How Castore transformed peak trading performance with a retail-first iPaaS

The challenge: why peak trading exposed hidden risk in Castore’s retail stack

Castore operates in one of the most demanding retail environments imaginable: high SKU turnover, extreme seasonality, and simultaneous trading across dozens of Shopify storefronts. During peak trading, pricing accuracy is mission-critical — yet it was also one of the most fragile parts of the operation.

Historically, retail pricing was prepared by merchandising teams using spreadsheets and SharePoint. Those files were then manually imported into Netsuite each evening using CSV uploads. Two team members would spend six to seven hours every night nursing those imports through — often locking the rest of the business out of NetSuite from late afternoon onwards.

Even then, success rates sat at 60–70%, with errors accepted as an unavoidable part of peak.

“We were in a place where a 30–35% error rate was considered acceptable — simply because the process was so long and so manual.” — Andy Richley, Head of Tech, Castore

Why ERP-led pricing couldn’t scale during peak trading

From a systems perspective, the problem wasn’t NetSuite itself — it was how it was being used.

Pricing logic didn’t live in the ERP. Validation didn’t live in the ERP. Reporting didn’t live in the ERP. NetSuite was being used primarily as a transport layer, simply because it sat in the middle of the architecture.

Once prices were finally imported, Patchworks would then pull that data back out of NetSuite and distribute it across 35 Shopify stores. This meant NetSuite APIs were under heavy load for hours at a time, pushing concurrency above safe thresholds and creating a constant risk of performance degradation — or forced licence upgrades.

For Andy, the red flag wasn’t just technical.

“When peak relies on people working late into the night just to keep systems moving, that’s not a scalable or healthy place to be.”

🤔 Did you know? Patchworks is a Shopify Certified tech partner

The turning point: rethinking how pricing data should flow

The breakthrough came from asking a deceptively simple question:

Why does pricing need to go through the ERP at all?

Castore already operated a mature Snowflake data lake. Core product data, publishing flags, and channel context were all available there. Merchandising logic lived upstream. The ERP remained the source of truth for base pricing — but it wasn’t adding value to day-to-day trading execution.

Rather than rebuilding systems, the team chose to remove unnecessary complexity.

“It wasn’t about replacing NetSuite. It was about being clear-eyed about what it actually needed to control — and what it didn’t.”

The solution: using Patchworks as a strategic integration layer

Castore redesigned pricing execution around a cleaner, more purposeful data flow:

•  Pricing updates are ingested directly into Snowflake from SharePoint
•  Validation, exception handling, and margin checks are surfaced via BI dashboards
•  Formal approval is built into the workflow before any data is released
•  Patchworks runs a scheduled SQL query against Snowflake and deliver approved pricing directly into Shopify using high-performance GraphQL connections

Patchworks operates quietly in the background — executing governed data flows at speed — while also enabling a broader architectural rethink.

“What surprised me was just how effortless it was. Once the data was prepared properly, Patchworks moved it into Shopify incredibly fast.”

In addition to executing pricing directly into Shopify, Castore also uses Patchworks to populate Snowflake. Patchworks can either deliver raw JSON data for downstream processing or, as implemented here, populate Snowflake with fully structured, final-form data. This ensures that pricing and trading data arrives in the data lake exactly as it is intended to be consumed — consistent, governed, and immediately usable by BI and trading teams.

By shaping the data in-flow rather than downstream, Castore reduced complexity and eliminated ambiguity around definitions and transformations.

Snowflake became more than an analytical destination; it became an operationally trusted reflection of live commerce activity. Patchworks acts as the orchestration layer between execution systems and the data platform, enabling speed during peak without sacrificing control or clarity.

Results: faster pricing, lower risk, and peak-ready scalability

Operational performance gains

•  Pricing updates reduced from ~5 hours to 2–3 minutes
•  Worst-case execution time during peak: 6.5 minutes
•  Error rates dropped from 30–35% to 1–2%
•  Zero-error days achieved during peak trading

Commercial and risk impact

•  NetSuite concurrency remained stable throughout peak
•  No emergency licence expansion required
•  Formal approval workflows removed pricing fraud risk
•  Fewer customer service issues caused by mis-priced promotions

Team productivity and leadership confidence

•  Manual overnight pricing work eliminated
•  Data and ecommerce teams shifted from execution to oversight
•  Trading calls evolved from firefighting to reassurance

“Halfway through peak, our IT update on trading calls became: ‘no problems to report.’ That was new.”

Beyond peak: operationalising the data lake for future growth

The impact extended well beyond a single peak trading period.

What had previously been an “aspirational” data lake became an operational asset. Confidence grew — not just in the systems, but in the organisation’s ability to scale without adding fragility.

“Now the conversation isn’t ‘should we use the data lake?’ — it’s ‘which use case do we tackle next?’”

For peak 2026 and beyond, Castore enters planning with something rare in retail technology: headroom.

Why Patchworks can help you

Patchworks is a retail-first iPaaS, designed for the realities of modern commerce: multi-store Shopify environments, complex data flows, and extreme peak-trading pressure.

By enabling clean, high-performance integration between systems like Snowflake and Shopify, Patchworks helps retailers simplify execution, reduce risk, and scale with confidence — without forcing costly platform rebuilds.