Who Owns The Number? A Data Ownership And Stewardship Model For Caribbean Firms

Here is the meeting everyone has sat in. Finance presents one revenue figure, operations presents another, and the two do not match. The next twenty minutes go not to what the business should do but to whose number is correct, and the meeting ends without resolving either. Nobody was careless. Both figures were pulled honestly from systems that define revenue slightly differently, and no one is authorised to say which definition is the company's. This is what an ownership vacuum looks like from the inside, and no amount of better software fixes it, because the missing piece is a decision about accountability, not a tool.

Data governance has a reputation for being heavy, bureaucratic and built for banks. The version most Caribbean firms actually need is none of those things. It is a short answer to one question asked of every important number: who owns this, and where is the real one? Getting that answer written down is most of the battle.

Three Roles, Clearly Separated

The confusion that stalls governance is usually a blurring of three distinct jobs. Keep them apart and the model almost builds itself.

Almost every firm already has custodians. IT keeps the systems running and the backups current. What most are missing is the owner and the steward, which is precisely why data quality has no home and every complaint about "our messy data" evaporates without anyone accountable to fix it. The owner is a business leader, not a technologist, because the questions that matter, what counts as an active customer, whether a stale record may drive a credit decision, are business calls with money attached.

Name The One True Number

The second move is to declare a single source of truth for each number that matters. If "monthly revenue" can be pulled from the accounting system, the sales platform and a finance spreadsheet, and the three disagree, then the business has no revenue figure, it has three candidates. Pick one system as authoritative for that number, write down the definition, and rule that every report and dashboard draws from it. The disagreement in the meeting does not get resolved faster next time. It stops happening, because there is only one number to bring.

A Forum, Not A Department

Governance decisions have to land somewhere. In a large organisation that is a data governance council: a small cross-functional group, chaired by a senior leader, that agrees definitions, settles cross-department disputes, approves quality standards and sets priorities. In a mid-sized Caribbean firm it does not need to be a new committee at all. It can be a standing 30-minute item on an existing management meeting, with the same authority: this is where we decide what a term means across the business and who owns what. The cadence matters more than the formality. A forum that meets and decides beats an elaborate charter that sits in a drawer.

Start Small On Purpose

The failure mode of data governance is trying to govern everything at once, producing a giant catalogue nobody maintains and a policy nobody reads. Do the opposite. Take the three or four data sets behind your most important decisions, the customer master, the price list, the loss or exposure register, whatever yours are. Name an owner for each. Agree the source of truth and the definition for the key numbers they produce. Put a five-minute quality check on the monthly agenda. That is a real, working governance programme, it can be live in weeks, and it earns the right to invest in tooling later by proving the value first on the numbers that carry real weight.

The risk framing is what makes this more than tidiness. In a volatile economy, the decisions that can hurt you most, extending credit, pricing exposure, holding stock against a storm, all run on data. If no one owns the exposure figure and no one has agreed what it means, then the number behind your largest bet has no accountable steward. Governance is simply what makes a risk number trustworthy enough to act on. It will not remove the risk. It removes the second, avoidable risk of not being able to trust your own read of the first.

One honest limit. Ownership on paper does nothing if the owner has no time, no authority and no consequence for a data set that stays broken. The model works when owning a data set is a real line on a real objective, reviewed like any other. Name owners who cannot act and you have built a governance diagram, not governance.

Frequently Asked Questions

What is the difference between a data owner, a data steward and a data custodian?

A data owner is the accountable senior person who decides what a data set means, who may use it and what quality it must meet. A data steward maintains it day to day and applies the owner's rules. A data custodian, usually IT, stores, secures and backs it up. One is accountable, one maintains, one safeguards.

What is a single source of truth and why does a business need one?

It is the one agreed system that holds the definitive version of a number, so revenue or a customer count has exactly one authoritative value rather than three that disagree. Without it, meetings argue about whose spreadsheet is right instead of deciding what to do.

Do small and mid-sized Caribbean companies need data governance?

Yes, and they can start small. A small firm needs a named owner for its handful of critical data sets, an agreed source of truth for each key number, and a short monthly quality check. Governance scales down as cleanly as it scales up.

What does a data governance council do?

It is a small cross-functional forum, chaired by a senior leader, that sets definitions, resolves disputes about whose number is right, approves quality standards and owns priorities. In a smaller firm it can be a 30-minute standing item on an existing meeting.

How do you start data governance without a big project?

Pick the three or four data sets that drive your most important decisions, name an owner for each, agree the source of truth and definition for their key numbers, and put a short quality check on the monthly agenda. That is a working programme, live in weeks.

How does data governance relate to managing risk?

Every risk decision leans on data, so ungoverned data is ungoverned risk. If no one owns the credit exposure or loss figure, the numbers behind your biggest decisions have no accountable steward and no agreed definition. Governance makes a risk number trustworthy enough to bet on.

Related reading across the Caribbean AI network