Peak trading exposes technology weaknesses at precisely the moment when retailers have the least room for failure
Carl Andrews, MBA, CEO, Original Software
Software that runs perfectly well in October can start to buckle in November. That’s rarely because peak trading breaks anything new. More often, it exposes weaknesses that were there all along, quietly tolerated during quieter months. It’s telling that only 48% of enterprise IT leaders say they’re fully confident their testing would catch critical issues before they hit customers.
For the other half, peak trading is a live test of systems nobody was sure would hold up.
For most of the year, retail systems operate with a fair amount of slack. Promotions change infrequently, so stock levels rarely move fast enough to expose timing gaps between systems.
That slack disappears during Black Friday and Christmas. Volumes spike, pricing changes hourly, and fulfilment runs close to capacity while stock shifts unpredictably across channels.
Three-quarters of enterprise IT leaders say software updates are now happening faster than they can test them, and nowhere is that gap more exposed than in the run-up to peak trading, when pricing and promotions can change several times a day.
A pricing engine that mishandles discounts under load gets found out. So does an inventory feed that lags behind real sales, or a fulfilment process built for calmer conditions, all at the exact moment the cost of failure is highest.
Most peak-readiness testing still asks the wrong question. It checks whether a site stays up, not whether it’s correct. A site can load fast and handle heavy traffic while customers are shown the wrong promotional price, told an item is in stock when it isn’t, or hit with a payment that fails for no reason.
None of that takes the site down. It still costs revenue and trust, at the worst possible time. Sometimes the infrastructure gives way: Australian retailer Harvey Norman’s website buckled under Black Friday 2023 demand, unable to handle visitor volume and leaving customers in a virtual queue tens of thousands deep for over ten hours, at exactly the moment the retailer needed it most.
Given limited time before the trading surge, where should testing effort go? Pricing and promotions deserve close attention, particularly where multiple offers, discount codes and loyalty schemes interact in ways only obvious once tested together under realistic conditions. Inventory accuracy across channels matters just as much.
Retailers running combined online and in-store stock pools are especially exposed, since a lag in updates can mean orders get accepted for stock that no longer exists.
The fulfilment chain, from order confirmation through to dispatch, is worth scrutinizing too. It’s often the least tested part of the system, sitting across multiple platforms and teams.
Conventional performance and availability testing won’t catch most of this. Load testing confirms a system stays up under pressure, but it has little to say about whether the right discount applied, whether stock was correctly reserved, or whether an order reached the warehouse system in usable form. Those are process questions, not infrastructure ones. A retailer can pass every performance benchmark and still find, on Black Friday morning, that a promotion is stacking incorrectly or that a batch of orders never made it to the fulfilment team.
This is where business users add something technical testing tends to miss. Technical teams test systems the way they were designed to work. Merchandising staff, customer service teams and store colleagues tend to use systems the way people actually use them, edge cases and workarounds included. Think of a merchandiser adjusting a promotion at short notice, or a store colleague processing a return against an online order, or a customer service agent trying to amend a delivery address after the order’s gone through.
These are ordinary moments, and they’re often where a process that looks fine on paper turns out not to work end to end.
Peak readiness comes down to whether the processes sitting on top of the technology hold up under the pressures of the season. That means testing the journey rather than isolated steps: applying a promotion to a real order, checking stock reserves, and confirming it reaches fulfilment intact, with the people who’ll run these processes involved before volumes arrive.
There’s only one run at Black Friday. The retailers who go in with confidence tend to be the ones who found their weak points in September, not the ones hoping the systems hold.
