For a business of 10 to 250 people, a data strategy is one page that answers six questions: which decisions matter most this year, which numbers inform those decisions, where each number comes from, who owns each one, how you'll know each number is right, and what you'll do first. That's it. It fits on a side of A4, the leadership team can read it in five minutes, and it tells everyone what to build, fix or ignore. Anything longer than that won't be used. A 40-page document gets approved at a board meeting, filed, and never opened again, while the one-pager gets pinned up and argued over, which is exactly what you want from it.
What goes in a one-page data strategy?
Six sections, one per question. The table below is the template: copy the first two columns onto a page and fill in the third for your own business. The example column is a made-up 50-person wholesale distributor, there to show the level of detail that works. It isn't a client.
| Section | The question it answers | Example line (fictional 50-person wholesaler) |
|---|---|---|
| 1. Decisions | Which three to five decisions would we make better this year with better numbers? | Which customers stay on 30-day account terms; when to reorder the top 200 stock lines; whether to add a second delivery van in the spring. |
| 2. Numbers | Which one number informs each decision? | Overdue debt per customer over 60 days; weeks of stock cover per line; deliveries per van per day. |
| 3. Sources | Where does each number come from today? | Debt: the Sage 50 sales ledger. Stock cover: Sage 50 stock records plus the buyer's reorder spreadsheet. Deliveries: the route planner's weekly export. |
| 4. Owners | Who owns each number, by role, and answers for it when it moves? | Debt: the credit controller. Stock cover: the purchasing manager. Deliveries: the transport manager. |
| 5. Checks | How will we know each number is right, and how often do we check? | Overdue debt ties to Sage's own aged debtors report every Monday. "Overdue" is defined once, in writing. Stock cover is spot-checked against a count of 20 lines each month. |
| 6. First step | What happens in the next 90 days, and what does "done" look like? | One report on overdue debt by customer, updated daily, tied to the ledger, read by the MD and FD at the Monday meeting. Stock cover is next. |
Source: my own working template, built from the leadership reporting I do for UK businesses of this size. Adapt the rows to your business; keep the six sections.
Add a review date at the bottom (every six months is plenty) and the names of the people who agreed it. That's the whole document.
Notice that there's no row for "tools" or "platform": everything follows from the decisions. And the first step is small enough to finish in a quarter, which keeps the room interested.
Why does a data strategy have to fit on one page?
Because the people who have to use it are busy running the business. An MD reads it once properly, then checks each new request against it: does this help one of our five decisions? If not, it waits. That test only works if the page is short enough to hold in your head.
One page also forces the useful arguments early. If the leadership team can't agree on the five decisions that matter, or can't agree what "overdue" means, you've found the real problem before spending a penny on software. A long document tends to bury that argument under chapters of background. A page is also easy to change: once the van decision is made, it comes off and something else goes on.
How do you write one?
For a business this size it should be a half-day session for the leadership team, not a three-month project. The order matters more than the time:
- Start with the decisions, not the data. Ask each director which calls they make regularly where they're guessing, or waiting too long for a number. Write them as decisions ("whether to...", "when to...", "which..."), not as topics ("sales", "stock").
- Pick one number per decision. The one you'd look at first. Note any others for later.
- Find where each number lives today. Usually the accounts system, the CRM, an operational system and at least one spreadsheet somebody maintains by hand. Write the spreadsheet down honestly.
- Put a name against each number. A role is fine on the page, as long as everyone knows who it means. If nobody will own a number, it probably isn't an important one.
- Decide how you'll check it. The simplest check is that the number agrees with the source system's own report. Write down the definition too, because two people meaning different things by "margin" is where most reporting arguments come from.
- Choose the first step. One decision, one number, one report, done in 90 days. Pick the one with the cleanest data and the most frequent decision.
That sequence is really the answer to "data strategy for leaders": leaders own steps 1, 2 and 4, because they're business questions. Steps 3, 5 and 6 are where outside help earns its keep, if you need it at all.
Where do UK businesses this size stand today?
Businesses of this size are generally further along on collecting data than on using it. The UK government's UK Business Data Survey 2026 (Department for Science, Innovation and Technology, published 18 June 2026, fieldwork October 2025 to January 2026; figures as of September 2026) found that among UK businesses that handle digitised data, 35% of small businesses (10 to 49 employees) and 49% of medium businesses (50 to 249 employees) said they analyse data to generate new insights or knowledge, compared with 69% of large businesses. The same survey found that 51% of small and 58% of medium businesses said they used AI-based technologies. So more businesses in that size band are using AI than are analysing their own data.
That isn't a criticism. Large businesses have analysts; a 50-person firm usually has a finance director with a spreadsheet and not enough hours. The one-page strategy closes that gap directly by naming the few numbers worth analysing and who looks after them.
Sage's research points the same way. In its 23 September 2026 press release, Sage reported that 42% of UK SMEs say their financial information is too spread out to make better use of AI, and that only one in ten can see their full financial picture in one place (as of September 2026). Row 3 of the template, "where does each number come from today?", is the row that starts to fix that.
Do you need a data strategy before you use AI?
You need something like one, yes, and this page is it. AI assistants that answer business questions (Copilot in Microsoft 365 and Power BI, and the chat tools built into accounting and CRM systems) answer from your data. If two systems disagree about revenue, the assistant will confidently quote whichever one it reads. If nobody has written down what "active customer" means, it will guess.
So the six questions on the page are the AI readiness work: agreed decisions, agreed numbers, known sources, named owners and a check that each number is right. When the page is filled in and the first few numbers reconcile, you're in a good position to let people ask questions of the data in plain English. If you already run Power BI and want to test whether its answers will come out right before switching Copilot on, that's what the AI-Ready Power BI Assessment is for.
It's also a safety question. FSB research published in June 2026 found that the biggest AI concern among small firms, at 54%, is AI producing inaccurate responses. A one-page strategy doesn't remove that risk, but the "checks" row is exactly how you know whether an answer is wrong.
What should a small business data strategy leave out?
Leave out the 40-page document. Background, market context, maturity models and vision statements belong in someone else's slide deck. If a paragraph doesn't answer one of the six questions, cut it.
Leave out the tool shopping list. Choosing software before you know the decisions is how businesses end up with three reporting tools and one trusted number. Tools come after row 6. Power BI, Excel with Power Query and the reporting built into your accounts system all do good work, and often the answer is to make better use of what you already pay for.
Leave out the data warehouse on day one. A central database is sometimes the right answer for a business of this size, usually once several systems need joining every day. It's rarely the right first step. Get one number trusted and in front of the right person, then decide what the next ten need.
Leave out the jargon. If a director can't read a sentence aloud to the team and have it understood, rewrite it in plain English. "The credit controller owns overdue debt and checks it against Sage every Monday" is a better policy than a page of framework language.
Leave out the data you don't use. The page lists the few numbers that drive decisions, not everything you hold. If you suspect some of those numbers aren't right, check the symptoms first (two reports that disagree, manual fixes at month end, a figure nobody quite trusts), which I've written up in How to Tell If You Have a Data Quality Problem.
Where do you start?
With row 6 on your own page: one decision, one number, one report.
If you already have Power BI and aren't sure the numbers in it are right, start with a Power BI Health Check. It's £950 fixed: six areas, 48 scored checks, each Healthy, Watch or Critical against a written standard, with a prioritised roadmap. It fills in row 5 for what you've already built.
If you don't have Power BI yet, or have it and nobody uses it, the first concrete step is the First Report: one system you already run, one report answering one written question, the headline figure reconciled back to source to the penny, published to your own workspace, £1,950 fixed, ten working days from the data arriving. In the example above, that's the overdue debt report on the Monday meeting.
If you'd like to talk the page through first, book a free 20-minute call. Bring the decisions you'd put in row 1 and I'll tell you honestly which of those two fits, or whether you need anything from me at all.
Common questions
What is a data strategy for a small business?
It's a one-page plan that says which decisions matter most, which number informs each one, where each number comes from, who owns it, how you'll check it's right, and what you'll do first. For a business of 10 to 250 people that's enough. Anything longer tends not to get read or used.
Is there a data strategy template I can use?
Yes: six sections, one per question. Decisions, numbers, sources, owners, checks and first step, with one line per decision in each. The table in this post is the template, with a filled example for a fictional 50-person wholesaler you can copy the level of detail from.
How long does it take to write a data strategy?
For a business this size, about half a day with the leadership team, provided you start with the decisions rather than the data. The first step on the page should be finished within 90 days.
Do I need a data strategy before using AI?
You need the answers a data strategy gives you: agreed numbers, known sources, named owners and a way to check each number is right. AI tools answer from your data, so if the data disagrees with itself, the answers will too. The one-page strategy is the AI readiness work for a business of this size.
Who should own the data strategy?
The MD or finance director should own the page itself, because the decisions on it are business decisions. Each number on it needs its own owner, usually the person whose job depends on it (credit control for debt, purchasing for stock).
Should a data strategy include which software to buy?
Not at the start. Tools come after the decisions and sources are known, and often the answer is to make better use of the accounts system, Excel or Power BI you already pay for. Buying software first is how businesses end up with several reporting tools and no trusted number.