The owner of a sixteen-person services firm sat down on a Tuesday morning to settle a question that had been irritating her for weeks. Her CRM said she had 487 active customers. Her accounting system reported 423. Her email marketing tool insisted there were 510 people on the list. Three numbers, three tools, three answers to “who are our customers”. None of the systems was broken. Each had been chosen for a good reason. Nobody had ever decided which one was meant to be right.
She is not behind the curve. Sagacity’s survey of SMEs found that only two in five hold customer data in a CRM or database at all, and of those that do, 92% know they should be cleaning it while a quarter still do not. The underlying problem has a name in the enterprise world. It is called master data, and the discipline around it costs Fortune 500 firms multiple millions of pounds in dedicated platforms and multi-year programmes. The good news for a five-to-fifty-person business is that the principle is independent of scale. The version you need is two pages of decisions and a calendar reminder.
What does master data actually mean for a small business?
Master data is the record that says what is true about a thing your business tracks. In a small firm, that thing is almost always a customer, a supplier, an employee, or a product or service you sell. The CRM has a field for customer address. The accounting system has one too. When they show different values, master data is the rule that says which one wins.
IBM defines master data management as a comprehensive approach to managing an organisation’s critical data across the enterprise. The sentence sounds large-scale, and at enterprise scale it is. The principle underneath is the same at every size: someone needs to own the truth about who your customers are, where they live, what they owe you and what you owe them. When no one does, decisions get made from conflicting numbers. Marketing chases prospects already acquired by sales. Finance reconciles invoices against mismatched records. The Information Commissioner’s Office expects UK businesses of every size to keep data “clean and accurate or delete it”, and when a data subject access request lands, the answer to “show us your one authoritative record” cannot be “we have three”.
Why does this matter for your business?
Gartner’s research puts the average cost of poor data quality at $12.9 million a year for a large organisation. The proportionate share for a £500,000 to £5 million services firm lands somewhere between £5,000 and £50,000 of annual waste. The waste is rarely visible as a line item. It sits in chased invoices, duplicate marketing, support staff working without full history, and the time tax of cross-checking records.
The compliance side is sharper. GDPR makes you accountable for breaches and data subject requests regardless of how many systems the data lives in. When a customer asks you to delete every record of them, you have to locate every instance. The firm with one authoritative customer record and a documented sync to its secondary systems can answer in an hour. The firm with three uncoordinated lists spends a day, often misses one, and exposes itself to a fine. Master data sits underneath credible compliance, and it is the difference between a calm GDPR response and an emergency.
Which four entities does an SME actually need to get right?
Four categories of master data carry the load in a typical small business: customers, suppliers, employees, and products or services. Each has a natural home. Customers usually live in the CRM, suppliers in the accounting system, employees in payroll or HR, products in a pricing document. The point is that you pick one home per entity and tell everyone which it is.
Customers shape revenue, retention and regulatory compliance, which is why they sit first. Suppliers are the quietest of the four and the one where drift causes procurement errors, missed contract renewals and modern slavery reporting gaps. Employees matter less at five people than at fifty, but the principle still holds: payroll, email administration and access control all hold employee records, and offboarding without a single source leaves orphaned access rights. Products or services are core to invoicing, sales conversations and revenue recognition. If your sales team quotes a different price from the one your accounting system invoices, the gap is a product master data problem.
What goes on the two-page decision?
The first page is an ownership matrix. Four columns, four rows. Entity type, authoritative system, what flows where, and update frequency. The customer row in a small services firm reads CRM authoritative, with quarterly syncs to accounting and real-time updates to email marketing and support. The supplier row points at accounting, the employee row at payroll, the product row at a pricing document. You are making a choice with the tools you already own.
The second page assigns ownership and reconciliation responsibility using a RACI block. Who is Responsible for updating the authoritative record? Who is Accountable, usually the owner or operations lead? Who needs to be Consulted, typically the administrators of the secondary systems? Who needs to be Informed, the teams that use the data? Ambiguity about ownership is what causes data to degrade. When someone owns it, it gets done. The Sagacity survey is clear that SME data quality fails less on knowledge and more on ownership; the missing piece is almost always the named person, not the better tool.
When does this stop working and what comes next?
The proportionate version will carry a five-to-fifty-person firm for years. The signals that it has outgrown itself are specific. Quarterly reconciliation starts taking longer than an afternoon and consistently finds more than a dozen mismatches. The number of systems holding overlapping data passes five or six, usually after an acquisition or a new business unit. An auditor stops asking whether you clean your records and starts asking for a documented data governance policy.
When those signals land, you are at the boundary that Gartner’s data governance maturity model places between level 3, proactive, and level 4, coordinated. The next step is a named data steward, even part-time, who is explicitly accountable for data quality and governance enforcement. McKinsey’s research on data governance points at this directly: the commonest failure mode is a strong governance strategy that never gets operationalised, and naming someone closes that gap. Alongside the role, you write a four-to-six-page governance document covering purpose, scope, roles, classification, standards and review cadence. Modest tooling, such as a data quality checker or a small-scale MDM tool, comes only when volume genuinely warrants it. The worst position to be in is a hundred-person firm with chaotic data and no governance at all. The best is a ten-person firm that started early. Two pages and a calendar reminder is where that starts.
If you want help drawing the matrix for your own firm or deciding whether you have hit the point where proportionate stops working, book a conversation.
