The Store Is Retail's Hardest Edge in Tech
In a single store, store technology is a small experiment. Across a fleet, every decision multiplies in cost and risk. That is what makes the store hard.
A screen on a store wall costs around 1000 euros. In a single store that is a small experiment. In 400 stores it is 400,000 euros and a commitment for several years. That difference is where many people misjudge retail technology.
The common view is that the store is the low-tech part of retail, the place where new technology arrives last. My view is the opposite. The store is the hardest part of the stack to get right, mainly because of what happens when you scale across a fleet.
Retail is ahead, not behind
The idea that stores lag behind is wrong. Retail, and fashion retail in particular, is one of the fastest adopters of in-store technology.
You can see it in the spend. Worldwide retail technology spending is heading towards 388 billion dollars by 2026. Digital signage on its own is a market of more than 30 billion dollars, and retail is the largest part of it. Within retail, fashion and apparel leads. The fitting-room screen, the electronic shelf label, the RFID tag that tracks every garment: most of these arrived on the shop floor first, often in apparel, because that is where the pressure to stand out is highest.
So the store is not behind. It is often where new technology is tried first.
Everything multiplies by store count
This is the point that gets underestimated. Whatever you put on the floor, you do not do it once. You do it for every store, or for a subset of stores depending on channel or brand.
It sounds obvious, but the numbers add up fast. The 1000-euro screen becomes 400,000 across a mid-sized fleet. A small change to the checkout flow becomes a coordinated rollout to every location. The public examples show the scale. Walmart is rolling out electronic shelf labels to around 2,300 stores. Morrisons is fitting about 10.8 million labels across 497 stores. Co-op is moving from 1,500 stores to all 2,400 by the end of the year. Each of these started as one per-store decision.
So on the floor there is no small decision. A cheaper bracket, a different cable, a slightly heavier device: you make that choice hundreds of times. The multiplication that helps you when something works also hurts you when something is wrong.
One store is not four hundred
The second problem is the gap between a pilot and a full rollout.
A proof of concept in one flagship store usually goes well. Good network, motivated staff, someone from the project team nearby. It works, everyone agrees, and the pilot is approved. Then it has to work everywhere.
Be mindful of the other side of this. I have been in a situation where there were so many proof of concepts that nobody could tell me what was running where. Every few months we would hear about another one, and it would surprise all of us. Often the people who set it up had already left the company. It took several years to clean this up and move everything back to standard.
Real stores are not like the pilot. Layouts change often, because they are meant to: new collections, seasonal resets, refits. The network ranges from fibre to an old and slow line in an older building. The busiest Saturday of the year puts a different load on the system than the quiet Tuesday you tested on. The hardware is not the same everywhere, because you bought it in waves over several years. The pilot proves the idea. The rollout is the real work, and it is much harder.
The purchase price is only part of the cost
There is also the cost that does not show up on the slide.
The price of the hardware is only a part of it. The rest is installation, maintenance contracts, field service, spare parts, and the simple fact that someone has to be on site in hundreds of buildings to install the device and to repair it when it fails. A change in the cloud is a deployment. A change in the store is a van driving to each location.
It does not stop after go-live. Store hardware ages on a clock you do not fully control. Payment terminals have PCI security certificates that expire, and after that the device can no longer legally accept a card. The expiry date sets the timing, across the whole fleet, whether it suits you or not. Most retailers replace this hardware every 3 to 7 years, usually region by region so the business keeps running. The pilot is the visible part. The recurring cost underneath it is the reason the store is the hardest edge.
What this means
To be clear, I am not against in-store technology. It does need to be planned properly, with the full fleet in mind and not just the demo.
A few things help:
- Ask what the cost and the risk become at full fleet size before you start the pilot, not after.
- Budget for installation, service and replacement, not only the purchase price.
- Assume the layout will change and the network will be worse than you hope.
Teams that do this build technology that works across hundreds of real stores. Teams that only look at the flagship pilot get a good demo and a hard rollout.
In my view the store is not the slow part of retail technology. It is the part where ideas meet real scale and real operations. That makes it the most demanding area to work in, and for me the most interesting.
Sources: Forrester US Tech Forecast 2026 for Retail; WWD/Sourcing Journal (retail tech spend); Grand View Research (digital signage, ESL); Technavio (retail digital signage); SNS Insider via GlobeNewswire (ESL deployments); YCR and SZZCS (POS lifecycle and PCI). Full links in 01-hardest-edge.md.