significant enough to justify serious investment in merchandising quality.

This is also why go-to-market execution feels harder than it used to. Traffic is more expensive. Competition is higher. The margin for error on what happens after the click is smaller. Merchandising is one of the few levers that improves the economics of acquisition without requiring you to spend more on media.

Measuring Merchandising Performance

Merchandising performance is measurable, but it requires you to set up the measurement framework before you make changes, not after. The metrics that matter most are: revenue per session by category, search conversion rate, zero-results rate for on-site search, and average order value by entry point.

Revenue per session by category is the most direct measure of merchandising quality. It captures both the conversion rate and the basket size, which gives you a fuller picture than conversion rate alone. A category with a high conversion rate but a low average order value may be well-merchandised for entry-level products but missing cross-sell and upsell opportunities.

Zero-results rate for on-site search is a leading indicator of missed revenue. Every zero-results page is a customer who arrived with intent and left without finding what they wanted. In most cases, the product exists in the catalogue. The problem is a synonym gap, a spelling variation, or a product that has not been tagged correctly. These are fixable problems with measurable revenue impact.

Growth strategy frameworks that treat ecommerce as a pure acquisition challenge often miss this. The growth strategy resources on this site explore how on-site performance connects to the broader commercial picture, including where merchandising fits relative to acquisition and retention investment.

The broader point about measurement is one I hold firmly: analytics tools give you a perspective on reality, not reality itself. Your merchandising data will tell you what customers did. It will not always tell you why. Build a testing discipline that lets you make changes, measure outcomes, and iterate. That is more valuable than any single optimisation.

Where to Start If Your Merchandising Is Underdeveloped

If you are starting from a low base, the priority order is: fix on-site search first, then category sort logic, then product data quality, then cross-sell and upsell logic. This sequence is deliberate. Search users are your highest-intent visitors. Fixing search has the fastest revenue impact. Category sort logic affects every browse session. Product data quality underpins everything else. Cross-sell and upsell logic is valuable but depends on the other three being in reasonable shape first.

The temptation is to start with the most visible element, which is usually the homepage or the hero category page. These matter, but they are also the parts of the site that most teams already pay attention to. The higher-value work is often in the long tail of categories and the search results that most teams have never audited.

Tools that support growth experimentation, like those covered in Semrush’s overview of growth tools, can be useful for building the testing infrastructure that good merchandising optimisation requires. But the tool is secondary to the discipline. A structured testing process with basic tooling will outperform an unstructured process with sophisticated tooling every time.

The teams that make the most progress are the ones that treat merchandising as an ongoing commercial practice rather than a setup task. They review category performance weekly, audit search results monthly, and test sort logic changes against revenue outcomes rather than click-through rates. It is not glamorous work. But it compounds in the same way that poor merchandising compounds against you.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions