Find The Stage That Is Losing You Orders Before You Redesign Anything

Ecommerce UX is how easily a shopper gets from landing on your store to receiving what they ordered, and it's measured in orders rather than compliments.

If someone has told you your conversion rate is a UX problem, the question worth answering is which part of your shop is losing the money.

I run Mint SEO out of Manchester and I've spent eleven years inside ecommerce funnels, most of them on Shopify. A pattern repeats. An owner reads a list of ten improvements, picks whichever looks cheapest and changes it, and nothing moves.

The problem with that list was its order, and the leak sat somewhere else entirely.

This article does the ordering. It covers what the discipline includes, where losses cluster across the journey, and what to change at each stage.

Short version: find your worst stage from your own analytics. Then work discovery, the product page, checkout and mobile in whatever order your data puts them.

Where The Order Actually Gets Lost, Stage By Stage

Start by finding which stage is yours, because these four failures are rarely the same size in any two stores you look at. A benchmark tells you what normal looks like. Only your own numbers tell you where you sit against it.

Sort The Funnel Before You Touch The Theme

Contentsquare publishes a 2026 Digital Experience Benchmark covering 99 billion sessions across more than 6,000 sites. It puts desktop conversion 74% higher than mobile while mobile carries 69.9% of traffic. For an owner that's the most useful line in the report, because it means your headline conversion rate averages two very different experiences.

Traffic source splits the picture again. In that same research, paid search converts at 2.8% and organic social at 0.7%. If a chunk of your sessions arrive from social, your rate is being dragged down by an audience that was never close to buying.

Splitting this in Shopify analytics takes me about twenty minutes. Segment sessions by device, then by traffic source. Look for the steepest drop between viewing a product, adding to cart and reaching checkout.

Your widest drop is the starting point. We have written the longer version of that diagnosis separately as an ecommerce CRO audit.

Where The Money Goes In Most Storefronts

Across the storefronts I audit, losses cluster in four places, and most of them come from theme defaults nobody chose deliberately.

Stage	What fails	How you spot it
Discovery	Shopper cannot find the product	Heavy exits from homepage and category pages
Product page	Shopper cannot judge or trust the product	Product views healthy, add-to-cart rate is not
Cart and checkout	Costs or friction arrive too late	Adds to cart healthy, completed orders are not
Mobile	Whole journey is harder on a phone	Mobile converts far below desktop on comparable traffic

Work them in the order your own funnel ranks them. What each stage needs, in one line:

• Cut the menu down to the categories people pick, and give the category page a routing job.
• Show enough of the product to judge it, and move the rating beside the price.
• Put the delivery cost and timescale before the cart, and drop compulsory account creation.
• Fix what a thumb has to hit, and remove the scripts that block interaction.

How Shoppers Find Products, And Where Navigation Loses Them

Fix discovery first if exits cluster before anyone reaches a product. A shopper who cannot find the item never enters the funnel you have been staring at, which is why this leak never shows up as a drop in a conversion report.

Cut The Menu To What People Actually Pick

A menu exposing everything makes the shopper do sorting you should have done for them. Baymard Institute's homepage and category navigation benchmark found 60% of sites hand shoppers an undivided category list, and 95% never show them where they sit inside it.

Provenance matters before you act on a number like that. Edward Scott, Research Lead at Baymard Institute, led that research and holds a PhD in technical communication and information design. In the same research his team rated 58% of desktop sites and 67% of mobile sites mediocre to poor for homepage and category navigation.

Given a failure rate that high, the fix we reach for is subtraction. Look at which categories receive clicks, group the rest, and give each level the heading a shopper would use for it.

Give The Category Page A Job

Between menu and product sits a routing decision, and most stores waste it. In that same navigation research, 76% of sites fail to lead shoppers with subcategories and 70% give no direct route to products shown in inspirational photography.

Two real storefronts show the difference. UpCircle Beauty puts a full-width search field directly under the header on a 390px viewport.

Above it a free-shipping threshold sits in the announcement bar, with a star rating and review count immediately below. Over on Wild, all navigation collapses behind a hamburger icon with no visible search field.

Ordering is the point of the comparison, because Wild's branding is stronger than most stores manage. Its discovery layer asks the shopper to open a menu before anything else can happen.

Category structure carries SEO weight too, which we cover in ecommerce category pages.

A Product Page Has Two Jobs, And Most Do One

A product page earns its order when a shopper can judge it from the page itself, which puts ratings inside that judgement, beside the price. In the audits I run, most pages describe a product well and prove very little about it.

What Product Imagery Has To Show

Imagery does the job that picking something up does in a shop. It has to answer scale and fit, how the thing looks in context and what turns up in the box. Nielsen Norman Group's product page report runs to 428 pages and 108 design recommendations drawn from more than 350 tested sites.

Practitioners put it less politely. In a thread on the Shopify subreddit a Shopify Expert posting as fathom53 told an owner asking why nobody was buying: 'Just 1 pic on each SKUs. No other pics of the product or on other models.'

Harsh but hard to argue with. One photograph per product reads as a shop that hasn't finished opening.

Put The Rating Where The Price Is

Social proof only works where a decision gets made. On 6th August 2026 I loaded 45 direct-to-consumer storefronts on a 390×844 iPhone viewport and confirmed 21 stores were running Shopify from the rendered markup. Following the first product link on each of those 21 homepages, I read what the product pages say. Nine of the 21 carried no visible rating or review count at all.

For the alternative, look at UpCircle Beauty's foaming face wash page, where review count and star rating sit beside the price. Reading it top to bottom, you meet the proof and the price in the same glance.

Moving a rating you already collect into the same glance as the price costs nothing to build. That makes it the cheapest trust fix here.

Put The Delivery Answer Before The Cart, Not After It

Answer the delivery question on the page where a shopper is still deciding. Meeting an unexpected cost at checkout is the largest recorded reason a filled cart gets abandoned, which is where your most recoverable money sits.

Say What Delivery Costs While They Are Still Deciding

Baymard Institute's running aggregate of 50 cart abandonment studies puts average abandonment at 70.22%. That same research estimates a large store can gain a 35.26% conversion increase from checkout design. In the same research, extra costs being too high accounts for 40% of abandonments and mandatory account creation for 18%.

Checkout leads most ecommerce funnel work because of what those reasons have in common. A shopper builds an expected total while browsing. A delivery charge revealed at the last step invalidates it exactly when they are asked to pay.

Being unable to work out a total cost up front accounts for another 12% of abandonments in that research. Look at a Nkuku product page and the delivery line sits directly under the add to cart button, at 3 to 5 working days for £5.95. Over on Wild it is £1.99, stated the same way. Neither is springing a cost on a shopper twenty minutes into building a basket.

A rotating announcement bar is the common alternative and it's weaker. On Gymshark the free-shipping threshold shares one bar with the email discount and the student discount, so your delivery answer is on screen part of the time.

Stop Charging People An Account To Buy Something

Compulsory account creation costs a fixed share of shoppers who have already decided to buy. An account is a cost your shopper pays for a benefit you receive, demanded at the point of highest intent.

Returns belong in this conversation too. An unsatisfactory returns policy takes another 13% of abandonments in the same research, which makes returns a checkout-stage decision.

On Shopify the supported route for that work changed this summer. Checkout Components reached general availability for Shopify Plus in the Summer 2026 Editions, which matters because on Plus a visible delivery line is now a configuration choice.

Every plan has that lever. Whether an account is compulsory sits in the Shopify admin under Settings, then Checkout, then customer accounts.

That release closed the old route as well. Shopify's Summer 2026 Editions retired Shopify Scripts in favour of Shopify Functions. Merchants had to migrate their checkout logic before 30th June.

If your store still runs discount or delivery logic through a Script, unsupported code is sitting on your highest-value page.

How Much Traffic Is On A Phone, And How It Converts

Treat mobile as its own piece of work, because it carries more traffic than any other device in the benchmark data and converts the worst. The fixes here are what a thumb can hit and how fast a page answers it.

Fix What The Thumb Has To Hit

Most mobile stores fail on the physical business of hitting the right thing. That same navigation research rated 67% of mobile sites mediocre to poor and found 51% failing to make clickable hit areas clear.

Persistent interface competes for the same screen. On Wild's mobile homepage a sticky 'Choose a product' panel sits over the lower third of a 390×844 viewport, so content behind it gets scrolled past.

Such panels only show up as a problem on the device itself. Load your own store on a real handset, and try buying something using one thumb. Everything you fumble is a fix.

Responsiveness Before Raw Load Time

How fast a page answers a tap matters more than how fast it first painted. Google replaced First Input Delay with Interaction to Next Paint because the old metric measured only the delay before a response began.

The thresholds are specific, and worth knowing exactly. Interaction to Next Paint has to come in under 200 milliseconds, and all three Core Web Vitals have to hold for at least 75% of real visits. The Web Almanac's performance chapter puts 48% of mobile sites past all three, against 56% on desktop.

Chasing a load-time score while taps hang optimises a number your shopper never experiences. Cut the scripts blocking interaction before you compress another image, which is the order we work in on a store-wide site review.

What Ecommerce UX Covers, And Where The Limits Are

Ecommerce UX is the whole path from landing page to returns process, which is wider than most people use the term. It also stops well short of fixing the business underneath it.

What The Term Actually Includes

Every point where a shopper judges whether to trust you and carry on is part of that experience. Navigation and imagery are in it. So are forms and speed.

The parts happening after payment count too, because a bad returns experience decides whether a second order ever lands.

Scoping it properly matters practically. When UX becomes a synonym for the theme, work gets handed to whoever owns the theme.

Delivery messaging, returns policy and product data then never get looked at, which is one reason we treat ecommerce web design as broader than a template.

Where A Redesign Stops Being The Answer

Better interface work can't fix a product nobody wants or a price the market won't pay. It changes how easily a shopper acts on an intention without creating the intention itself.

Knowing which way the tide runs helps too. Desktop conversion declined 7.9% year over year in that Contentsquare benchmark, and rates fell on both devices. A flat year isn't automatically a failure.

That same subreddit thread carries an honest version of this from fathom53: 'You should worry about fixing your site before getting more traffic and customers.'

An inverse also holds. If your store is sound and nobody's turning up, that's a traffic problem, and a redesign is an expensive way to avoid saying so.

Some Of Your Shoppers Are Not People Any More

A rising share of traffic reading your product page is software, and what it reads is your structured product data. That deserves attention now and doesn't yet justify a rebuild.

What An Agent Reads Instead Of Your Layout

Agent-based discovery expanded in Shopify's Summer 2026 Editions. Structured product data can be interpreted by AI systems, so stores need consistent attributes and catalogue structure alongside accurate inventory and metafield definitions.

None of what persuades a person is visible to an agent. That makes product data quality part of the UX.

An agent comparing your product against three others reads attributes and stock state, so an incomplete metafield is its equivalent of a blank product photo.

Do Not Redesign For Robots Yet

Published in March 2026, that same Contentsquare research recorded AI referred traffic growing 632% year on year while converting at only 1.3%. Both halves of that sentence matter, and the second half is the one that usually gets left out.

The proportionate response is therefore a data job. Getting the product data right gives an agent something to act on, and it helps a person comparing you against a competitor in another tab.

Rebuilding a template that already converts for people, to court a channel converting at that rate, gets the order backwards.

Where A Second Opinion On Your Funnel Helps

Deciding your fix order is the ecommerce UX judgement that's hardest to make from inside your own store. You know where everything is, which is exactly what stops you seeing where a first-time shopper gets stuck.

Making that call is what our SEO audit and strategy work is for. I have spoken at the SEO Mastery Summit and I run every retainer personally.

Mint SEO runs on eight disciplines doing one job. Keyword strategy, technical SEO, content and digital PR are four of them.

Core Web Vitals and page speed sit inside technical SEO, so responsiveness work above gets a named owner.

It works when a diagnosis is right. On a Shopify home brand we ran product and collection page optimisation plus buyer intent content across a twelve week seasonal window. That produced 256% more sales from Google Search.

If you'd rather talk it through than read another list, our free 30-minute consultation is a call directly with me. We cover your store's current rankings, its biggest quick wins and a rough 90-day direction. Book your free consultation and bring your analytics.