29 minutes reading time (5830 words)

E-commerce checkout: the basket that empties itself

Umberto Mazza at Ozone Coffee in London, completing an online purchase on his smartphone with his laptop open beside him

Tuesday morning, Ozone Coffee, Leonard Street. Outside, Shoreditch is waking up under a lead-grey sky; inside, the smell of freshly roasted coffee drowns out the damp coats. I buy online all the time, for work and for home, and while I'm waiting for my coffee I order a Christmas present from a small British shop on my phone.

Email. Postcode. The address pops up on its own in a list; I tap it. I pay with my phone's digital wallet.

Three taps. The coffee hasn't even arrived and the order's already confirmed. Nobody asked me for a tax code.

The week before, I was in Santa Maria Capua Vetere, in my studio on Via Melorio. Late afternoon, the hum of the building's lift going up and down behind the wall, a forgotten coffee on the desk. On screen, a training-course shop built in VirtueMart. A year after launch, without spending a single euro on advertising, purely on organic search traffic, that shop had started getting its first orders and its first quote requests. And right then, the course codes had fallen out of sync: a paid order risked being attached to the wrong course.

Umberto Mazza in his studio in Santa Maria Capua Vetere, checking an e-commerce shop's orders on his computer in the late afternoon

A year to reach the till. One wrong detail to lose it all in the final metre.

In Naples you know it well: the coperto, the cover charge in small print at the bottom of the menu that you only really notice when the bill arrives. In London it's the 12.5% service charge lurking at the bottom of the receipt. Online, the same thing happens every day, with one difference: the customer hasn't got their jacket hanging on the back of the chair. Their thumb is a millimetre from the "back" button.

Checkout isn't a form: it's your last conversation with the customer. Every extra field is a question put to someone who's already said yes. On Black Friday you don't lose customers on price. You lose them because you interrogate them.

The bill arrives: how many baskets get left halfway

Berto rings me every year around mid-November. He's got an online shop, he looks at his stats, sees dozens of full baskets abandoned, and convinces himself the problem is his competitor's prices. Bobby, over in Salford, looks at the same charts and asks a different question: which line of the form did they stop on?

Bobby's right.

The Baymard Institute, an independent research centre for e-commerce usability, works out an average basket abandonment rate of 70.22% across 50 studies. On its checkout research page, the same institute gives 70.19%, the result of fourteen years of tracking: the two figures come from different aggregations, so I'm reporting both. Baymard is redoing its checkout testing from scratch and expects to publish the new research in the second quarter of 2027, so these are today's figures, not eternal truths.

Before you slash your prices, a moment of honesty. According to Baymard, 42% of people who abandon were just browsing or weren't ready to buy. You won't win those baskets back, and you shouldn't try: anyone who walks into a shop to have a look has every right to walk out empty-handed.

Black Friday revenue is decided by everyone else. Leave out the browsers, and these are the reasons for abandonment Baymard has measured:

  • 48%: extra costs too high (delivery, tax, fees);
  • 24%: the site forced them to create an account;
  • 22%: delivery too slow;
  • 18%: didn't trust the site with their card details;
  • 17%: checkout too long or complicated;
  • 16%: couldn't see the total cost before reaching checkout.

Have a proper look at that list. Nearly every item is about how the form is designed and written. Not the product. Not the photo on the home page.

In The site that doesn't convert I went through the whole journey of a website that pulls in visits and doesn't sell. Here the focus is narrower: the last three screens, those thirty seconds when the customer has already made up their mind and is just looking for the door to pay at.

The $300 million button: compulsory registration

The story's told by Jared Spool, founder of User Interface Engineering. It came about while Luke Wroblewski was writing Web Form Design: Filling in the Blanks, one of the go-to books on web form design.

A big American retailer with $25 billion in turnover. Spool doesn't name it, and nor will I. At the till, customers had only two options: log in with email and password, or register. No way to buy without an account.

Watching people in front of that screen, the research team spotted two sticking points. New customers couldn't be bothered to invent yet another password for a purchase worth a few dollars. Regulars couldn't remember which email they'd registered with, got their password wrong and got lost in the password-recovery process.

The designers took the Register button away. In its place they put "Continue", with a message: you don't need an account to buy.

Purchases went up by 45%. An extra $15 million in the first month, $300 million in the first year.

Three hundred million left standing on the doorstep, all for one more password.

What to do in practice

  • Guest checkout is the main route. At the till, the first field is email, then delivery. Anyone with an existing account finds a discreet link at the top to sign in.
  • Offer the account after payment. On the confirmation page you've already got their name, email and address: just ask for a password to save them. The customer doesn't retype a thing and the order's already in the bank.
  • The password is no longer the only way in. Passkeys, the standard that lets you sign in with your face or fingerprint from your phone, or a one-time code by email, get rid of the thing people forget most often.

Field by field: the anatomy of a form that doesn't wear people out

A badly made checkout form is like the counter at a government office: three different forms asking for the same details, filled in by pen, one after another.

In Baymard's testing, a complete checkout can get by with 12–14 elements, meaning 7–8 actual fields to fill in. The average US checkout shows 23.48, of which 14.88 are fields. Another study by the same institute, counting text fields only, puts the average at 11.3: the two measures count different things, but they tell the same story. Almost everyone asks for twice what they need.

Infographic of an e-commerce checkout with eight fields, the label above each field and the matching autocomplete attribute value

HTML that does the work for you

The browser can do most of the typing, if the code lets it. Safari, Chrome, Edge and Firefox fill in name, address, phone and card on their own, as long as every field declares what it holds with the autocomplete attribute, defined in the WHATWG HTML specification. A field called campo_1 with no attribute forces the customer to type everything out with their thumb.

<!-- Contact -->
<label for="nome">First name</label>
<input id="nome" name="nome" type="text" autocomplete="given-name" required>

<label for="cognome">Surname</label>
<input id="cognome" name="cognome" type="text" autocomplete="family-name" required>

<label for="email">Email for your order confirmation</label>
<input id="email" name="email" type="email" autocomplete="email" required>

<label for="telefono">Mobile (for the courier only)</label>
<input id="telefono" name="telefono" type="tel" autocomplete="tel" required>

<!-- Delivery -->
<label for="indirizzo">Street and house number</label>
<input id="indirizzo" name="indirizzo" type="text" autocomplete="address-line1" required>

<label for="cap">Postcode (CAP)</label>
<input id="cap" name="cap" type="text" inputmode="numeric" autocomplete="postal-code" required>

<!-- Card: text field with numeric keypad, never type="number" -->
<label for="carta">Card number</label>
<input id="carta" name="carta" type="text" inputmode="numeric" autocomplete="cc-number" required>

<label for="scadenza">Expiry date (MM/YY)</label>
<input id="scadenza" name="scadenza" type="text" inputmode="numeric" autocomplete="cc-exp" required>

<label for="cvc">Security code</label>
<input id="cvc" name="cvc" type="text" inputmode="numeric" autocomplete="cc-csc" required>

<!-- One-time code received by SMS or email -->
<label for="codice">Verification code</label>
<input id="codice" name="codice" type="text" inputmode="numeric" autocomplete="one-time-code">

Two details make all the difference on a phone. type="email" and type="tel" bring up the right keyboard, with the @ front and centre or the number pad. And for card numbers and Italian postcodes, you never use type="number": that type is designed for quantities, adds side arrows that change the value if you brush the scroll wheel, and mangles leading zeros. Rome's CAP, 00100, knows all about it. The right combination is a text field with inputmode="numeric".

A word of warning before you copy the code. If you take payments through Stripe, Mollie or PayPal, the card fields are usually provided by the gateway, in a secure frame inside your page. Coding them by hand means card numbers pass through your server, and then the security obligations under the PCI DSS standard get a lot heavier. The card fields above are there to show how the browser thinks, not to replace the gateway's own.

Labels above, never inside

It's still fashionable to drop the label and write the name of the field in light grey inside the box. The moment you tap the field, the text vanishes.

Picture someone filling it in on their phone on the regional train between Caserta and Napoli Centrale, or standing on the Northern Line. They type half the address, look up to check the stop, look back at the screen: a half-filled box and no label. Was it delivery or billing? To find out, they have to delete the lot.

Baymard doesn't mince its words: on mobile forms, labels never go inside the field. They stay above it, always visible.

Validation as you type

According to Baymard, 31% of sites have no real-time validation at all: you only find out about the error after hitting the final button, when the page reloads with a list of messages in red.

But badly done validation is even worse. The two mistakes I come across most often are mirror images of each other. The first: the field turns red while you're still typing, three letters into your email, before you've even got to the @. It's like someone interrupting you after your first word. The second: you fix the error and the message stays put until you click somewhere else.

The rule's simple. Check the field when the user leaves it, and clear the error the instant they fix it. Baymard cites Fitbit as an example: it updates the validation the moment the input becomes valid.

On the card number, nearly all checkouts run the Luhn check, a formula that verifies the digits are consistent: 15 of them for American Express, 16 for most cards, up to 19 for some. It catches two swapped digits instantly. What it can't tell you is whether the card's expired or has no funds: only the bank knows that.

When a form responds like that, with the right timing, it stops being a counter and becomes a conversation. It's the principle I talk about in Microinteractions: when your website stops being a brochure and starts talking back.

Italy vs the UK: the same form won't suit both

Anyone selling in Italy with a theme born in English almost always makes the same mistake: they translate the labels and leave the fields as they are. It's like taking a jacket cut for a client in Manchester and handing it to one in Caserta, having only changed the label inside. It pulls across the shoulders. As I explain in A bilingual website isn't a translated website, localising a purchase journey means adapting it to the postal system, the tax rules and the habits of the people buying.

The address

Two smartphone checkout screens side by side: on the left, address lookup with a British postcode; on the right, the Italian form with CAP, province and intercom name

In the UK, the postcode, five to seven characters mixing letters and numbers, identifies a street, a stretch of street or sometimes a single building. Type EC1V 2NX and the form offers you the list of addresses on that stretch of City Road: pick 124 City Road and you're done. Two seconds, zero typos.

Italy doesn't work like that, because the CAP, the Italian postcode, is far coarser. In Santa Maria Capua Vetere, 81055 covers the whole municipality; in Caserta, 81100 covers the whole city; Portici is 80055. Naples, on the other hand, has dozens of different CAPs, from 80121 to 80147. In no case does the CAP alone give you the street: the form has to ask for street and house number, town, province code (CE for Caserta and Santa Maria Capua Vetere, NA for Naples and Portici) and an optional field for staircase, flat number or the name on the intercom. Anyone who's sent a parcel to a gated residential complex in the Caserta area, or to a palazzo in the Quartieri Spagnoli, knows the drill: without the staircase and intercom name, the courier leaves a card and the parcel goes back to the depot.

The codice fiscale: only ask the people who need it

Then there's the very Italian affliction: the codice fiscale, the personal tax code, made compulsory for anyone buying a pair of shoes or a book. When I ask why, the answer's nearly always the same: "My accountant told me to, it's because of e-invoicing."

That's a misreading, and it costs orders. Selling goods online for home delivery, so-called indirect e-commerce, is treated by the taxman like mail-order selling. That's why an invoice to a private customer isn't compulsory, unless the customer asks for one by the time of the transaction (Art. 22 of Presidential Decree 633/72). The Agenzia delle Entrate, Italy's tax authority, has confirmed this several times, most recently in a 2022 ruling: these sales are also exempt from issuing receipts and from electronically transmitting daily takings. No invoice requested, no codice fiscale to ask for.

The right design is straightforward. No tax details by default, and a checkbox under the address: "Mi serve la fattura" (I need an invoice). If the customer doesn't tick it, and they almost never do, they pay and they're done in ten seconds. If they do, the form opens up the right fields: the codice fiscale for a private customer; for a business, company name, VAT number and either the recipient code for the SDI, the Sistema di Interscambio that handles Italian e-invoicing, or a PEC, a certified email address.

Before-and-after comparison of an Italian checkout: on the left, a compulsory codice fiscale field showing an error; on the right, the optional Mi serve la fattura (I need an invoice) checkbox that opens the tax fields

Two caveats from the shop floor. First: this rule applies to goods that are shipped. For services, such as the courses or timestamps in the two cases I'll come to later, and for business-to-business sales, the rules change: there, you build the framework with your accountant, not with a plugin. Second: if your client also sells in a physical shop, I'd advise against issuing invoices from the e-commerce site. The site collects the details; the invoice comes from the management software they already use. Otherwise they end up with two different numbering sequences in the same year, one from the site and one from the shop. It's not a legal requirement: it's admin common sense.

In the UK this step simply doesn't exist. Nobody asks a British consumer buying online for any tax document, and your National Insurance number is for work, pensions and personal tax, not for buying a lamp. It's the biggest gap between the two ways of designing a checkout: in Italy we start from the back office's fears, in London from the convenience of the person holding the card.

The phone number

The phone field hides one small bit of friction and one big one. The small one is format: +39 or 0039, a UK mobile with or without the leading zero. You sort it with a country-code selector preset to the delivery country and validation that's forgiving about spaces and zeros.

The big one is suspicion. Nobody wants to hand over their number only to get sales calls three days later. Baymard found that 39% of the largest e-commerce sites ask for a phone number without explaining what it's for; in its testing, 14% of users said they'd abandon checkout when the field was compulsory with no reason given. One line under the field does the trick: "We'll only use this if the courier needs to contact you about your delivery."

The digital cover charge: hidden costs are now illegal

Back to the table. Online, the coperto trick has a technical name: drip pricing. I show you a bag at €49. At step three, along come €7.90 for delivery, €2.50 for "handling", €1.90 for "packaging". By the summary, the price has gone up by a quarter.

For years it was seen as a clever bit of salesmanship. Today it's banned, on both sides of the Channel.

In the UK, the Digital Markets, Competition and Consumers Act 2024 banned drip pricing from 6 April 2025: mandatory charges have to be included in the price shown from the start.

In Italy, the Consumer Code (Codice del consumo) has long required traders to state, before the consumer is bound, the total price including taxes and delivery charges (Art. 49(1)(e)). And then there's paragraph 6 of the same article, which hardly anyone knows about: if the trader hasn't told the customer about additional charges, the customer doesn't have to pay them. Not a fine for the site: the cost simply isn't owed.

Just how fine the line is shows in the Ticketmaster case in the UK, after the sale of tickets for the Oasis reunion in 2024. The CMA, the UK's competition authority, investigated and secured formal undertakings from Ticketmaster to give fans clear information about prices and ticket types before and during purchase. Voluntary undertakings, with no admission of liability. And according to the trade press, the investigation ruled out dynamic pricing. No scam: what was missing was clarity about what people were buying and for how much. That alone is enough to land you under investigation and in the papers.

The same goes for pre-ticked boxes that slip a warranty, insurance or a donation into the basket. Harry Brignull, a British designer, gave these practices a name back in 2010, dark patterns: among them, hidden costs and "sneak into basket", the item slipped into your basket on the sly. I've written about it in The trick that's now against the law: dark patterns, the DSA and deceptive design. Here, one rule will do: under Art. 65 of the Consumer Code, if consent to an extra payment was obtained through a default option the customer had to opt out of, the customer is entitled to a refund.

The final button: the words the law requires

The last centimetre. The coloured rectangle where the customer rests their thumb.

Plenty of people treat it as a style choice and write "Confirm", "Proceed", "Send". Soft words, chosen so as not to scare anyone off. They can make the order non-binding.

In Italy, Art. 51(2) of the Consumer Code states that the order button must bear only the words "ordine con obbligo di pagare" (order with obligation to pay) or an equivalent, unambiguous wording. If it doesn't, the consumer isn't bound by the contract. With its judgment of 30 May 2024 in Case C-400/22, the Court of Justice of the European Union closed off the last loophole too: only the wording on the button counts, not explanations placed above or below it.

In the UK the rule's the same. The Consumer Contracts Regulations 2013 require the button to be labelled only with "order with obligation to pay" or an equivalent phrase (regulation 14, paragraph 4), and there too, without it, the consumer isn't bound (paragraph 5). The UK government's guidance gives "pay now" as a clear example.

In practice:

  • In Italian: "Ordina con obbligo di pagare" (order with obligation to pay), "Acquista e paga" (buy and pay), "Paga ora" (pay now).
  • In English: "Pay now", "Buy now", "Order with obligation to pay".
  • Avoid: "Confirm", "Proceed", "Continue", "Send order", "Complete" (in Italian: "Conferma", "Prosegui", "Continua", "Invia ordine", "Completa").

Order summary screen with all costs visible and the final Acquista e paga (buy and pay) button, next to a wrong button reading Conferma (confirm)

Next to the button, in November, "Pay in 3" often pops up too. Careful: that's credit advertising, with its own rules. I covered it in "Pay in 3": from 20 November, it's credit advertising.

Payment: 3-D Secure, digital wallets and the single click

Between the tap on "Pay now" and the money landing, there's still a way to go. And some of the abandonment happens right there, in silence.

Amazon's single click

In September 1999 Amazon was granted a US patent on one-click buying: anyone who'd already bought could order again with a single tap, using their saved address and payment details. The patent expired on 11 September 2017; some sources say the 12th. In Europe it never got off the ground: the European Patent Office rejected the application in 2007 for lack of inventive step, and the decision was upheld in 2011. In the US, on the other hand, even Apple took out a licence from Amazon in 2000.

Eighteen years of exclusivity, then the technique became fair game for everyone. If a shop today makes a customer coming back for the third time retype their address and card, it's keeping a door shut that nobody's asking it to keep shut any more.

3-D Secure without the panic

When the customer pays by card, 3-D Secure comes into play: the step where the bank asks for confirmation, usually through its app with a fingerprint or face scan. In Europe it's required by the PSD2 payment services directive, through strong customer authentication; the UK has equivalent rules.

There's an exemption for small amounts: under €30, or £25 in the UK, verification can be skipped, up to a limit of five consecutive transactions or €100 cumulatively. Two caveats: the exemption has to be requested by the payment provider, it isn't automatic, and the customer's bank can always turn it down.

On my clients' online shops I always integrate PayPal and, depending on the project, Stripe and Mollie too. Three providers means three different verification screens. Someone who rarely shops online suddenly sees their bank's window pop up, with logos that aren't the shop's, and thinks it's an error or a scam. That's why I always put a line on the shop's page, just before that step: "Your bank may now ask you to confirm in its app or with an SMS code." One sentence. It prepares the eyes for what they're about to see.

And when something goes wrong, the form mustn't forget. According to Baymard, 34% of sites wipe the card details just entered after a validation error. Making someone who's just hit a snag type sixteen digits all over again is the quickest way to see them off.

Digital wallets at the top, not the bottom

On lots of Italian online shops, the digital wallet buttons sit at the bottom, after name, address, CAP and phone number. By then, the customer has already done all the legwork those buttons were meant to save them.

Two smartphone checkouts side by side: express payment buttons at the bottom of the form, and the same buttons at the top, above the address fields

A digital wallet isn't just another payment method: it already holds the card and the address, confirmed with a fingerprint. It goes at the top of the checkout, above the form, with a clear divider: "or fill in your details below". One tap, and the order's done without filling in a thing.

Accessibility: the checkout everyone has to be able to finish

In Italy, since 28 June 2025, the European accessibility rules (Legislative Decree 82/2022, which transposes the European Accessibility Act) also apply to the online shops of businesses with at least 10 employees or more than €2 million in turnover. Anyone already online before that date has a transition period until 28 June 2030. In March 2026, AgID, Italy's Agency for Digital Italy, published the guidelines for applying them.

In the UK there's no European law, but there is the Equality Act 2010: anyone providing services to the public, websites included, mustn't discriminate against disabled people and has to make the reasonable adjustments needed (sections 20 and 29).

Within checkout, two criteria from WCAG 2.2, the international web accessibility guidelines, decide whether the customer makes it to the end:

  • 3.3.7, redundant entry. Information already provided in the same process mustn't be asked for again. The classic example is the billing address: if it's the same as the delivery address, a tick box is enough, not filling it all in twice.
  • 3.3.8, accessible authentication. You can't impose cognitive tests to sign in, like remembering a password unaided or solving a puzzle. In practice: the login code field has to accept pasting and password managers. The criterion covers login; but blocking paste on the payment verification code is still a mistake, and I still come across it on sites that ought to know better.

Clear labels, readable contrast, everything usable by keyboard, no repetition: they're not just about avoiding penalties. They make checkout faster for everyone, as I explain in Web accessibility 2026: from legal obligation to competitive advantage.

Two real shops: what I fixed

Checkout theory looks tidy on paper. Then you open up a real shop, with its own tax rules and its own data, and find that the worst problems crop up where the interface meets the database.

The VirtueMart course shop

The first is the training-course shop from the opening scene, which I'm not naming for confidentiality reasons. It runs on Joomla with VirtueMart. A year after launch, on organic traffic alone, the first orders and the first quote requests are coming in for course packages with several attendees.

Quote request form for a corporate course package, with number of attendees, company details and participants to be added later

This shows something standard themes ignore: checkout isn't just for paying. A company that needs to enrol five employees doesn't pay with a personal card and often doesn't have all the names yet. For them, the right route is a quote request, with fields designed for a business: number of attendees, company details, and the option to name the participants later. In training, attendees' details genuinely matter, so there the codice fiscale and the ATECO code, Italy's classification of economic activities, are legitimate. The point is to ask for them at the right moment and help people fill them in, not to block the purchase. Next to those fields I added guidance on the correct format, instead of a generic error: the same principles as in Dashboard UX: how to design management software people can actually use.

Then there's the problem from the opening scene. The course codes had fallen out of sync with the packages, and an order could end up linked to the wrong course. The problem had started upstream, outside the shop. I sorted it last week.

The lesson applies to everyone: you can design the most beautiful checkout in the world, but if the data behind it doesn't match up, it breaks exactly when the orders start coming in.

Tempomark: one order, one file

The second is Tempomark, a shop selling digital timestamps for long-term rental contracts, for a Bulgarian client. Whoever buys the service uploads the document to be timestamped. I designed and built the site myself.

And here the mistake was mine. In development, file uploads worked; in testing, everything ran smoothly. Then the site went live, real customers turned up, and, as always, real-world use found the limit I hadn't spotted. The upload was unclear and, above all, it never locked: the file could be uploaded, changed or deleted at any time. So lots of customers were paying for one timestamp and uploading several documents to be timestamped within the same order.

No internal test would have shown me that. I knew how that form was meant to be used; customers didn't, and had no reason to. It's the most useful lesson in this whole article: you can't test a checkout through the eyes of the person who built it.

The fix has two parts. First: uploading, editing and deleting files are locked once the order moves to "in preparation" or "closed/delivered". Second: the upload became clearer to use. The principle to take away: the interface must know at every moment where the order stands, and change what it allows accordingly.

Diagram of the three states of an order and the file permissions: free upload, files locked during preparation, read-only once the order is closed or delivered

The numbers. Up to 10 February 2025, the shop had taken 29 orders in total. Since I've been managing the orders, from then to today, 753 have come in.

This is where I put the handbrake on, because I'm not going to spin you a yarn. That jump isn't down to the file lock alone. The lock closed a gap I'd left open myself; careful, day-to-day order management did the rest. Inflating a cause-and-effect link to take all the credit is the quickest way to lose credibility.

Checkout in 15 checks before 27 November

Black Friday falls on Friday 27 November. You don't need to rebuild the site in two weeks. You need to go over the purchase form with a fine-tooth comb. These are the checks I run before every peak:

  1. Visible guest checkout: a new customer gets to payment without creating an account.
  2. Total including delivery already in the basket: no costs popping up at the last step.
  3. Optional codice fiscale: it only opens with the "Mi serve la fattura" (I need an invoice) box.
  4. autocomplete on every field: check it in the browser's developer tools.
  5. Well-tuned validation: the error shows up when you leave the field and disappears as soon as you fix it.
  6. Card details kept: after an error, the customer doesn't have to retype them.
  7. Legally compliant final button: "Acquista e paga", "Ordina con obbligo di pagare", "Pay now".
  8. Zero pre-ticked boxes: no insurance, warranties or donations ticked from the start.
  9. Digital wallets at the top: above the form, not at the bottom.
  10. Delivery before Christmas clearly stated: anyone buying on 27 November is already thinking about presents.
  11. Pasteable verification code: try copying and pasting it.
  12. Numeric keypad for card number and CAP: open the checkout on your phone and check.
  13. Labels always above the fields: never just grey text inside.
  14. Test on a real phone, on mobile data: switch off the Wi-Fi, head out into the street, complete an order.
  15. Test with a screen reader: VoiceOver on iPhone and Mac, NVDA on Windows; every field must announce its label and its errors.

And one more line, for anyone who's also got a shop on the high street: issue invoices from your management software, not from the website.

One last check: a form with eight perfect fields that loads four megabytes of script takes six seconds to respond on a phone on the move, and by the fourth second the customer's already gone. For page weight and speed, have a look at The beautiful website Google won't forgive: Core Web Vitals and the problem that started in Figma.

The last conversation

Back to the bill. The undeclared coperto in Naples, the service charge that pops up in London: it's not the couple of euros that stings. It's the feeling that someone waited until you couldn't get up any more before telling you the real price.

Online, the customer gets up whenever they like. They don't even have to call the waiter over: a thumb's all it takes.

By the end of November, your customer has already made it through the noise of the offers, compared, and put their trust in you. They've already said yes. Every pointless field, every hidden cost, every codice fiscale demanded from someone buying a present, every card wiped after a hiccup is an excuse you're handing them to take it back. Checkout is the last conversation before the handshake. Whoever asks the fewest questions wins.

Before Friday the 27th, try this: pick up your phone, turn off the office Wi-Fi and buy one of your own products as if you were a stranger. If you get irritated, if you have to zoom in to read a label, if you wonder what a piece of information is for, you've just found where you're leaving money on the table.

If you'd like us to run that check together on your shop, drop me a line. I'll say it up front, in the interest of transparency: I've got an obvious conflict of interest, because designing and fixing online shops is what I do for a living. But precisely because I've got my hands in them every day, between Caserta and London, I know where the forms that empty baskets are hiding.

And to be ready for the whole season: Back to work is won in August, Black Friday is won in September and The discount is decided on 28 October (not in November).

Note: the guidance on Italian and British law (Consumer Code, DMCC Act, Consumer Contracts Regulations, invoicing, accessibility) is up to date as of 12 November 2026 and is for general information only. It doesn't replace advice from a solicitor or an accountant on your specific case.

Sources

Data, cases and legislation cited in the article, in the order they appear. Legal information up to date as of 12 November 2026.

Frequently asked questions

What is an e-commerce checkout?

It's the journey that takes the customer from the basket to a confirmed payment: contact details, delivery address, delivery options, payment and the final summary. Just a few screens, but they decide whether the order comes through or gets left halfway.

What's the difference between the basket and the checkout?

The basket is the summary of the products chosen, where you change quantities and see the total. Checkout starts when the customer decides to pay and has to provide their details for delivery and payment. Delivery charges and the final total should already be shown in the basket, so the customer doesn't get a nasty surprise later.

Is the codice fiscale compulsory at checkout?

For online sales of goods shipped to private customers in Italy, no: an invoice is only compulsory if the customer asks for one, and the codice fiscale is only needed in that case. The right solution is a "Mi serve la fattura" (I need an invoice) checkbox that opens up the tax fields. For services, courses and business-to-business sales the rules are different and need checking with your accountant.

How much does it cost to improve a checkout?

It depends on the platform and how customised the form is. Plenty of fixes cost very little, though: autocomplete attributes, labels above the fields, an optional codice fiscale, the right wording on the final button. Deeper work, such as a quote flow for businesses or locks tied to order status, needs bespoke development.

How do you optimise checkout for smartphones?

With few fields, labels that are always visible, the right keyboards (type="email", type="tel", inputmode="numeric" for card number and CAP), digital wallets at the top and a verification code that can be pasted. The decisive test is buying from a real phone, with the Wi-Fi off, the way a customer out in the street would.

When does it make sense to make customers register?

Hardly ever on a first purchase. An account is useful when the service genuinely needs one, such as subscriptions, members' areas or online courses. Even then, it's better to take the payment first and offer registration on the confirmation page, with the details already filled in.

What tools do you use to check a checkout?

The browser's developer tools, to check attributes and fields; a real smartphone on mobile data; a screen reader such as VoiceOver on iPhone and Mac or NVDA on Windows; and a test payment with a deliberate error, to check the card details don't get wiped.

How to write a Christmas story that doesn't taste ...
Pay in 3 is credit: Italy joins the UK from 20 Nov...

Related Posts

 

Comments

No comments made yet. Be the first to submit a comment
Already Registered? Login Here
Tuesday, 17 November 2026

By accepting you will be accessing a service provided by a third-party external to https://insightadv.it/

Your weekly dose of Insight!

All our latest work. Every Monday in your inbox.

From Creative Writing to Photography, taking in Marketing and Design along the way. Every day we're busy prepping exclusive content to give you some fresh inspiration. Sign up so you don't miss a thing: every Monday at 09.30am you'll get the full weekly round-up. No spam. Just proper quality content, and you can unsubscribe whenever you like.

 

Satisfied customers

What they say about us

We started our collaboration with the InsightAgency, entrusting them with a strengthening of our Corporate Identity. They advised us on the creation of a new Brand and Logo and supported us in buil...
2013-10-02
Read more
Massimiliano Leone
CEO & Founder / Esigitaly srl
I have known Umberto for a few years and I can confirm that he is a professional in his field! Over time a good friendship has also been established which I am sure will last over time: it is prec...
2013-09-24
Read more
Pietro Sciannamblo
CEO & Founder / Esigaretta Italia
I have worked with Insight Agency for more than ten years in my previous role (Trade Marketing Manager of a well-known red mobile phone brand) and have always appreciated their honesty, speed and c...
2015-07-16
Read more
Giampaolo Moscardi
Titolare / Kar di Giampaolo Moscardi

About

Insight Adv Ltd is a full-service advertising agency. We offer our customers Graphic and Web Design, Marketing and Strategic Communication services.

We create websites, e-commerce and fad platforms, commercials and promotional videos and applications for smartphones and tablets. We also offer digital & direct marketing, social media and content management services. 

Pills...

" They say that advertising is the soul of commerce, and it's true — a product that isn't known is a product that doesn't exist. "
Stephen Littleword

Follow us on