How Data Analytics Turned a Mid-Size Company Into a Data-Driven Powerhouse
A manufacturing company with 200 employees built an analytics function from scratch and transformed how every major decision in the business was made. Here's exactly how they did it.
The Starting Point: Decision-Making by Instinct
Four years ago, a 200-person precision manufacturing company was making most of its significant decisions, pricing, capacity planning, supplier selection, sales targeting, on gut instinct and monthly spreadsheet reports that took three days to compile and were outdated before they were distributed.
The CEO knew something had to change when the company lost a major contract because a competitor could respond to pricing enquiries in hours while their team needed days to run the margin analysis.
This is the story of how they built a data function that now drives every strategic decision in the business, and what it actually took to get there.
Phase 1: Getting the Data Foundations Right
The company's data lived in four places: a legacy ERP (partially implemented), spreadsheets on individual desktops, a disconnected CRM, and paper-based production records.
The first step was not building dashboards. It was data consolidation. Over three months, they:
- Migrated production records from paper to a structured digital system
- Completed the ERP implementation for purchasing and inventory
- Connected the CRM to the ERP via an integration layer
- Established a central data warehouse (Snowflake) as the single source of truth
This phase was unglamorous, expensive, and unpopular. It was also completely essential.
Phase 2: The First Dashboards That Changed Behaviour
With clean, centralised data, the analytics team built three dashboards that immediately changed how the business operated:
1. Production Efficiency Dashboard: Real-time OEE (Overall Equipment Effectiveness) by machine and shift. For the first time, production managers could see exactly where capacity was being lost, and they were in a competition to improve their numbers.
2. Customer Profitability Report: Margin by customer, product line, and order size. The analysis revealed that their three largest customers by revenue were in the bottom quartile by profitability. Sales conversations changed immediately.
3. Supplier Performance Scorecard: Delivery reliability, quality rejection rates, and price trends by supplier. Procurement moved from relationship-based to performance-based sourcing.
Phase 3: Predictive Analytics
Eighteen months in, with clean data and a team that trusted the analytics function, they introduced forecasting:
- Demand forecasting at SKU and customer level, reducing finished goods inventory by 22%
- Machine maintenance prediction based on production cycle data, reducing breakdown downtime by 31%
- Customer at-risk scoring identifying accounts showing early signs of churn based on order frequency changes
The Results After Four Years
- Gross margin improved by 6.2 percentage points through rational pricing and customer mix optimisation
- Inventory working capital reduced by £1.4M
- On-time delivery improved from 78% to 96%
- Decision cycle time for pricing (the original trigger) reduced from 3 days to 4 hours
What Made It Work
Three factors distinguished this transformation from analytics projects that never deliver:
CEO sponsorship, the analytics function had board-level backing and authority to access all data sources.
Starting with decisions, not data, every dashboard was built around a specific decision that needed to be made better, not around what data was available.
Building internal capability, they hired an internal data analyst in year one and a data engineer in year two, rather than remaining perpetually dependent on external consultants.
Conclusion
Being data-driven is not a technology investment, it is an organisational transformation. The technology is the enabler. The hard work is data consolidation, cultural change, and building the internal capability to act on what the data reveals.