All 42 stores live · Real-time inventory sync operational · Azure Monitor dashboards & store manager training complete across 3 rollout waves
Who we worked with
A US-based retail chain operating 42 stores across the Midwest, managing operations on an 11-year-old custom-built .NET Framework 3.5 retail management system hosted on a single dedicated server in their corporate office in Columbus, Ohio.
The system had no API layer, no e-commerce integration, and inventory was tracked via nightly CSV exports emailed between store managers. Stockouts were costing an estimated $1.3M/year in lost sales, and their Shopify online store had zero connection to in-store inventory — driving 220+ customer complaints per month about items showing "in stock" online but unavailable at pickup.
| Metric | Before | After | Change |
|---|---|---|---|
| Inventory visibility | Nightly CSV exports (24hr delay) | Real-time across 42 stores + online | 60% accuracy improvement |
| Admin portal load time | 8.2 seconds | 1.6 seconds | 80% faster |
| Repeat purchase rate | 18% | 25.2% | 40% increase |
| Online sales | $1.8M/year | $2.43M/year | 35% increase |
| Stockout-related lost sales | ~$1.3M/year estimated | ~$480K/year | 63% reduction |
| Customer complaints (inventory) | 220+ per month | 28 per month | 87% reduction |
Running a 42-store chain on spreadsheets and gut feel
No API, no real-time data, no e-commerce connection — an 11-year-old system that was costing the business over $1.3M a year in stockouts alone, with a single server in Columbus and zero redundancy.
Inventory Blindness Across 42 Stores
Each store manager exported a nightly CSV and emailed it to corporate. The corporate team manually consolidated these files every morning — a 2-hour process that was frequently inaccurate. At any given time, corporate had inventory data 12–36 hours old. Stockouts ran at ~$1.3M/year in lost sales; overstock in slow-moving locations tied up roughly $400K in dead inventory.
$1.3M/year in lost sales · 36hr inventory delayDisconnected Online and In-Store Channels
The Shopify online store operated as a completely separate system with its own inventory counts maintained manually by one employee. Zero real-time connection to in-store availability resulted in 220+ customer complaints per month — customers ordering online for pickup only to find items unavailable. The client estimated a 12% cancellation rate on online orders as a result.
220+ complaints/month · 12% order cancellation rateAging Technology with No API Layer
The .NET Framework 3.5 application had no REST API, no webhook capability, and no way to integrate with modern tools. Every integration was done via CSV or manual data entry. The admin portal averaged 8.2-second load times, and the system froze during peak hours. The single Columbus server had no redundancy — a 2023 power outage took the system offline for 9 hours.
No API · 8.2s load times · 9hr outage in 2023No Customer Data Unification
Online purchases couldn't be connected to in-store history. Loyalty programs were paper-based in stores and email-based online with no crossover. The marketing team ran blanket promotions to their entire 68,000-person email list with no segmentation capability. Email open rates had dropped to 11% and were still declining.
68K emails · 11% open rate · no segmentationFive phases, real-time inventory across 42 stores
A 24-week delivery — assessment through store-by-store rollout — with a 3-wave go-live (5 pilot stores → 15 stores → 22 stores) so every store manager was trained and ready before launch.
Assessment & Roadmap
Wks 1–3Azure Infra & Data Migration
Wks 3–8App Rebuild & API Layer
Wks 6–16AI Forecasting & CRM
Wks 14–22Rollout & Training
Wks 20–24- Legacy Health Score™ assessment — scored 38/50 (Concerning, bordering Critical); mapped all data flows between the legacy system, 42 stores, Shopify, QuickBooks, and shipping partners
- Set up Azure App Service, Azure SQL Database, Azure Cache for Redis, and Blob Storage — fully managed, no dedicated server
- Migrated 2.8M transactions, 145K customer records, and 22K product SKUs from legacy SQL Server to Azure SQL
- Ran a 2-week parallel operation to validate complete data integrity; each store's local data consolidated into a central cloud database for the first time
- Rebuilt core application in .NET 8 with a clean REST API layer (87 endpoints) and a responsive Vue.js admin frontend with role-based access for store managers, regional managers, and corporate
- Implemented real-time inventory sync via Azure Service Bus — any sale, return, or transfer at any store updates all connected systems within 30 seconds
- Integrated with Shopify Storefront API for live stock levels on product pages and full BOPIS (buy online, pick up in-store) functionality — 0 to 340 pickup orders/month within 3 months
- Deployed a demand forecasting model via Azure Machine Learning, trained on 3 years of sales data across all 42 locations with seasonal, promotional, and weather adjustments
- Model generates weekly SKU-level restock recommendations per store — reduced stockouts by 63% and identified ~$160K of overstock to redirect between locations in the first 6 months
- Built a unified CRM module merging Shopify and in-store purchase history into single customer profiles; integrated with Mailchimp for segmented campaigns by behavior, location, and preference
- Wave 1 (Wk 20): 5 pilot stores — validated system under live conditions, captured store manager feedback before wider rollout
- Wave 2 (Wk 22): 15 stores — refined onboarding process based on pilot learnings
- Wave 3 (Wk 24): Remaining 22 stores — all 42 stores live simultaneously on the new platform
- In-person training for store managers at each wave; video training library built for new staff onboarding; Azure Monitor dashboards and dedicated Slack support channel live for first 4 weeks post-launch
Managed — no in-house ML team required
2.8M transactions · 42 store locations
Spend 2.3× more than single-channel customers
Every choice made for retail operations at scale
| Technology | Role | Why This Choice |
|---|---|---|
| Microsoft Azure (App Service, SQL, Redis) | Cloud infrastructure | Client's IT team had Azure familiarity; strong .NET ecosystem integration |
| .NET 8 | Backend API & services | Natural upgrade path from .NET Framework 3.5; 87 clean REST API endpoints |
| Vue.js | Admin frontend | Lightweight, fast rendering for data-heavy retail dashboards with role-based views |
| Azure Service Bus | Event-driven messaging | Real-time inventory sync across 42 stores — updates within 30 seconds of any transaction |
| Azure Machine Learning | AI demand forecasting | Managed ML; weekly restock recommendations by SKU and location with no ML team needed |
| Shopify Storefront API | E-commerce integration | Live inventory feeds on product pages and BOPIS functionality for all 42 store locations |
| Azure SQL Database | Primary database | Managed SQL Server familiar to the client's DBA; centralized all 42 stores for the first time |
| Mailchimp | Email marketing | Segmented campaigns based on unified CRM data — open rates from 11% to 24% |
| Azure Monitor | Operations monitoring | Performance dashboards and alerting for corporate and store-level visibility |
Results that transformed the business
Across inventory accuracy, online revenue, customer engagement, and operational efficiency — all delivered in 24 weeks with a 3-wave rollout and zero store disruption during go-live.
Saved Annually
Increase
Per Month (New)
Complaints
5 engineers, end-to-end retail delivery
A lean, full-stack team covering solution architecture, .NET backend, Vue.js frontend, Azure DevOps, and quality — with in-person store manager training and a 3-wave rollout built into the engagement.
Solution Architect
Legacy Health Score™ assessment, Azure architecture design, data flow mapping across 42 stores, Shopify, QuickBooks, and shipping partners
Backend Developer
.NET 8 REST API (87 endpoints), Azure Service Bus real-time sync, Shopify Storefront API integration, unified CRM data model
Frontend Developer
Vue.js admin portal with role-based access for store managers, regional managers, and corporate; responsive design for in-store tablet use
DevOps Engineer
Azure App Service, Azure SQL, Redis Cache provisioning; CI/CD pipeline; Azure Monitor dashboards and alerting for all 42 store locations
QA Engineer
End-to-end inventory sync validation, BOPIS flow testing, load simulation across 42 concurrent store connections, 3-wave rollout monitoring
24-Week Full Delivery
3-wave store rollout · In-person training per wave · Video library for new staff · All 42 stores live with real-time inventory sync
What the client said
"For 11 years we ran this business on spreadsheets and gut feel. I didn't realize how much money we were leaving on the table until I-Verve showed us the numbers. The demand forecasting alone paid for the entire project in the first year. And our store managers — who were skeptical at first — are now the system's biggest advocates because it actually works when the store is busy."