Master Data: Why Your Customer List Has Five Versions of One Client
When one client appears five times in your system, invoices, credit limits and reports all suffer. Why duplicate master data happens, and a practical way to clean it and keep it clean.

Search your system for one of your biggest customers. There is a fair chance you will find it more than once: in English and in Arabic, with and without “Co.”, under the old trading name, and once more as a duplicate a new salesperson created last month.
This is the problem that master data management exists to solve. Each version looks harmless on its own. Together they split the customer’s history, confuse credit control and make every report on that customer slightly wrong.
What master data is, and why it matters
Master data is the core reference information that transactions depend on: customers, suppliers, products, employees, price lists, chart of accounts, locations. It changes slowly, but every order, invoice and payment points to it. When master data is wrong or duplicated, the error spreads into every transaction that uses it.
Consider a pharmaceutical distributor in Amman that merged the customer lists of two sales teams, one serving pharmacies and one serving hospitals. Some private hospitals bought through both channels, so they now appear twice, with different payment terms and credit limits. One record is over its credit limit and blocked. The other is not, so orders keep flowing. Finance sees overdue balances on both records but cannot see the total exposure without exporting and combining them by hand.
How duplicates creep in
Duplicates are rarely the result of carelessness alone. Most come from predictable situations:
| Source of duplicates | What it looks like |
|---|---|
| Spelling and transliteration | The same family or company name written as Mohammad, Mohammed or Muhammad, or with and without the “Al” prefix |
| Two languages | One record created in Arabic by a branch, another in English by head office |
| Legal and trading names | A company recorded under its registered name and again under the name on its shopfront |
| Branches and sites | Each branch of a customer created as a separate customer instead of a delivery address under one account |
| Weak search before creating | A salesperson searches, misses the existing record because of a spelling difference, and creates a new one |
| System migrations and mergers | Lists from two systems or two companies combined without matching |
The same patterns affect products, where one item may exist under a supplier code, an internal code and a slightly different description, and suppliers, where the same vendor is set up separately by purchasing and by finance.
Cleaning the data you already have
A one-off clean-up is the starting point. Work through it in a controlled way, rather than letting everyone merge records as they find them.
- Extract and profile. Export all records of one type, such as customers, and count blanks in key fields like tax number, phone, email and address. This shows how much matching information you have to work with.
- Find likely duplicates. Match on unique identifiers first, then on phone numbers and email addresses, then on normalised names. Normalising means removing common words such as “company”, “trading” and “est.”, standardising “Al” prefixes and converting to a consistent spelling. Power Query in Excel has a fuzzy merge option that suggests pairs whose names are similar but not identical, which is a useful starting list for review.
- Review suggested matches with the business. Sales and finance should confirm which records are truly the same customer, and which are separate legal entities that happen to look alike. Automated matching proposes; people decide.
- Choose the surviving record. For each group, keep the record with the most complete and correct information, and copy across anything useful from the others, such as a missing tax number or an additional delivery address.
- Merge or block, depending on your system. Some systems allow true merges that move transaction history to the surviving record. Others only let you archive or block the duplicate. Either way, make sure no new transactions can be posted to the retired records.
- Check open balances and credit limits. After merging, review balances, credit limits and payment terms for each surviving customer, since these may have been set differently on each duplicate.
Do the clean-up in batches, starting with the customers who generate most of your revenue or most of your credit risk. Cleaning the top few hundred records properly delivers most of the benefit quickly, and the method can then be applied to the remaining records with less supervision.
Keeping master data clean afterwards
Without prevention, duplicates return within months. A few controls make the difference:
- Restrict who can create master records. A small team, or a request form that a data owner approves, rather than every user.
- Make key fields mandatory. Tax number, a primary phone or email, and the customer’s preferred language should be required before a record can be saved.
- Check for duplicates at the point of entry. Many ERP (enterprise resource planning) and CRM (customer relationship management) systems can warn when a new record matches an existing one on tax number or phone. Where they cannot, a simple automated check can flag likely duplicates daily for review.
- Use parent and child structures. Record customer branches as delivery addresses or child accounts under one parent, so credit and reporting work at the right level.
- Assign a data owner. One person per data type, accountable for its quality, with a monthly review of new records and flagged duplicates.
Where customer records include individuals, remember that this is personal data. Jordan’s Personal Data Protection Law of 2023 is one reason to collect only what you need and to control who can see and change it.
One record for each customer, and rules that keep it that way
A quick way to size the problem is to export your customer list and count how many tax numbers appear on more than one record. That number, and the balances attached to those records, usually makes the case for a clean-up better than any argument.
Data quality and data architecture are a core part of our Enterprise Systems service. We help you clean existing master data, design the structures and controls that keep it clean, and automate the checks so problems are caught at the point of entry.



