E-commerce has stiff competition and ever-shrinking margins, and most of the sales you lose, you never see leave.
They slip out through a confusing checkout, a product page that doesn't answer the obvious question, a redirect chain nobody's maintained since 2022.
Baymard's research puts the average cart abandonment rate at 70.19%, which means that for every ten people who add to cart, seven leave without buying. Some of that is unavoidable browsing behavior, but a lot of it isn't.
When we audit an online store, the same mistakes come up again and again, and almost none of them are exotic. They're small, fixable, and often very expensive. Here are the ones that cost the most, grouped by where they happen, with the fix for each.
This is where to look first, because it's the closest point to the sale and the easiest place to lose one you'd already won.
The average US checkout asks for nearly 24 form fields when about half that will do. Every extra step, every field that makes someone stop and think, is another exit.
Baymard estimates a better-designed checkout can lift conversions by around 35% which means checkout isn't a place to leave "good enough," it's one of the highest-return things you can fix on the whole site.
The fix leans into cutting fields to the minimum, showing progress, not asking for the same information twice and not making people hunt for shipping costs after they've entered their billing address.
Always try and get a fresh pair of eyes (ideally not someone who has seen it before) to buy something on your site and watch where they hesitate.
Making someone register before they can buy is one of the most reliable ways to lose a sale. They came to purchase, not to start a relationship, and a mandatory account feels like a toll booth.
The fix here is to offer guest checkout, always. You can invite them to create an account after the order's placed, when the friction costs you nothing.
The single most common reason people abandon carts is extra cost that shows up later into the purchase (things like shipping, taxes, fees) which is data cited by about 48% of shoppers.
The number a customer sees on the product page becomes the number they expect to pay. When checkout quietly inflates it, trust evaporates and so does the order.
The fix is to always be honest about cost early. Show shipping estimates on the product or cart page. If you can offer free shipping over a threshold, say so loudly - roughly two-thirds of shoppers now expect free shipping and many treat unexpected shipping fees as a dealbreaker.
If you can't offer it, at least remove the surprise.
If someone's preferred payment method isn't there, a chunk of buyers simply won't improvise. This is especially true on mobile, where one-tap options have become the default expectation.
The fix here is to offer the methods your audience actually uses - cards, PayPal, and the express wallets (Apple Pay, Google Pay, Shop Pay) that collapse checkout into a single tap.
These also tend to convert better precisely because they remove form-filling.
Mobile now drives around 59% of e-commerce sales, yet it still converts below desktop almost everywhere and a clumsy mobile checkout is a big reason why.
Tiny tap targets, a number pad that doesn't appear for the card field, a coupon box that eats the screen: each one costs orders from the majority of your traffic.
The fix is to treat mobile as the primary checkout, not the afterthought. Test the real thing on a real phone. The express wallets above do a lot of the heavy lifting here.
Product page mistakes (where trust is won or lost)
People can't touch what you sell. The product page has to do the reassuring that a physical store does for free.
It's easy to paste the manufacturer's copy when you're loading products. The problem is that every other retailer selling that product pasted the exact same words, so you're competing for rankings with identical content and giving the shopper no reason to buy from you.
The fix is obviously to rewrite it. Use real language, set context, go past the spec sheet to why a feature matters. Add your own images and link to related products. Be creative. This is also how you stay legible to AI shopping tools, which lean on distinctive, well-structured copy (more on that below).
The only thing worse than canned copy is none. A product with no description won't sell, and it gives search engines and AI tools nothing to work with.
The fix - every product gets real copy. No exceptions.
I shop online a lot, and this is the thing that most often makes me abandon an order: titles that don't match descriptions, images that don't match the product, shipping times that seem to change between pages. Anything that triggers a "wait, is that right?" calls the whole store's trustworthiness into question.
The fix is to keep titles, descriptions, images, specs, and shipping details consistent across every page they appear on. One contradiction can sink an otherwise-ready sale.
The biggest blocker to an online purchase is "I just need to see it in person." People want to know what the thing actually looks like, how big it is, how it sits in a real space, all before they commit. One catalog photo doesn't cut it.
The fix is to always show more. Multiple angles, scale references, detail shots, lifestyle context, video where it helps. Give people enough to build the confidence a fitting room or a shelf would.
Around 95% of shoppers read reviews before buying. If you're not collecting and showing them, you're missing both the trust signal future customers want and the keyword-rich, ever-refreshing content search engines love.
The fix is to actively ask for reviews after purchase and display them. Go beyond stars where it helps. For example, H&M's "true to size" feedback is a good example of structured review data that answers the exact question a hesitant buyer has.
When customers keep asking the same things like "is this compatible with…", "what's the shipping weight", "what's in the box" - that's not noise, it's a map of what your product page is missing. Unanswered questions become uncompleted orders.
The fix is to try to answer product questions on the product page and order/shipping questions in clear FAQs. Mark them up with FAQ schema so they're eligible to surface directly in search and AI answers.
These are the unglamorous ones we find on almost every audit. They don't announce themselves, but they bleed visibility.
Having several H1s is, to Google, what typing in all caps is to a reader: if everything's shouting, nothing stands out. A category page should have one H1 (the category name); a product page should have one H1 (the product name). Everything else is a subheading.
That new product model with a near-identical page to last year's version?
If both stay indexed, you're splitting your own ranking signals and diluting your shot at the short-tail terms for that product line. Picture the iPhone 17 page competing with the 16, the 15, and every variant for "buy iPhone."
The fix is to build a real strategy for product versions and variants while taking care of canonical tags, consolidation, and redirects, all for the effort of not bidding against yourself.
Pulling a product from your category and listing pages keeps the storefront tidy, but if the page stays live with no navigation to it, it becomes an orphan which means it is sitting there unmaintained, found only by someone clicking an old link or stumbling in from search.
It's a poor first impression and a dead end.
The fix is to decide deliberately what happens to discontinued products. Keep and update them, redirect them, or retire them properly. Don't just hide them and forget them.
Deleting old product pages without redirects is worse. Someone clicks an old link or a search result and lands on "page not found." They hit back and they're gone before you can show them an alternative.
The fix is to redirect discontinued product URLs to the closest relevant product or category. A 301 keeps the link equity and keeps the visitor.
Redirecting old pages is the right instinct until that page gets replaced and redirected again, and again, and you've built a chain of hops that slows the page and dilutes its value. Nobody remembers all the legacy redirects.
The fix is to keep a documented list of redirects. Crawl the site periodically with a tool like Ahrefs or Screaming Frog to find chains and broken redirects, then clean them up.
If your store generates URLs that are a string of product-ID numbers, you're leaving value on the table. A clean URL with the product name reads better, shares better, and carries a little more SEO weight.
The fix is to use readable, keyword-aware URL structures that include the product name.
Alt tags are one of the most-missed opportunities on e-commerce sites, and it’s usually because images get imported in bulk and never optimized afterward.
It's tedious work (unless you learn how to leverage LLMs), but alt text helps your products show up in image search and strengthens the page's organic relevance.
The fix is to always add descriptive alt text to product images. Build it into your product-upload process so it's not a someday job.
Buying guides, comparisons, how-tos, and blog posts do double duty: they help your SEO and answer engine optimization, and they help shoppers make confident decisions. A store that only has product pages has nothing to rank for the questions people ask before they're ready to buy.
The fix is to build content around the decisions your customers are making. It's also the raw material AI tools pull from when they recommend products.
Getting the sale is one thing. Not leaving easy revenue on the table afterward is another.
You may not ship everywhere, but blocking entire regions outright throws away sales you could have made. I'm a Canadian living abroad, and when I tried to buy appliances for delivery in Canada on a Canadian card, one major retailer simply locked me out, while their competitor detected my location, set a Canadian store for me, and won the order. One of those companies understood that "not local" doesn't mean "not a customer."
The fix is to, instead of blocking, adapt. Detect location, set expectations on shipping, and offer the options you can. Make your shipping policy clear rather than slamming the door.
Letting visitors share products is free reach you're declining. For visual and lifestyle products especially, social platforms like Pinterest, Instagram, and TikTok have become genuine discovery and search engines in their own right.
Pinterest pins still surface prominently in image search, and a growing share of younger shoppers start product searches on TikTok rather than Google.
The fix is to add share options to product pages, and show up on the platforms where your audience actually discovers things. Don't treat social as an afterthought to search.
Re-marketing someone the product they browsed (sometimes with a nudge discount) is one of the most cost-effective ways to recover a sale. But relentless, uncapped re-marketing just annoys people, and still showing ads for a product someone already bought is wasted spend.
The fix is to set frequency caps across campaigns, and build an exclusion list for recent purchasers so you stop paying to reach people who've already converted.
Seven in ten carts get abandoned, and a good chunk of those people were a single nudge away.
A timely email follow-up like an offer of help, a reassurance, a gentle bit of urgency, or maybe even a small incentive work wonders in recovering orders you'd otherwise write off.
The fix is to set up an abandoned-cart sequence. Lead with help ("questions about sizing, shipping, or returns?") before you reach for a discount, so you're not training people to abandon on purpose.
If you're not running tests, you're guessing. The status quo never improves on its own. Featured products, page layouts, the checkout flow, the "Add to Cart" button's text and color, your cart-recovery messaging - all of it is testable, and small wins compound.
The fix is to always have at least a couple of tests running. You don't need a lab; you need the habit. Let your analytics tell you what to test next instead of relying on opinion.
This is the section that didn't exist when this article was first written, and it's the one most stores are getting wrong right now, because the rules changed faster than most teams noticed.
A fast-growing share of product discovery now starts inside an AI tool, not a search bar.
ChatGPT has shopping features and hundreds of millions of weekly users; Perplexity and Gemini recommend and compare products with citations; some can begin checkout on the shopper's behalf.
Retail traffic from generative-AI sources grew enormously through 2025. If your product information is thin, locked in images, or buried in scripts these tools can't read, you're absent from the exact moment a purchase decision is forming. The fix is to treat AI assistants as a storefront. Give them clean, specific, text-based product information they can actually parse, and make sure your pages are crawlable.
This is the search strategy for AI shift in an e-commerce context where the goal is to be the source the AI recommends, not just to rank.
Dominoes gets a lot of love for their one-click (and zero-click) checkout process. What you may have missed is their multi-device checkout options that allow you to order pizza on your Xbox. Many retailers are simply not paying attention to their audience to identify other platforms that could be leveraged for ordering capabilities.
71% of consumers now expect personalized experiences, and around 76% get frustrated when they don't get them.
Generic, one-size-fits-all merchandising is no longer neutral because it actively underperforms against stores using AI to tailor recommendations, search results, and offers.
The fix is to use the personalization tools your platform offers, such as tailored recommendations, dynamic search, and behavior-based email, in order to show people things they're actually likely to want. Start with the highest-traffic surfaces (home page, cart, post-purchase) and expand from there.
Most of these mistakes don't come from not caring. They come from a small team wearing too many hats, with no time to maintain redirects, rewrite product copy, run tests, and learn how AI shopping works all at once. That's normal. It's also fixable.
If your store is leaking sales through any of the above, that's exactly the work we do. Our e-commerce marketing team has worked on hundreds of online stores across Shopify, WooCommerce, HubSpot and more. We can help you find the leaks and close them.
Why is my e-commerce store getting traffic but not selling?
Usually the problem is at the bottom of the funnel: a high-friction checkout, surprise shipping costs, thin product pages, or missing trust signals like reviews. Start at checkout, since that's the closest point to the sale, then work backward to product pages and pricing transparency.
What is a good e-commerce conversion rate?
Most stores land between 2% and 4%, with desktop converting higher than mobile. Rather than chasing a benchmark, track your own rate over time and improve it through testing. Even small lifts compound as traffic grows.
How do I reduce cart abandonment?
Simplify checkout, offer guest checkout, show all costs upfront, add the payment methods your audience uses, and follow up with an abandoned-cart email sequence.
The average cart abandonment rate is around 70%, so even modest improvements here move real revenue.
How do I get my products to show up in AI tools like ChatGPT?
Give AI tools clean, specific, text-based product information they can read, implement product and review schema, and keep your pages crawlable. Distinctive content and structured data are what make a store eligible to be recommended inside AI shopping answers.