Cart Abandonment on a Small Business Website: Where the Sale Actually Dies

No email arrives. Nobody rings. A customer picks three items, reaches the page where the total is finally assembled, decides against it and closes the tab, and the owner of the shop learns nothing at all. Sales that never happen leave no paperwork.

That is the awkward thing about an abandoned basket. Every other failure on a business website eventually announces itself. A broken contact form produces a phone call from somebody who gave up on it. Downtime produces three. This one is silent, and it is the single most common outcome of a customer reaching your basket in the first place.

I have spent twenty years watching business owners investigate the wrong problem because it was the one that made noise. So this is a reference piece: what the abandonment figure actually measures, the six places a basket dies, which of those you genuinely control, and how to walk your own checkout the way a stranger would.

What the Abandonment Figure Actually Measures

The number everybody quotes comes from Baymard Institute, which puts the average documented cart abandonment rate at 70.22%. Before that goes anywhere near a business plan, look at how it is built. It is an average across 50 separate studies, and those studies disagree violently with each other. The lowest, from Forrester back in 2010, recorded 55%. At the other end, SaleCycle in 2020 recorded 84.27%.

A spread of nearly thirty points is not a benchmark. It is a band.

There is a second problem with treating 70% as a target to beat. Baymard's own survey work with US online shoppers found that 42% abandoned because they were just browsing and not ready to buy. That is not a fault in your website. It is somebody comparing prices on the bus, and no amount of checkout work reaches them. The fixable portion of abandonment is a good deal smaller than the headline suggests, which is exactly why the headline gets quoted so enthusiastically by people selling checkout services.

Baymard also estimates that the average large e-commerce site could gain roughly a 35% conversion improvement from checkout design alone. Worth reading carefully: that figure is drawn from large sites with elaborate checkouts, and Baymard sells checkout research for a living. Directionally useful. Not a promise.

The Irish backdrop matters too. CSO figures for 2025 put 85% of internet users buying goods or services online, while only 37.5% of enterprises here had e-commerce sales, dropping to 34.9% among small enterprises. Most local firms are not selling online at all. The ones that are compete with everybody, including the platforms that have spent a decade filing the friction off their own checkouts.

The Six Places a Basket Dies

Baymard's survey of shoppers who abandoned a checkout in the previous quarter produces a fairly consistent ranking. The percentages below exclude nobody, so they include people who gave more than one reason, and they describe US shoppers rather than customers here. In my experience the ordering travels.

Where it dies	Share who cited it	Who controls it
Extra costs appear late	39%	You
Delivery too slow or unclear	21%	You and your courier
Site asks for an account	19%	You
Card details do not feel safe	19%	You
Checkout too long	17%	You
Card declined	8%	The bank

Five of those six sit on your side of the line. That is better news than it reads.

The Total Changes at the Last Step

Extra charges such as shipping, taxes and fees are the largest single cause, at 39%. The mechanism is worth stating plainly, because it is not really about the money. Buyers do the arithmetic in their head on the product page. Your checkout then produces a different answer. What they feel at that moment is not expense, it is a small, sharp sense of having been handled, and people who feel handled do not negotiate. They leave.

Consider a case of the kind that plays out constantly. A fly-fishing tackle shop in Cavan sells rods, reels and lines around the country. Somebody builds a basket on a Sunday evening: a reel, a spool of backing, a rod. The rod ships in a nine foot tube, the courier prices that tube like a piece of furniture, and the delivery cost is calculated at the final step. Three items go in. Nothing comes out. The owner opens the shop on Monday to no orders and no explanation, and concludes that the week was quiet.

The fix is not to make delivery cheaper. It is to make it earlier. A shipping figure, a range or even a plain sentence about how large items are priced, shown on the product page, converts a nasty surprise into a known cost. Irish law already has a good deal to say about how delivery charges must be presented on your website, and meeting that standard properly tends to solve the commercial problem as a side effect.

The Delivery Promise Is Vague

Delivery being too slow accounted for 21%. Slow is relative to what was promised, which means the failure is often in the wording rather than the logistics. "Dispatched within three to five working days" reads, to a customer buying a present, as an unknown quantity that might be next week or might be the week after.

Give a date. Even an estimated one, even a cautious one. A person deciding whether to buy from you or from a marketplace that will tell them precisely when the box lands is making a comparison, and vagueness loses it.

The Site Asks Them to Create an Account

Roughly one in five abandoned because the site required an account before it would take the order. Consider what is actually being proposed there. Somebody who wants one reel is being asked to enter a relationship with your shop, choose a password, and hold a login they will never use again. Against that, the marketplace alternative takes their money in four taps.

Guest checkout is the answer for most owner-managed shops, and the usual compromise is a good one: take the order as a guest, then offer account creation on the confirmation page once the money is in. You still hold the order record. The email address is still yours. The customer simply was not asked to commit to anything before they had a reason to.

The Payment Step Does Not Look Safe

Another 19% abandoned because they did not trust the site with their card details. Trust at that moment is not built by a padlock in the address bar, which nobody looks at, and is not built by a badge in the footer either. It is built by the things around the payment field looking like they belong to a real business: an address, a phone number, a returns policy that reads as though a person wrote it, and a payment step that behaves like every other payment step the customer has ever used.

Getting the mechanics right underneath that is a separate discipline. The point for this article is narrower. A checkout that looks improvised will be treated as improvised, and the customer's only defence against an improvised checkout is to close it.

The Form Is Longer Than It Needs to Be

A checkout that was too long or complicated accounted for 17%. Baymard's benchmarking of live sites found the average flow in 2024 running to 5.1 steps and 11.3 form fields. That is an improvement on 11.8 in 2021 and 12.7 in 2019, and it is still comfortably above the eight or so fields most shops actually need to fulfil an order.

The extra fields are rarely decisions anybody made. They arrive with the plugin. A second address line nobody fills in. The company name field on a shop that sells to consumers. Worst of all, a separate billing address block presented to a customer paying with a card registered at the address they just typed. Each one is a small demand for attention at the precise moment the customer is deciding whether you are worth the bother.

The Card Is Declined

The smallest category, at 8%, and the one you control least. What you do control is what the customer sees next. A page that says "payment failed" and nothing else ends the sale. One that keeps the basket intact, explains that the bank refused the transaction rather than implying the customer did something wrong, and offers a second payment method, converts some proportion of those into orders. The decline belongs to the bank. The dead end is yours.

Walking Your Own Checkout in Five Passes

Most owners have never bought from their own shop under realistic conditions. That is not carelessness. It is that the person who built the site is the worst possible test subject for it.

I signed off on a checkout years ago because I got through it in about forty seconds and declared it fine. My browser had my address saved, my card saved and my email saved, so I had personally typed almost nothing. A colleague tried it cold on a phone and hit eleven fields. That is the sort of error you only make once.

Run these five passes in order. Do the first one in a staging environment so nothing you break reaches production.

1. Clone. Copy the live site to a staging environment before touching a single field, so any change you regret can be discarded rather than rolled back under load.
2. Arrive cold. Open the shop on a device that has never seen it, signed out, with no saved cards and no autofill. This single pass surfaces more than the other four combined.
3. Count. Write down every field the checkout asks for, then mark the ones you have genuinely used in the last year. Anything unmarked is a candidate for removal.
4. Verify the total early. Go back to the product page and ask whether a customer could work out the final figure, including delivery, before they reach the payment step. If not, that is your largest leak.
5. Deploy one change at a time. Push a single change to production, leave it for a fortnight, and compare. Four changes deployed together tell you nothing about which one worked.

The whole exercise takes an afternoon. Pass two takes about four minutes and is the one nobody does.

Who Controls the Checkout You Are Using

Everything above assumes you can actually change these things. That assumption does not hold everywhere, and it is worth knowing which side of the line your shop sits on before you plan a single fix.

On a hosted website builder, the checkout is a component the vendor ships. You get the fields they chose, the flow they designed and the payment methods they support. When you want the company name field gone, the route available to you is feature feedback. Plenty of owners discover this only once they have a reason to care, which is usually after they have read something like this article and gone looking for the setting.

WordPress treats the checkout as yours. It powers around 43% of the web, and part of the reason is that it refuses to draw that line: fields, steps, wording, payment methods and the order of the whole flow are configurable, and the entire plugin ecosystem is available when you need something the core does not do. That is the practical difference between renting a shopfront and owning one. You can move the counter.

Doing that safely needs two things beyond the ability to edit. It needs a staging environment, so a checkout experiment never runs on live orders, and it needs a backup you can restore from when an experiment goes badly. Both are included at Web60 alongside everything else a working shop needs for sixty euro a year, on Irish infrastructure, with no separate line item for the staging site. The owner who wants to remove two fields on a Tuesday afternoon can clone, edit, verify and deploy without booking anybody's time.

Where a Hosted Checkout Is the Better Call

Some shops should not be running their own checkout at all, and pretending otherwise would be dishonest.

If you sell one product, or a handful of workshop places a month, or a single annual membership, a hosted payment link is genuinely the better tool. The vendor maintains it, the payment provider carries the compliance, and there is no checkout to audit because there is barely a checkout. Everything I have described above only pays for itself once you have enough order volume that a few percentage points of abandonment is real money, and enough product variety that delivery and options get complicated. Below that threshold, a full shop is overhead you are carrying for no return.

The calculation changes when the catalogue grows, when delivery stops being one flat rate, or when you want the customer relationship rather than a receipt issued by somebody else's brand.

What a Better Checkout Cannot Do

Two honest limits, because a checkout project sold as a cure disappoints predictably.

The first is intent. That 42% who were just browsing remain unreachable. If your traffic is largely people who arrived from a social post, looked, and left, your abandonment rate will stay high no matter how good the form is, and the work that pays off is earlier in the journey rather than at the payment step.

The second is a genuine trade-off, and it catches people who read an article like this one and start deleting fields enthusiastically. Some of what a checkout collects is evidence. A full delivery address, an order record, a customer history: those are what you produce when a card payment is later disputed, and the card schemes are strict about what counts. Before you strip a field, check it is not one you would need in a chargeback, where the money leaves your account before anybody asks you for proof. Shorter is better right up until the point where it is not.

Conclusion

The reason checkout abandonment goes unexamined in small shops is not that owners do not care about it. It is that it is the only failure mode on a website that produces no complaint, no ticket and no phone call, so it never reaches the top of anybody's list. Meanwhile the total that appeared too late, the account nobody wanted and the eleven fields that could have been eight go on doing quiet damage every week.

You do not need a redesign to find out. What you need is one cold walk through your own checkout on a device that has never seen it, and a pen. Whatever makes you hesitate in those four minutes is doing the same thing to everybody else.

Frequently Asked Questions

What is a normal cart abandonment rate for a small online shop?

The figure most often quoted is Baymard Institute's average documented rate of 70.22%, but that is an average across 50 separate studies whose results range from 55% at the low end to 84.27% at the high end. Treat it as a rough band rather than a target. A small shop with a narrow product range and returning customers can sit well outside it in either direction, so your own trend over several months tells you more than any industry average does.

Why do customers abandon a basket at the payment step?

Baymard's survey work with US online shoppers puts extra charges such as shipping, taxes and fees at the top, cited by 39%. Delivery being too slow accounted for 21%, not trusting the site with card details 19%, being forced to create an account 19%, a checkout that was too long or complicated 17%, and a declined card 8%. Separately, 42% said they were simply browsing and never intended to buy, which is not a design problem at all.

Should my online shop allow guest checkout?

For most owner-managed shops, yes. Roughly one in five shoppers in Baymard's research abandoned because the site required an account, and a first-time buyer of a single item rarely wants an ongoing login with your shop. The usual compromise is to take the order as a guest and offer account creation on the confirmation page, once the money is already in. You still capture the email address and the order record either way.

How many fields should a checkout have?

Baymard's benchmarking found the average checkout flow in 2024 ran to 5.1 steps and 11.3 form fields, down from 11.8 in 2021 and 12.7 in 2019, while most sites need around 8 fields to complete an order. Count yours. Fields that exist because the plugin shipped with them, such as a second address line, a company name or a separate billing block for a card paid at the same address, are usually the first candidates to remove.

Can I change the checkout on a website builder like Wix or Squarespace?

Only within the limits the platform sets. On a hosted builder the checkout is a fixed component: you get the fields, the flow and the payment options the vendor decided to ship, and when you want a field removed you submit feature feedback rather than make a change. A self-hosted WordPress shop treats the checkout as your code, so fields, steps, wording and payment methods can all be altered and tested before they reach production.

Does site speed affect cart abandonment?

It contributes, though it is rarely the whole story. A slow checkout adds waiting to a moment when the customer is already weighing up whether to proceed, and it makes a payment step feel less reliable. The larger causes in the survey data are informational rather than technical: what the total turns out to be, when the goods will arrive, and how much the site asks of the customer before it will take the money.

Sources

Baymard Institute, cart abandonment rate statistics compiled from 50 studies

Baymard Institute, reasons for cart abandonment and checkout usability benchmark

Baymard Institute, average checkout form fields and flow length

Central Statistics Office, Household Digital Consumer Behaviour 2025, internet purchases

Central Statistics Office, Information Society Statistics Enterprises 2025, e-commerce