It feels like every piece of advice in the ecommerce growth world published in the last 2 years has a liberal wedge of AI optimisation, GEO, AEO and a full roster of shiny new acronyms for good measure.
The truth is that the ‘new and shiny’ rarely moves the needle as much as we’d like to believe – and focussing heavily on AI optimisation by burning precious resources to get Claude and ChatGPT to reference your brand is costing you dearly elsewhere.
In fact, I’m seeing this now more than ever; brands are focussing on AI and neglecting the core components and work streams that consumers (and ironically LLMs) care about.
To shed light on this, I’ve put together a list of the 3 most frequent gaps that ecommerce brands of all sizes are constantly making in 2026.
For context, I’m an ecommerce consultant who specialises in delivering measurable sales and profitability growth for brands selling online; either exclusively or as an accompaniment to their other channels.
From blue chips and start-ups to household names, I’ve been fortunate enough to work closely with hundreds of brands across many verticals and at various stages.
I don’t mention this to gloat. I’m sharing my experience with you to show that, with the rise of the new kid on the block (LLMs and AI), I’m seeing brands make the same mistakes over and over.
What’s even more concerning is that I’m also seeing from a firsthand perspective just how damaging this is becoming, particularly when you add in the fact that ecommerce is a tough space to be in right now. Just a few real-world challenges that now come to mind:
- CPMs and clicks are more costly than ever, eroding margins
- Ad performance is volatile and at times completely unpredictable
- Customers expect more yet have less time, shorter attention spans and less disposable income to spend
- Brand loyalty feels elusive, there’s always someone cheaper
- Fuel and shipping costs are unstable
Here’s an analogy I like to use – forewarning, it’s a weird one…
Cookware / pan technology has been transformed over the years, but no pan in the world replaces good technique, seasoning and the right temperature when cooking a good steak.
It’s the same for ecommerce – you can bolt on all the shiny new AI widgets and inject LLMs where you see fit, but it can’t replace the fundamental pillars of ecommerce and selling online.
So, without further ado, here are the three biggest mistakes I’m seeing ecommerce brands make in 2026:
Poor category structure and hierarchy
Shopify made ecommerce effortless, accessible and hugely approachable. Because an ecommerce site can be set up with so little effort, its foundations have been somewhat forgotten.
Categories are the lifeblood of an ecommerce store, particularly if you’re selling a breadth of items across multiple categories and departments.
If you sell skincare, simply having a face, body, eyes and lips as categories just isn’t satisfactory. It’s a disservice to more than you might think.
Not only will your customers have a hard time navigating and finding what they’re looking for, but your visibility in organic search and LLMs will also suffer substantially.
Your conversion will be throttled, your average order value will be stagnant (which is a real profitability lever).
It goes deeper than that too – your ad performance is throttled as your categories aren’t well thought out so customers face friction when browsing.
Another way to think of categories – supermarkets spend and make millions of pounds each year ensuring their customer journeys are as optimised as possible. Granted, they aren’t always optimised in the way we, as customers, would like, but they’re designed to ensure you spend as much time in-store as possible and spend more of your hard-earned money there too.
The bigger the store, the more intensive that area of focus becomes.
The supermarket aisle is a physical manifestation of your website’s categories.
I’m not just talking about building those categories correctly in the first place, but continually maintaining, pruning and managing those categories too – it’s a living and breathing entity in the context of ecommerce.
It’s important to remove, redirect or delete categories that no longer serve their purpose, ensure stock levels in those respective categories are healthy and that products aren’t misplaced in the incorrect sections.
Going back to the skincare analogy, yes, you need to break down your categories by application (face, lips, neck etc) but also spend significant amounts of energy and time curating those ‘aisles’ in a way that’s truly helpful to your customers.
For example, organise them by the problem areas customers are trying to tackle, brand, skin type, key ingredient and so on; a personal favourite of mine that brands still haven’t adopted well is allergens to exclude.
If there’s one place in ecommerce where AI is criminally underutilised it’s here – optimising and carving taxonomies based on the products sold and customer behaviour, shopping habits and analytics data.
Not in generating waves of low-rent ad creative, not in writing product descriptions, it’s right here.
Relying too heavily on category URLs to drive sales and customers
This feeds nicely into the first mistake above – with the right category architecture and taxonomy, you have the underpinnings to create the right links for landing pages.
Staying with the skincare theme – rather than just sending a heavy amount of paid/email/organic (delete as appropriate) traffic to a ‘Face’ category page, with the right, well-thought-out category structure, you could instead create a Face landing page.
This would allow customers to explore all of the facial skincare on offer at a glance, while also acting as a product wizard to help potential customers filter through to the products that matter to them.
For example, your landing page could have banners, creative and highly targeted product categories for key concerns such as:
- Acne
- Rosacea
- Teenage skin
- Anti-ageing
- Hyperpigmentation
- And so on…
Better yet, you could have landing pages for each of these areas too – so a dedicated page for Acne skincare that further drills down into:
- Acne cleansers
- Acne moisturisers
- Acne toners, serums and accessories
A well-curated landing page for customers with acne will always perform exponentially better than simply placing those customers onto an ‘All Acne Skincare’ product category.
I cannot tell you how many brands are making this mistake and how much cash and customers they’re leaving on the table as a result.
To add – brands of all sizes and stages are making this mistake and it’s such a huge opportunity if you get this right.
Inaccurate tracking that’s unable to paint the right picture
I can’t tell you how astonishing it is to hear variants of this same story over and over across different ecommerce brands that are running ads, whether agency-managed or operated in-house:
“We spend ££££ per month on ads and we feel it’s working but we don’t really know for sure how or where those sales are coming from or if our ad spend is profitable.”
Or something similar:
“Our Meta/Google Ads/TikTok etc. tracking isn’t tracking conversions properly, but we’re still spending significant fistfuls of cash every month on ads.”
Absurd isn’t it – there isn’t another place on the entire marketing spectrum that would be able to consume so much capital without comfortably proving a return.
Both agencies and in-house teams are guilty of this and over the years, I’ve worked with countless PPC teams on both sides of the fence that have muttered these or very similar statements.
The point I want to make here is that server-side tracking for ad platforms and analytics is a must. Server-side tracking is no longer a luxury or a nice-to-have component of your tech stack.
Regular client-side analytics just doesn’t cut it anymore. Revenue and transaction data are frequently dropped, and crucial conversions in ad platforms go unrecorded. Browsers and ad blockers cripple client-side tracking, resulting in significant gaps in data.
You cannot successfully run an ecommerce business if your data collection is compromised.
I’ve moved countless ecommerce brands across to a server-side setup – not only to collect more of the data they’re missing, but also to provide insight into optimisation levers and opportunities.
It allows brands to fully understand whether the cash they’re spending on ads is effective and what they need to address if it isn’t.
My point here is that too many brands neither understand nor trust the data they’re collecting. For me, this is the biggest of the three mistakes ecommerce brands are consistently making – even in 2026!
In summary
I don’t want you to come away from reading this and assume this is an exhaustive list – it isn’t. However, in my experience working firsthand with ecom brands, these are the biggest hurdles that are most often overlooked in favour of AI, LLMs and working to secure brand mentions in these spaces online.
I’m not saying that AI should be dismissed, not at all. I’m simply saying that shifting too many precious resources into those areas and using AI improperly could actually be costing you dearly elsewhere.
And no, before you ask, AI didn’t write a single bastard word 😂
