How a technical wholesaler with 8,000 items and three warehouse locations transformed from four competing inventories to a learning operating system – and now manages procurement, assortment, and pricing in a data-driven manner, instead of reacting afterward.
Current Situation
The company supplies craft businesses, industrial customers, and increasingly end customers through an online shop – with fastening technology, tools, and consumables. Thousands of items, dozens of suppliers, three warehouses, customers who order yesterday and expect delivery tomorrow. The ERP theoretically knew the inventory. The WMS knew the storage locations. The online shop showed availabilities. And somewhere in between was the moment when someone called someone else to ask if the item was really there. The operational reality:
- The ERP said 200 pieces, the warehouse counted 140 – the truth was only discovered at picking time
- Reorders were based on experience and Excel, not on sales data – some items lay around forever, while others were constantly missing
- Variants, BOMs, special conditions, and customer-specific pricing created manual workarounds in every single order
- ERP, WMS, online shop, and supplier portals had their own logics – data was manually reconciled daily
- Customer and margin analyses took days because numbers had to be aggregated from multiple systems
The problem was not the individual system. The problem was that there was no seamless data flow between procurement, inventory, sales, and shipping. The team learned to fill in the gaps manually. This worked – but every new item, every new customer, every new sales channel created proportionally more effort.
More revenue did not mean more margin, but less clarity.
Three Stages to Operational Freedom
The project followed our 3-stage model. Each stage builds on the previous one. After 12 months, all three stages were in productive use.
| Stage 1 – Order | Stage 2 – Automation | Stage 3 – Freedom |
|---|---|---|
| Centralizing items, inventory, suppliers, customers, and orders; connecting systems | Automated interfaces, intelligent reordering, pricing rules | AI demand forecasting, assortment optimization, dynamic pricing, supplier rating |
Stage 1 – Order
Before we could automate anything, all departments needed a common data foundation. The crucial question was not which system is the right one – but which data is correct and which is not. In Stage 1, we established:
- An inventory reconciliation across all three warehouses – we identified where ERP inventory deviated from physical inventory and which discrepancies were systemic
- A clean item structure – we removed duplicates, consistently modeled variants, and consolidated BOMs
- A central assortment logic – we defined item classes, set pricing rules, and mapped special conditions in the system
- An inventory truth – a single inventory per item, updated in real-time across all warehouses and channels
The most challenging phase was not the technical migration. It was the cleanup. We tidied up item records that had grown over years with inconsistent names, duplicate entries, and orphaned variants alongside procurement and inventory – line by line.
The ERP had 8,400 item numbers. After the cleanup, 7,800 remained – and the team could trust each individual number for the first time.
The implementation was gradual: system audit with procurement, inventory, sales, and management. Key users from each department took on the role of multipliers. We activated the modules one by one and planned six weeks of parallel operation for order processing. After four months, all departments were working on a common data foundation.
Stage 2 – Automation
Based on structural clarity, we specifically added automation. The biggest impact was seen in procurement: instead of manually reconciling minimum stock levels in Excel weekly, the purchasing manager now receives reorder suggestions based on actual sales data, lead times, and current inventory. Seasonal patterns and customer behavior automatically feed in. What was once weekly manual work has changed into a data-driven process that operates daily. Other automations in this phase include:
- Automatic data flows between ERP, WMS, online shop, and supplier portals – no one is copying manually anymore
- Inventory changes in the warehouse update availability in the shop in real-time
- Special conditions and customer-specific prices are stored as rules in the system, no longer in the sales department's heads
- An order entry through the shop automatically triggers picking, shipping preparation, and invoicing
After Stage 2, daily operations were calmer, more accurate, and faster. But the strategic questions – which assortment the company should actually carry, at what prices, and with which suppliers – still lay with humans.
Stage 3 – Freedom
With a stable data foundation and functioning automation, the system reached a point where it can co-responsibly address strategic questions. Stage 3 has changed how the company manages its assortment.
AI Demand Forecasting per Item
For each of the 7,800 items, the system now forecasts the demand for the next weeks and months. Seasonality, customer behavior, market trends, and individual sales patterns are included. Reorders are no longer based on yesterday – they anticipate tomorrow. Overstocks and stockouts have decreased. The tied-up capital in the warehouse is dropping, even as business grows.
Automatic Assortment Optimization
The system proactively reports which items procurement should remove from the assortment. The criteria: sales velocity, capital commitment, margin, strategic relevance to the core business. Every quarter, the system provides a suggestion with candidates for delisting – and vice versa with candidates for inclusion, based on customer inquiries, search behavior in the shop, and supplier offers. Procurement decides – but based on data, not intuition.
Dynamic Price Adjustment
The system continually adjusts prices – based on margin pressure, competitor prices, demand developments, and customer segments. Management centrally defines rules and limits; the system executes. The margin per item has measurably increased without losing customers.
Data-Driven Supplier Rating
The system continuously evaluates each supplier – delivery reliability, quality, price development, complaint rate. The assessment is no longer a subjective evaluation but a measurable metric. New awards, volume negotiations, and failure scenarios are based on this data. The procurement side is now as transparent as the sales side.
Measurable Results After All Three Stages
| Area | Before | After Stage 2 | After Stage 3 |
|---|---|---|---|
| Inventory Management | ERP and warehouse do not match | A truth across all warehouses and channels | Optimized inventory, capital commitment significantly decreased |
| Disposition & Procurement | Experience and Excel | Reorder suggestions from sales data | Demand forecasts per item, several weeks in advance |
| Assortment Management | Static, rarely checked | Transparent, rule-based | Quarterly automatic optimization proposals |
| Pricing Management | Once a year, manual | Pricing rules in the system | Continuous, data-driven, per customer segment |
| Supplier Management | Subjective, experience-based | Central master data | Automatic rating based on measurable metrics |
| Control Mode | Reactive – after the fact | Proactive – with early signals | Predictive – the system thinks along |
Changes for All Stakeholders
The system has transformed how the company manages its business. Not just operationally – but also strategically. Procurement conducts supplier negotiations based on robust data instead of in Excel. Contracts improve because the company's position is transparent before every conversation. The warehouse works with inventory that is accurate – which means also: with fewer total items, because forecasting is more precise. Sales sees, per customer, which items bring margin and which only bring revenue – and provides more targeted advice. Management now has real-time insights into assortment performance, inventory turnover, and customer margins. Strategic decisions – new product lines, new customer segments, new sales channels – are made by the leadership team based on real-time data, not on the monthly report from controlling.
Previously, complexity grew with revenue. Today, margin grows with revenue – and the complexity stays where it belongs: in the system.
And perhaps most importantly
The company has decoupled its size from its manageability. More items, more customers, more channels no longer make the business more confusing. They make it better – because each additional data point makes the overall system more precise.
What This Project Demonstrates
This project showcases a pattern particularly prevalent in retail: ERP, WMS, online shop, and supplier portals all exist – but they create competing realities. Inventories do not match, procurement orders based on experience, and every special condition becomes a manual workaround. The first step was not a better ERP or a new WMS. It was the decision to create an inventory truth – one data point per item that all departments can trust. Only this order enabled automation. Only this automation has allowed for freedom.
Order enables automation. Automation enables freedom.
If you manage an assortment with high item diversity and find that growth generates more complexity instead of more margin, you should ask yourself a question: Do procurement, inventory, and sales see the same inventory – or each their own? If the answer is “each their own,” the lever is not in an additional system. It lies in finally creating a common truth – and in what the company can build on once it's there.