For fast-growing retailers, peak trading performance is determined by how quickly and accurately pricing, promotions, and product data can be executed across increasingly complex commerce stacks.
This was the challenge facing Castore, a global performance sportswear brand operating 35 Shopify stores simultaneously across international markets. With high SKU turnover, extreme seasonality and intense promotional pressure, pricing accuracy during peak trading was business-critical.
By rethinking how pricing data flows through its retail architecture and deploying Patchworks as a retail-first integration platform as a service, Castore reduced pricing execution times from hours to minutes, cut pricing errors by more than 90 percent and stabilised peak trading without adding operational strain. Once the pricing execution approach was agreed, the solution moved from concept to live in days rather than months. The workflow was designed, tested, and deployed in under two weeks, without the need for bespoke development or extended change cycles with Patchworks ‘low code no code’ ability.
Peak trading exposed hidden risk in Castore’s retail stack
Historically, pricing at Castore was prepared by merchandising teams using spreadsheets and SharePoint. Each evening, those files were manually imported using CSV uploads. During peak trading periods, two team members routinely spent six to seven hours overnight managing those imports, often locking the wider business out of systems from late afternoon onwards.
Even with that level of effort, success rates during peak trading sat at just 60 to 70 per cent. Pricing error rates of 30 to 35 per cent were treated as an accepted cost of doing business.
“We were in a place where a 30 to 35 per cent error rate was considered acceptable simply because the process was so long and so manual,” said Andy Richley, Head of Tech at Castore. “When you are spending most of the night just trying to get pricing through, you stop asking whether the process itself makes sense.”
The risk extended beyond technology. Peak trading success depended heavily on people working late into the night to keep systems moving, creating fatigue, operational fragility and limited confidence in the organisation’s ability to scale
Why ERP-led pricing execution could not scale
From an architectural perspective, the issue was not the ERP itself, but how it was being used. Pricing data was routed through it because it sat at the centre of the architecture. Once pricing data was finally imported, it was then pulled back out and distributed across 35 Shopify stores. This placed sustained load on APIs for hours at a time, pushing concurrency beyond safe thresholds and introducing the risk of degraded performance or forced licence expansion during peak periods.
“What became clear was that we were using a system of record as an execution engine simply because it was there,” Richley said. “That’s a very different role, and it’s where the cracks start to show under peak conditions.”
Rethinking how pricing data should flow: a cleaner execution model using Patchworks
According to Jim Herbert, CEO of Patchworks, Castore’s results highlight a broader shift in how retailers are reassessing integration performance as a strategic capability rather than a background function.
“Like many fast-growing retailers, Castore had gradually pushed more execution workloads into its core systems of record,” Herbert said. “Pricing data was being validated, transported and executed through the same platforms, creating unnecessary coupling and exposing peak trading operations to risk. The issue was not governance, but execution.”
“By separating pricing governance from pricing execution, Castore was able to leave its core systems doing what they do best, while moving high-volume data movement and orchestration into a retail-first iPaaS,” Herbert added.
Measurable gains during peak trading
The impact during peak trading was immediate and sustained. Pricing execution times were reduced from approximately five hours to two to three minutes, with worst-case execution times of 6.5 minutes during peak periods. Pricing error rates fell from 30 to 35 per cent to just 1 to 2 per cent, with zero-error days achieved during peak trading.
Operationally, concurrency remained stable throughout peak periods, with no emergency licence expansion required. Formal approval workflows reduced pricing fraud risk, and customer service issues caused by mispriced promotions were significantly reduced. Manual overnight pricing work was eliminated entirely, allowing data and ecommerce teams to shift from execution to oversight. Trading calls evolved from reactive firefighting to confirmation of stability.
“What surprised me was just how straightforward it became,” Richley said. “Once the data was prepared properly and approved upstream, Patchworks moved it into Shopify incredibly fast. That was the moment we knew we had the right execution layer.”
Share