If you've ever had an argument in a board meeting about whether the revenue numbers are right, I want to tell you something: that argument almost certainly wasn't about the numbers. It was about the definitions.
The finance director's revenue figure and the sales director's revenue figure are both correct. They're just measuring different things — and nobody ever agreed on which version to use. So when the two numbers don't match, you get a 40-minute meeting about data integrity that's actually a 40-minute meeting about the fact that nobody ever wrote down what "revenue" means in your business.
A data dictionary solves this. It's not a technical tool. It's a shared agreement — a document that says: in this business, this is what we mean when we use these words. That's it. And it's one of the most valuable things you can build in an afternoon.
What a Data Dictionary Actually Is
Forget the enterprise definition. For an SME, a data dictionary is a simple, structured document — it could be a spreadsheet, a shared Google Doc, a Notion page — that defines your key business terms in one place.
For each term, you record: what it means, how it's calculated, where the underlying data lives, and who is responsible for keeping the definition accurate. That's the core of it. Nothing more complex than that.
The value isn't in the sophistication of the document. It's in the fact that it exists and everyone has agreed to it. Once you have that, you've eliminated an entire category of data argument from your business.
Why SMEs Need One More Than Anyone
Enterprise businesses have data governance teams, data stewards, and formal taxonomy frameworks. They've over-engineered it. SMEs have gone the other direction — no definitions at all, just the assumption that everyone means the same thing.
That assumption causes real, specific problems:
- Inconsistent reporting. The same metric appears differently in different reports because different people calculated it differently. Nobody trusts any of them.
- New staff take forever to ramp up. They spend their first month figuring out what things mean rather than doing their job. Half the time they get it wrong and nobody corrects them until months later.
- Data arguments in meetings. Time that should be spent making decisions gets spent arguing about whether the numbers are right — which they are, just defined differently.
- System migrations become nightmares. When you switch from one CRM to another, or one accounting platform to another, nobody can map across what the old fields meant. You lose data or duplicate it.
All of these problems have the same root cause: you never wrote down what things mean. One afternoon fixes that.
The Terms That Always Cause Problems
Before you start building, it's worth knowing which terms are most consistently undefined in SMEs. These are the ones that cause arguments in almost every business:
- Customer. Is a prospect who's had a sales call a customer? Is someone who bought once two years ago and never came back still a customer? What about someone mid-onboarding? The word "customer" appears in every report, and almost nobody has defined it precisely.
- Revenue. Gross or net? Invoiced or received? Recognised or cash? Does it include VAT? What happens when a refund is issued — does it reduce the period of the original sale or the period of the refund? Every finance team has its own answer.
- Lead. Is a lead someone who filled in a form, someone who booked a call, someone who was manually added to the CRM by a salesperson? The answer affects every sales conversion metric you'll ever calculate.
- Active user. For software businesses: logged in this month? Used a core feature? Paid in the last 30 days? "Active" is one of the most abused terms in business metrics.
- Conversion. From what to what? Lead to customer? Visitor to lead? Proposal to close? "Our conversion rate is 12%" means nothing without the definition of both the numerator and the denominator.
Start with these. If you can get clean, agreed definitions for just these five terms, you've done more for your data quality than most SMEs ever manage.
How to Build One in an Afternoon: Step by Step
Step 1: Start with your most argued-about metrics
Don't try to define everything at once. Ask your finance, sales, and operations leads one question: "What data do we disagree on most often?" The answers will point you straight at the definitions that matter most.
You're looking for the metrics that appear in multiple reports, get used by multiple teams, and have generated at least one meeting argument in the last six months. Prioritise those. Everything else can wait.
Step 2: For each term, document five things
Set up a simple table — a spreadsheet works perfectly. For each term, capture:
- Name — the exact term as it appears in your reports
- Definition — plain English, one or two sentences
- Calculation — how it's derived (formula, logic, data source)
- Data source — which system holds the underlying data
- Owner — who is responsible for this definition and its accuracy
Don't overthink the format. The important thing is that all five fields are completed, not that the table looks impressive.
Step 3: Document the edge cases
This is where most data dictionaries fall short — they define the normal case but ignore the exceptions, which are exactly where the arguments happen.
For every definition, ask: what are the boundary cases? For revenue: "Excludes VAT. Excludes refunds issued within 30 days of the original transaction — these are netted in the same period. Refunds issued after 30 days are recorded in the period of the refund." For customer: "A customer is any account with at least one completed purchase in the last 24 months. Lapsed customers — no purchase in 24+ months — are tagged as Inactive and excluded from active customer counts."
The edge cases are where the arguments live. Document them explicitly, and the arguments go away.
Step 4: Get sign-off from each department head
This is the step people skip, and it's the one that makes the whole thing work. A data dictionary that one person wrote in isolation is just an opinion document. A data dictionary that finance, sales, and operations have all reviewed and agreed to is a shared standard.
Call a short meeting — 30 minutes — and walk through the definitions together. The goal isn't consensus on every nuance. The goal is that everyone agrees to use the same definitions going forward, even if they privately think a different definition would be slightly more useful.
Get that agreement. Write it down. Move on.
Step 5: Store it somewhere everyone can actually find it
The graveyard of SME documentation is full of important documents that were created once, saved in a SharePoint folder called "Strategy" or "Internal Processes," and never opened again. Don't let your data dictionary end up there.
Put it somewhere people genuinely go: the front page of your team wiki, pinned in your main Slack channel, linked directly from your reporting templates. The test is simple — can anyone in your business find it in under 30 seconds without asking someone else? If the answer is no, it's filed in the wrong place.
Keeping It Current as You Grow
A data dictionary isn't a one-time project. Every time you launch a new product line, integrate a new system, or hire into a function that generates its own data, you need to update it. The owner field in each row is important for this reason — they're responsible for flagging when a definition needs to change.
Build a simple annual review into your calendar. One hour, once a year, to walk through the dictionary and check whether any definitions have drifted. That's enough maintenance for most SMEs.
The businesses with the cleanest data aren't the ones with the most sophisticated systems. They're the ones who wrote down what things mean and made everyone stick to it.
The SME Data Dictionary Builder — included free as a bonus with BI Without the BS — gives you a structured template to do exactly this: pre-populated with the most common SME metrics, with guidance on edge cases and a ready-to-use format that your whole team can work from.