The Associate App, one app for the whole shop floor.
- Associates
- 50,000
- Transactions
- 30% faster
- Associate NPS
- +25 points
- Kind
- Mobile point of sale, iOS
The till, the stockroom and the customer's history, in a phone in an associate's hand.
Everything a shop assistant does in a day, in one app: take the payment, look up stock in another store, order what is not on the shelf, and see who the customer is before they get to the counter. 50,000 people open it at work. I designed the first version, then hired the team that took it over and led it from there.
- How it was made
- 2017the year I designed the first version5 yrsholding the one app line against 90 companies1design system, which every NewStore app since sits on
Everyone else shipped three apps.
One to sell, one to fulfil, one to count stock. Associates paid for that in context switching, and retailers paid for it in training. I set the direction for a single app that did all of it, and then spent five years holding that line against feature pressure from ninety companies who each wanted their own screen.
Holding the line was the work. The first version was the easy part.
It was the right line. Sixty eight per cent of the associates who take payments in a month also do at least one other job in it: looking stock up, fulfilling an order, running the backroom. Two thirds of them would have been carrying a second app.
Five jobs, one app.
- Checkout
- On the floor, in a fitting room, at a pop up. Tap to pay, wallets, split payments, remote checkout.
- Clienteling
- Purchase history, preferences and direct messaging in hand, so an associate sells to someone they remember.
- Endless aisle
- Shelf items and network items in one cart, one payment, one receipt. Out of stock stops being a lost sale.
- Fulfilment
- An in-store order from pick to hand off without leaving the app, with Runner coordinating the backroom and the floor in real time.
- Inventory
- Counts, transfers and receipts done standing on the shop floor, rather than written down and typed in later.
The screen an associate lives in.
Scan it or search for it, build the basket, take the money. It is the first thing open at the start of a shift and the thing the rest of the app hangs off.
The total, at the size of the thing that matters.
One number, large, and nothing on the screen competing with it. This is the moment the sale is won or lost, so it does one job and the phone turns around to show it.
Selling outside the four walls.
What somebody bought, what they liked and what they said last time, in hand before the conversation starts, and a way to carry it on after they leave the shop.
Is it here, is it anywhere.
Stock in this shop, in every other shop and in the warehouse, answered on the floor rather than by walking the customer to a computer.
Every shop is a small warehouse.
Orders waiting to be picked, packed and handed over, queued by what is holding them up. This is the screen that turns a shop into part of the delivery network.
Nobody walks to a counter.
An associate who has to walk a customer across the shop to pay has already given back the part that made coming in worth it. The conversation ends, the queue starts, and the thing that was about to be bought becomes a thing being queued for.
So the sale finishes where the conversation happened: on the floor, in a fitting room, at a pop up stand. The whole design goal was to make the payment itself disappear, and the measure of that is the associate never looking at a terminal instead of a person.
Tap to pay on the phone itself, a card reader, or the customer’s own phone.
Wallets, cards, cash, gift cards, and a split across more than one of them.
Returns, exchanges and a digital receipt handled in the same place, in seconds.
A hundred and two seconds, median, from starting a checkout to a confirmed payment, across 4.2 million of them in a month. A third are done inside a minute. Thirty per cent faster than what it replaced, and the twenty five points of associate NPS are mostly this: it is the feature associates mention first, and the one nobody asks for by name.
It does something else that nobody asks for by name either. Paying inside the conversation is also the moment the shop can ask who you are, and the best retailers on it now capture a customer on nine transactions out of ten, against about three quarters of orders across the whole platform. Everything the clienteling screen does later is spent from that.
Three decisions, and what made them.
The price has to face the customer
In a lot of countries it is the law that the customer sees the price the moment it is scanned. The usual answer is a second screen bolted to the till, which is another object to buy, mount and break. We had one phone, so the price is drawn to be read from the other side: the associate turns the phone around and the law is satisfied by a gesture rather than by hardware.
We wanted two phones touching, four years early
Taking a card on the phone itself was the one thing the hardware would not let us do. We filed for it anyway in January 2018: a checkout where the customer’s phone and the associate’s phone authenticate between themselves, with a code scanned between the two because touching them together was not yet possible. It granted in July 2022 as US 11,379,813 B2, to NewStore, with five of us named on it.
Apple announced Tap to Pay on iPhone in February 2022 and we shipped it to retailers that September, among the first to do it. In iOS 27 Apple shipped Tap to Share, which is two phones touching to hand over a customer’s details. It took eight years for the thing we designed in 2018 to become something you could buy.
The cart is the shopping app’s cart, deliberately
The selling screen looks like the brand’s own shopping app because the customer had already trained the associate for us. Everybody who has bought anything online knows what a cart is. An associate is selling on the floor in about two minutes, and that is a training decision made in the interface instead of in a manual.
I designed it, then I gave it away.
The first version was my own work, as every NewStore product was at the start. As this one grew I hired product designers, handed each area over, and moved to setting direction and reviewing, which is a less satisfying job and the only one that scales.
It held together because I had drawn it myself first. I knew which requests were cheap and which would cost a year, and the designers who took it over inherited an argument rather than a style guide. The foundation became the design system every later NewStore app was built on.