150 million records,
99.5% accuracy, 36 hours
How a major European last-mile logistics company turned hundreds of millions of fragmented customer records into trusted single customer views — laying the foundation for everything that comes next.
Last-mile logistics & e-commerce fulfilment
Single customer view — 150M records resolved into 38.1M golden profiles
Hundreds of millions of records from dozens of retailers — via Amazon, eBay and internal databases
Implementation phase — with Pinboard Consulting
The data chaos that broke traditional solutions
A major European last-mile logistics company was drowning in customer data chaos. Orders poured in from dozens — potentially hundreds — of separate retailers, each feeding customer records through its own integration. Across marketplaces like Amazon and eBay and the company's own internal databases, hundreds of millions of records piled up — and their customer service teams couldn't get a single view of who their customers actually were.
The scale was staggering:
- ▹ 150 million customer records arriving from dozens of retailers sending orders
- ▹ Inconsistent data formats and missing information
- ▹ Marketplace integrations with heavily masked email addresses
- ▹ Complex address matching requirements for accurate deliveries
Previous attempts to solve this problem either couldn't handle the scale or sacrificed accuracy for speed. The company needed something different.
Initial analysis revealed 377 data quality issues in just 1,000 sample records. Yet Tilores successfully resolved:
- ▹ Multiple records for single customers across platforms
- ▹ Heavily masked eBay email addresses
- ▹ Inconsistent address formatting and name variations
- ▹ Complex household relationship mapping
Performance that defied expectations
Tilores' entity resolution platform was implemented by Pinboard Consulting in just three weeks. The results were immediate and impressive:
- ▹ 150 million records processed in just 36 hours
- ▹ 99.5% accuracy rate confirmed by comprehensive testing
- ▹ 8 milliseconds per record processing time
- ▹ 38.1 million golden customer profiles — one single customer view per real customer — created from deduplicated data
The three-week timeline from deployment to full-scale processing shows how Tilores' no-code interface and pre-built connectors enable rapid implementation without sacrificing sophistication.
Advanced matching intelligence
Tilores' proprietary matching engine combines comprehensive data transformation with sophisticated matching rules. The platform configured:
- ▹ 10 advanced matching rules covering all business scenarios
- ▹ 10 specialised matchers for different data types
- ▹ Custom ETL pipelines for seamless integration
- ▹ Real-time API endpoints for operational systems
Beyond the numbers
A unified, trusted customer view sets up powerful new capabilities:
- ▹ Enhanced delivery accuracy through improved address matching
- ▹ First-contact resolution from a single customer view
- ▹ Automated fraud detection through pattern recognition
- ▹ Personalised service delivery based on household preferences
And perhaps most importantly, it lays the foundation for future innovation — including AI/LLM integration and advanced analytics capabilities — on top of a single, trusted view of every customer.
Why this implementation matters
This case study represents more than impressive numbers — it demonstrates that modern entity resolution can deliver immediate business value while establishing the data foundation for tomorrow's opportunities. Real-time matching at hundreds-of-millions scale is no longer aspirational; here it is a measured proof point.
Transform your customer data —
at enterprise scale
See how Tilores handles entity resolution on your own data, in real time. Available on AWS Marketplace.
Moving goods? See the Logistics solution, or download the full case study PDF.