How a 14-location specialty retailer replaced a broken POS stack and launched Shopify POS in 6 weeks.
Northgate Outfitters (name changed for confidentiality) had been selling outdoor gear and lifestyle apparel across western Canada for over 20 years. By the time they contacted Kaizen Commerce, they had 14 stores operating on Lightspeed Retail POS and a WooCommerce site managed by a freelance developer. On paper, both systems worked. In practice, they had not meaningfully talked to each other in years.
The ecommerce store showed inventory that was wrong. The POS showed inventory that was sometimes right and sometimes not. When a customer asked a store associate whether they could ship an item from another location, the answer was always a guess. Transfers between stores were tracked on a shared Google Sheet. Receiving was done by one person who printed a report, walked the warehouse, and updated numbers by hand.
The company had grown to a point where the workarounds that got them to 14 stores were actively costing them money. A senior buyer estimated that the business was losing at least 3–4% of potential revenue to inventory mismatches and customer service failures caused by unreliable data.
"I'm spending more time managing the systems than managing the business. We've got 14 stores, a decent online store, and I still can't tell you with confidence what we have in stock right now."
— Founder, Northgate Outfitters
The Blueprint diagnostic — a 7-day paid discovery and architecture engagement — confirmed what the founder suspected. The problem was not the tools. It was the architecture. Five distinct operational failures were all running at the same time:
Lightspeed managed in-store inventory. WooCommerce managed online inventory. A third-party plugin was supposed to keep them in sync. It did not — at least not reliably. When the plugin failed silently, no one knew for days. The result: oversells online, confused associates in-store, and a customer service team that spent significant time each week manually reconciling orders and issuing apologies.
A customer who had shopped with Northgate online for three years had no profile in Lightspeed. A loyal in-store customer had no record in WooCommerce. There was no email capture at POS — associates had to ask manually and type names into a paper log. The marketing team was running email campaigns off a list that was months out of date and missing a significant portion of active buyers.
Buy-online-pickup-in-store existed as a policy. It did not exist as a workflow. When an online order came in with a pickup note, an email alert went to the store manager. The manager checked by hand. If the item was there, they called the customer. If it was not, they called the customer to apologize. There was no system. There was a person.
Every store had slightly different hardware configurations — a mix of Lightspeed-issued iPads, tablets purchased independently, card readers of varying ages, and receipt printers that may or may not work on any given morning. Two stores had no barcode scanners at all. Associates in those stores navigated by searching product names.
WooCommerce was hosted on a server the original developer had set up in 2019. Updates had been skipped to avoid breaking things. The site was slow, not mobile-optimized, and running two outdated plugins that the current freelancer did not want to touch. The founder had been quoted $18,000 to rebuild it by a previous agency — a quote that did not include POS, inventory, or any back-office changes.
The team modelled the cost of continuing on the current stack for another 12 months. Conservative estimates put the combined cost — lost revenue from inventory mismatches, staff hours spent on manual coordination, customer service failures from BOPIS breakdowns, and ecommerce underperformance — at over $85,000. The Blueprint paid for itself before the engagement ended.
The engagement ran in three parallel tracks. The most important thing Kaizen did before touching a single piece of data was design the target state — specifically, which system owns which data, and what the boundary between Shopify and anything else should be.
| Phase | Timeline | Focus |
|---|---|---|
| Blueprint | Weeks 1–2 | Architecture + risk mapping |
| Migration | Weeks 3–7 | WooCommerce → Shopify |
| POS Rollout | Weeks 4–10 | 14 locations, pilot-first |
| Hypercare | Weeks 11–16 | Stabilize + optimize |
The WooCommerce catalog had 4,200 active SKUs and another 2,800 legacy SKUs that had not sold in over three years. The first decision was to separate them: active SKUs would migrate to Shopify, legacy SKUs would be archived externally. This single decision reduced the migration scope by 40% and eliminated a significant source of catalog noise.
The catalog audit uncovered 340 products with missing or broken image URLs, 120 duplicate SKUs caused by years of manual entry, and 18 products with more than 3 variant options — Shopify's structural limit. Each was resolved before the first import wave ran. Kaizen ran every import through Matrixify, using dry-runs before every live import and pilot subsets before full waves.
Customer migration required a deduplication pass that found 2,400 duplicate email records — roughly 12% of the total customer base. These were consolidated before the import wave, and source system IDs were stored in Shopify metafields for traceability. 14,000 URL redirects were mapped and loaded before DNS was touched.
The location model design was the most important POS decision. Northgate had 14 selling stores, one central warehouse, and two overflow storage locations that had been incorrectly configured as selling locations in Lightspeed — meaning they were showing up as available stock when the inventory there was not sellable.
Kaizen designed the Shopify location model with 14 selling locations, one non-selling warehouse location, and clear rules for inventory allocation and transfer. Inventory was loaded from a freeze-window export taken on a Sunday morning before the pilot store went live.
Every location went through a pre-launch readiness checklist: products confirmed published to POS, barcodes validated, card readers paired with test transactions processed, smart grid designed, staff permissions tested by role, and a full test sale and return completed before the first real transaction.
Store 7 — the Kelowna flagship — was selected as the pilot. It had the most experienced store manager, a stable internet connection, and enough transaction volume to surface edge cases quickly. The pilot ran for two weeks before the broader rollout began.
The pilot surfaced three issues: variant navigation on the smart grid required too many taps (redesigned from 5 taps to 3), manager override permissions were too permissive (corrected before other stores went live), and receipt printer Bluetooth dropouts under high Wi-Fi traffic (switched to USB). All three were resolved before the broader rollout.
Staff training was built by role. Associates received one 90-minute session covering the standard sale, product search, discount application, and receipt options. Managers received a separate session on exception workflows. Each store received a two-page laminated quick reference card.
"The pilot surfaced real problems in a low-stakes environment. By the time we rolled out to the other 13 stores, the team felt like they already knew the system. The training landed because there was nothing surprising left."
— Operations Director, Northgate Outfitters
At the end of the engagement, Northgate Outfitters was running a single unified commerce platform across all 14 stores and their online channel.
For the first time, a sale at any store — online or in-person — decremented inventory in real time. Associates stopped guessing whether a product was available at another location. The buyer who had been spending 4–5 hours per week on manual inventory cleanup now spends under an hour.
Buy-online-pickup-in-store moved from an improvised process to a native Shopify workflow. Associates mark orders ready, customers receive automated notifications, and orders are fulfilled without a phone call. In the first month post-launch, Northgate processed over 180 BOPIS orders without a single customer service failure.
Shopify POS prompts for email opt-in at checkout — automatically for card payments, with a pop-up for cash. In the first four weeks after full rollout, Northgate captured over 4,400 new customer email addresses in-store. The marketing team now has a growing, high-quality list tied to real purchase history.
The most consistent feedback from store managers: staff moved faster. The smart grid eliminated the need to search for common products. The permission model meant managers were not being pulled to override routine transactions. Associates who had been nervous about the change were, within two weeks, more comfortable on Shopify POS than they had ever been on Lightspeed.
Real-time reporting across all 14 locations, accessible from a phone. Inventory by location at any time. Cross-channel customer profiles showing both online and in-store purchase history. The founder described this as the first time in five years that he felt like he knew what was actually happening in the business on any given day.
The most important decision was made before any data moved: designing the target architecture before touching the migration. Every downstream problem — from inventory accuracy to associate training to BOPIS design — becomes harder to solve if the platform model is wrong. Kaizen's Blueprint phase identified three structural decisions that, if made differently, would have created expensive problems during cutover.
Two months after the POS rollout was complete, Northgate returned with a new problem. The migration had exposed something the Lightspeed stack had been masking: the purchasing and receiving process was running entirely on email and spreadsheets.
Phase 2 was a Kaizen Commerce AnyDB back-office build: vendor onboarding workflow, purchase order tracking, receiving and QA inspection records, and inventory discrepancy case management. What had been running on spreadsheets and email threads became structured, trackable workflows with clear ownership, handoff gates, and exception handling.
The receiving team went from a 3–4 day lag between delivery and accurate inventory records to same-day. The buyer team eliminated 12 recurring email chains. And the operations director, for the first time, had a visible queue of open vendor issues rather than a growing inbox.
Northgate thought they had an ecommerce problem. They also had a POS problem. And an inventory problem. And a receiving problem. The Blueprint exists to map the full picture before any project starts — because fixing only one layer while leaving the others broken produces a more expensive failure, not a cheaper success.
At 14 stores, the temptation is to roll out everything at once and absorb the chaos. The pilot found three issues that would have affected every location. The six-week timeline for the first live store felt slow in the planning phase. In execution, it was what made the remaining 13 stores straightforward.
The technology decision — Shopify POS — was made in the Blueprint. Everything after that was operational: location model, inventory readiness, hardware, permissions, training, and the workflows associates need to do their jobs without thinking about the system.
* Client name and identifying details have been changed. All operational outcomes and metrics reflect the actual engagement. This case study is a composite informed by multiple Kaizen Commerce delivery engagements and represents the typical outcomes achievable with proper implementation discipline.