The Invisible Till in luxury

I've been thinking about the till a lot lately. Where it actually sits in the life of a retail business, and — more to the point — what it's for in luxury.

Why now? Because 2026 is the year I'm pulling out a piece of legacy. The same till that, across eight or ten years, started eating an absurd amount of engineering time. Patches, extensions, features, all of it, all the time.

Look at where checkout systems have gone in the last three or four years and two directions stand out. The first is self-checkout. Trendy, and it doesn't stick everywhere. It's crept into apparel a little, footwear less so. In grocery it landed perfectly — a real, big shift. There it belongs, and you see it everywhere, just scaled down.

The second is the all-in-one-on-the-go format. The till as minimum setup, maximum everything-in-one-box — the fiscal unit and the till software inside a single product. Buy it, plug it in, almost nothing to fuss over. Every integration, every module you need to sell is already switched on. It's a response to small businesses suddenly needing tills: the law won't tolerate trade without them anymore, which is fair. Standing up a halfway-sane till used to take real effort — a separate fiscal unit, support for it, till software, all made to agree with each other. It could balloon into its own project. Now you buy one thing. Scanners, fiscal unit, the till itself, the handoff to the tax authority. Give it 220 volts and internet and it does the rest. No cashier training required.

Neither of these fits luxury. And I'm an active part of luxury retail, so I notice — it has its own patterns that ordinary retail doesn't. Most of the process is identical, sure. But luxury is luxury precisely because of the small things that change nothing globally and everything locally. Self-checkout doesn't fit it. "Everything in one box" doesn't fit it either.

Which is logical, if you believe a till is just a way to close a sale cleanly and inside the law. My own experience ran the other way. That old legacy, in one of the countries I worked in, had its till software stretched way past that — my predecessors bolted "till plus inventory plus stock movements" and who knows what else onto the register, most of it duplicating functions that already lived elsewhere. And that's the real question. In luxury there's no special reason to keep the till as a pure receipt-punching device. But there's every reason to grow clienteling through it — to build and evolve the client journey.

There are ways to do that. Not many, but they exist. And here's my core view: in luxury retail the till is not a receipt printer. It's the seller's instrument for the client relationship. The rest of this is how I'd build one.

Keep the core with the vendor

Before the "why," a stop-factor — especially if you run IT. CIO, CTO, any head of IT.

Keep the direction of the law in the back of your mind, always. The receipt is, roughly, 95% of all B2C sale records. B2B is going electronic — e-invoicing, e-document flow — and everything drifts that way. Receipt and labeling rules keep tightening alongside it, and any country that reaches a certain technical level and wants to collect tax better will bring in stricter measures. So the stop-factor is this: your till has to have a core — the receipt-punching engine — and that core has to be supported by the vendor.

If at some point you decide to pull the system off vendor support and build your own, remember that receipt policy only ever gets stricter. You'll be spending on the till practically every time the state turns the screws. The limit is simple: the core stays with the vendor — don't let the receipt engine itself become your code. Do that and you step onto a treadmill, re-implementing the same thing year after year, one requirement then the next, while your legacy swallows a huge slice of the budget.

So bolt on whatever extra modules you want, but leave the core alone. Not every product lets you work this way, but plenty of modern ones do: a single canonical product catalogue, receipts punched against it, and everything you invent — opening a browser page, calling the cleaning crew from a button — added as separate modules. Never drop vendor support just because the commercial department wants a button on the till that files a ticket in a task-tracker so an engineer shows up. It isn't worth it. There are smarter ways. Keep the core intact and secure.

Simple enough for a five-year-old

Say we've found a system like that — its own core, but open to extension, by us or by a vendor-contractor. So what does luxury actually need from it?

First thing, before anything else: the cashier needs a beautiful, comfortable interface. So many people skip this. We've all seen the old grocery tills, tuned purely for speed — those keyboards where one key is "promo," one is "plus," one is "minus." There was no interface to speak of. In any retail where you work with a person standing in front of you — luxury clothing, footwear, electronics, anything where accuracy matters more than how fast you scan — you need a good UI, and it has to be intuitive. Get that right and you can nearly forget about training cashiers.

Nobody likes training. Honestly. There's no golden key that teaches everyone at once and ends the questions — it doesn't exist. Write a stack of perfect instructions and no one reads them. Run good webinars, teach everyone in person, and then staff turn over, and you're either training the new ones again or teaching in batches — and in a batch, someone stays half-trained. The questions come either way.

Most of them get answered by the technological context itself. Everyone has a smartphone, everyone uses apps, everyone is trained on certain patterns already. Someone who's never once stood at a till reads it as an app on their phone. They already know an X in the corner means close; a button lit up in color means it matters; red means stop, cancel. Go into your banking app and start a transfer — there are two or three buttons, "send" and "cancel." Now picture them with no color at all. You'd hesitate for a second. People carry a formed sense of UX around with them. If the till respects the ordinary patterns of good UI, you kill a huge number of questions before they're asked, and you skip writing tons of instructions nobody reads.

So before you ship any new feature, ask how sound it is as UX. Sound enough that it needs a manual to be usable at all? BI, CRM — fine, that gets genuinely complex, instructions earn their place there. But the till should be simple enough that you could sit a five-year-old down, say "here's the scanner, here's the barcode, scan," and they ring up a receipt. If they can't, the till isn't adaptive enough and the UX is built badly. That's it. That's my honest read.

Make it a clienteling tool

Good UX buys back mountains of support time. The next pillar is what the till is actually for.

What can a till even do for luxury? Nothing, if it just prints receipts while, off to the side, a separate system holds your clients, your stock, your availability, and you look people up by phone or email to see what they were into. You can run it that way: till for receipts, push them into the merchandise system, then BI, then the global dashboards, and keep the whole luxury client story elsewhere. It works. I just prefer the version where the till carries its full ordinary toolkit itself.

Start from the simplest operations, a sale and a return, then add what luxury specifically needs. Mandatory client-data capture, for one. The till should push the seller, by any means, to ask the client for something — any digital trace, a phone number, an email. There are ways to coax that out of a client, and they're worth building in.

And a single omnichannel base. Across every channel, and — this part matters — shared across every store. Picture the standard setup where each card is a thing unto itself: talks to nobody, knows nothing about what happened at other locations. That's bleak. You shop at one mall, buy something, walk into the store across town, and there they've never heard of you. A stranger, a new client, none of your preferences on file. So: every till wired into one whole, one record of purchases and receipts, everywhere. Ideally the seller types in your phone, or any identifier, and pulls everything they need — preferences, purchase history. Say you came in back in 2023, bought regularly before that, then nothing in our stores from 2023 to 2026. That's odd. That's a reason to start a conversation, a reason to let the client feel they're part of the brand, part of the story, that they're remembered — three or four years on, the history is still there and the seller can reach it.

And we're not talking about cashiers. We're talking about full consultants, sellers, who happen to also ring up the sale inside one tool. In luxury there are no pure cashiers. Anyone at the till is a seller first, and their job isn't to master the fine print of cash discipline — it's to know the client journey and to sell. Which is exactly why the tool has to be as simple as it can be. Don't burn a person's working memory on cash-discipline trivia, error types, and how to route around them.

Keep it invisible

On errors — this one matters. Any decent modern till handles its own. If your cashiers have to recognize errors and know how to work around them, that's bad. In luxury it's unacceptable. You're filling their head with data they don't need.

The other thing I care about is stock and products — a certain visibility. It's genuinely good when you can scan an item right at the till and see its card, the basics. Keep it in proportion, though. A store might have exactly one till computer and one in the back office. Every machine is worth its weight in gold, so you cannot overload the till. But you do want to hang small touches on it, and those are specific to each business — checking a task-tracker like Jira or YouTrack to see when your ticket for some fix gets closed, say. Bolt it on if you need it. And in 2026 a till simply has to be able to open a web app. Everything runs through the browser now, and a till that can't open one in a single window and behave is a sad thing.

As for the physical side — how compact, touch or not — for luxury I decided it barely matters. A customer-facing screen, visual merchandising on a separate 13-to-14-inch display turned toward the client with their basket glowing on it — that's mass-market. In luxury the average ticket is high and the client shouldn't be re-checking their basket mid-purchase; it reads as clutter. Hide the extra. A 15-to-20-inch monitor is about right, touch or not. Touch ideally — but I wouldn't drop the keyboard and mouse. All touch doesn't work; all keyboard-and-mouse doesn't work. Simple scanning is easier on touch. The moment you let the seller move into CRM and the like, keyboard and mouse are irreplaceable.

So for luxury I've settled on loading the store's interface as little as possible. Make the till unobtrusive, tuck the machine away, keep it all spare. The system should close real needs, not just punch receipts. Then everything around the client journey layers on top: sign-in and pulling up clients, their data, stock at other locations so you can tell the client when it'll arrive, and — if there's an online store — a comfortable way to handle web orders. Running through all of it: fault tolerance and as few errors as possible. And if an error does surface, let it be informative and obvious to the person reading it. Otherwise your support team spends forever teaching every new hire how to use the till.

Which is the whole point. In luxury I hold to one line: the till should be invisible to the client. It works only as a support to the seller — who, leaning on it, runs a clean and beautiful sale.