"I'll finish it on the laptop later." I hear some version of that every November, from someone holding a phone with three items in the basket and a delivery form that keeps zooming in when they tap it.
Later is a lovely word. In a shop's analytics it mostly looks like an abandoned cart.
Black Friday falls on Friday 27 November 2026 and Cyber Monday on Monday 30 November. That leaves about two weeks, which is enough to fix a checkout, not enough to rebuild a whole shop in a panic. So this is a list of the things I would look at first, in the order I would look at them, all on a phone.
Why does "I'll finish it on the laptop later" usually mean never?
A shopper who leaves a mobile checkout to finish on a laptop has to remember the shop, find the same products again, rebuild the basket and still want the thing hours later, when the sale banner may be gone. Every one of those steps loses people, and the phone was where they had the impulse.
The phone is not a side door any more. Adobe's report on the 2025 holiday season found that 56.4% of US online transactions between 1 November and 31 December happened on a smartphone, up from 54.5% in 2024, and on Thanksgiving Day the mobile share reached 61.6%. Those are US figures, but I see the same pattern on shops in Europe: the browsing starts on the sofa, on a bus, in a queue.
The Baymard Institute's list of cart abandonment statistics, updated on 28 October 2026, puts the average documented abandonment rate at 70.22%, the unweighted average of 50 studies. Plenty of that is window shopping and nobody can design it away. Baymard also notes that bigger or more complicated purchases often shift back to a desktop instead of being completed on a phone, which is exactly the "later" in the title. Some of those people do come back. The ones you lose are the ones who hit a small, avoidable annoyance and decided the laptop could deal with it.
Here is the doubt I cannot resolve from the outside. Unless a shopper signs in on both devices, your analytics cannot tell you whether the phone visit at lunch became the laptop order at 10pm, and the public reports count devices, not people. What would settle it for your shop is a test: fix the mobile checkout, then compare the share of mobile orders in the Black Friday week with last year's. If it barely moves, the phone was never the problem.
Are the buttons big enough for a thumb?
The fastest way to find out is to try paying with one hand while holding a coffee in the other. If you have to pinch, zoom or aim, the tap targets are too small or too close together. WCAG 2.2, success criterion 2.5.8 sets a minimum of 24 by 24 CSS pixels for a target, or enough space around a smaller one, and I treat that as the floor, not the goal. Quantity steppers, the little remove cross and the "apply code" link are the usual offenders.
Then look at the keyboard. A card number field that brings up the full letter keyboard is a small insult every single time. Google's guide to payment and address form best practices on web.dev says to use type="text" with inputmode="numeric" for card numbers, rather than type="number", and to add the right autocomplete values (cc-number, cc-name, cc-exp and cc-csc) so the browser can fill the card details it has already stored. The same applies to name, email, address and postcode fields. When autofill works, a checkout that took two minutes of thumb typing takes twenty seconds.
Then count the fields. Form length matters more on a phone than anywhere else. Baymard counts 23.48 form elements on the average US checkout and puts the ideal at 12 to 14. I would go through the form and ask of every field: what breaks if I delete this? A second address line can be optional. A phone field, if you need one for the courier, should say why. "Company name" on a shop that sells to people at home can go. Be ruthless.
Show the total price and delivery before the last step
In Baymard's survey of US shoppers, once you set aside the people who were just browsing, 40% said they abandoned because extra costs such as delivery, tax or fees were too high, and 12% because they could not see or work out the total cost up front. Nobody enjoys finding a delivery charge on the final screen, and on a phone, where going back means scrolling up through a long form, it feels like a trap.
Put the delivery cost, or a clear "free delivery over a set amount", on the product page and in the basket. Show the running total before you ask for an address. If delivery depends on the address, say so in plain words next to the estimate and show the cheapest option first. On Black Friday, when people compare three shops in three tabs, the one that shows the real number earliest wins the comparison.
What should sit next to the Pay button?
Trust has to be visible at the moment of paying, because that is when people hesitate. In the same Baymard survey, 19% of US shoppers who abandoned said they did not trust the site with their card details. On a phone you cannot rely on a footer full of badges, because nobody scrolls down there mid-payment. So put the reassurance right next to the button.
What works is plain and specific. The payment methods you accept, shown as the familiar card and wallet marks, with Apple Pay and Google Pay offered where your payment provider supports them, because they skip most of the typing. A one-line note on returns, linked to the full policy. The delivery estimate again. A way to get help, such as an email address or a short form, for the person who has a question before they hand over their card. Things I would remove: a countdown timer that resets on reload, a pop-up asking for a newsletter sign-up halfway through the form, and anything that makes the page jump while the shopper is typing.
One more check, and it is the one most shops never make.
Many card payments now send the shopper to a banking app to approve the payment and then back to the browser. Try it on your own phone. If the checkout loses the basket or shows an error when the shopper returns from the bank app, that is the moment you lose a sale that was already won.
Do shoppers need an account to buy?
No, and they should not be made to. Baymard's survey puts "the site wanted me to create an account" at 18% of abandonment reasons once browsing is excluded, and web.dev's guide says to make guest checkout the default and offer account creation once the order is complete. Ask for an email for the receipt, take the order, then offer to save the details with a password on the thank-you page. People who have already paid are much happier to make an account than people who have not.
Test it on your own phone tonight
You do not need a lab for the first pass. Pick up your own phone tonight, ideally an older one too, and buy something cheap from your shop as a stranger would.
- Open the shop from a link inside Instagram or Facebook too, not only from a browser bookmark. We check every site we build in those in-app browsers as well as in Safari and Chrome, on real phones, before launch.
- Add a product, change the quantity and remove one, with one thumb.
- Go to checkout and see whether autofill offers your name, address and card, and whether the keyboard matches each field.
- Find the full price including delivery and note on which screen it first appeared.
- Pay, approve in the bank app, come back and check the order is there.
- Do it all once more on mobile data rather than Wi-Fi.
Write down every moment you sighed. That list is your brief for the next two weeks. If you also need the sale page itself to be ready, I wrote about preparing a Black Friday landing page that Google can find in time.
When is a fix enough, and when does the checkout need a redesign?
Most of what is above is a fix, not a rebuild. Adding autocomplete values, enlarging buttons, moving the delivery cost up and turning on guest checkout are jobs of hours or days on most platforms, and they are worth doing before 27 November even if nothing else changes. I wrote up the search box separately, in what to do when your shop's search says 0 results.
It is a design problem when the checkout fights the phone at every step: a theme that was built for desktop and squeezed down, a basket that reloads the whole page on every change, or a form spread across five screens with no sign of how many are left. Patching those two weeks before Black Friday means more risk, not less. Leave them. In that case I would fix the worst two things now and plan a proper redesign of the shop for January, when traffic is quieter. If you are deciding what a new shop should be built on, my guide to building an online store covers the choice.
Web design here starts with the checkout on a real phone, because that is where the money changes hands. If you want a second pair of eyes before Black Friday, send the address of your shop to contact@magnuszgr.com and we will tell you which of the fixes above apply, and whether anything bigger is worth planning.
Your laptop shoppers will forgive a clumsy checkout. The phone shoppers just close the tab.

