First to Know, Not Last

A customer knew before we did. That's the sentence every Ops director dreads hearing back from their own team. Not because it's rare, but because it's common enough that most people in this role have lived it at least once.

In March 2024, McDonald's tills and kiosks went down across multiple countries. Customers were posting about it and queuing at closed counters before there was any public statement on what had happened. The next day, Sainsbury's experienced issues with contactless payments, SmartShop, and Nectar, while chip and pin continued to function. Shoppers were the ones piecing together which parts of the system were down and which weren't.

The visibility gap between the public experience and the internal knowledge is where trust erodes fastest. Once a customer is the one explaining your outage to you, you've lost the narrative—and usually the queue as well.

Here's the test I'd put to any Ops team running a distributed estate: not 'did something break?' Something will always break across enough sites eventually. The real test is: did you catch it in the signal, or did you catch it at the till?

Those are two different companies to run. One has a customer standing in front of a dead card reader, working out for themselves what's gone wrong, then telling a member of staff who has to relay it upward before anyone with the authority to fix it even knows there's a problem. The other has already seen the signal degrade, already knows which site, which system, and roughly what it's costing per hour it stays down.

The difference isn't about detection speed; it's about who's doing the noticing. A customer noticing on your behalf means your monitoring has already failed, whether or not the underlying system comes back up in five minutes or five hours. Fast detection after that point is damage control, not operational visibility.

Distributed retail estates generate a constant stream of small signals long before anything is bad enough for a customer to stand in front of it. A payment terminal that's slower than usual, a kiosk that's dropped connection twice this week, refrigeration drawing power outside its normal pattern—none of that is dramatic, but all of it is where the story usually starts.

The operators who avoid the 'customer knew first' moment aren't the ones with fewer things going wrong. They're the ones who've built the habit of watching the estate as a whole, not waiting for a specific site to escalate a specific complaint. The customer at the till shouldn't be your alert system.