Customer data cleansing & Match/Merge for a telecom operator
Real MDM engagement (client anonymized under NDA): design of cleansing rules and Match & Merge in TIBCO EBX for customer master data from two Central American countries.
Context
Customer records from two countries arrived with inconsistent phone numbers and addresses, duplicates across sources and no shared standard. The goal: return clean, deduplicated customer data to the operator from a single governed hub.
What I did
- ▸Researched the phone numbering plans and postal address structures of both countries to define which standardization rules could be applied to each one.
- ▸Implemented cleansing rules in TIBCO EBX: validations, constraints and automatic corrections on phone and address fields, with rejection handling for unrecoverable records.
- ▸Configured Match & Merge: matching policies, similarity thresholds and survivorship rules to consolidate duplicate customers into a golden record.
- ▸Delivered the cleaned dataset back to the client with quality metrics per country, ready for downstream consumption.
Stack
- ▸TIBCO EBX: data model, Data Spaces, validation rules, Match & Merge add-on
- ▸SQL for profiling and quality checks
- ▸Java extensions for country-specific validations
Confidentiality
Client name, volumes and specific rules are under NDA with Alldatum Business. The approach described is real; for technical depth request a private session.