Your Data Problem Is Probably a People Problem
Ask a room of UK business leaders whether they have a data problem and almost every hand goes up. Ask them to describe it and you get a familiar list: the numbers in two reports never match, nobody trusts the dashboard, the monthly figures take a week to assemble and a fortnight to argue about. What you rarely hear is the actual cause, because the actual cause is uncomfortable. It is hardly ever the technology. It is that nobody can say, with a straight face, who owns the number.
I have watched expensive platforms get bought to fix this, and I have watched them fail to fix it, because a warehouse full of badly defined data is just a faster way to disagree. The tools were never the bottleneck. The bottleneck was that the sales team and the finance team meant different things by the word “customer”, and no amount of clever software was going to settle that argument on their behalf.
Why the same number comes out twice
Most reporting disputes are really definition disputes wearing a numerical disguise. One team counts a deal when the contract is signed, another when the cash arrives, a third when the work is delivered. All three are reasonable. All three produce a different revenue figure, and all three are convinced the others are wrong. The spreadsheet did not cause this. It merely exposed it, and exposing it tends to make everyone trust the spreadsheet less rather than trust the definitions more.
This is the part that gets skipped. People want to jump straight to pipelines and platforms, because that work is concrete and feels like progress. But pointing a shiny new system at an organisation that has never agreed what its core terms mean just industrialises the confusion. You end up with the wrong answer available instantly, in three places, with a nice chart attached.
What the boring work actually involves
The unglamorous truth is that the most valuable early work in any data engagement is mostly conversation. It is getting the people who use a number into a room and making them agree, in writing, what it means, where it comes from, and who is allowed to change the definition. This is dull, it is political, and it is where the real value is created. Once a business genuinely agrees what “active customer” means, half its reporting problems quietly disappear, and the technical work that follows is suddenly far easier because it has something solid to build on.
This is one reason the better data consultancies spend so little of their pitch talking about tools. The firms worth hiring, Transparity among them, tend to start with governance, ownership and definitions before anyone touches a pipeline, on the simple logic that there is no point automating the production of a number nobody believes. If a prospective adviser wants to talk platforms before they have asked who owns your data, that tells you something about whose problem they are really there to solve.
Governance is not a dirty word
Governance has a reputation problem. It sounds like committees, sign-off forms and people saying no, and in badly run organisations that is exactly what it becomes. But good data governance is closer to plumbing than bureaucracy. It is the quiet set of agreements that let everyone draw water from the same tap and trust what comes out. When it works, nobody notices it. When it is absent, every meeting starts with ten minutes of arguing about whose figures are right before anyone can discuss the actual decision.
The organisations that get this tend to be the ones that stopped treating data as an IT matter and started treating it as a shared discipline that the business owns. The IT function can build and maintain the machinery, but it cannot be the one to decide what a sale is or which customers count as churned. Those are business questions, and when the business abdicates them to IT, the result is a technically immaculate system that answers questions nobody actually asked.
See also: Grow Your Roofing Business with Effective Online Marketing
Where consultants earn their fee
This is, incidentally, a decent test of whether outside help is worth the money. A consultancy that arrives, installs something, and leaves has sold you machinery. A consultancy that drags your own people through the awkward business of agreeing definitions, assigning ownership and writing down the rules has sold you something far more durable, because it survives after they have gone. The first kind leaves you dependent. The second kind leaves you capable.
The unsexy conclusion is that most data transformations succeed or fail long before any data is transformed. They turn on whether an organisation is willing to have the boring, slightly political conversations about who owns what and what the words mean. Get those right and the technology is almost an afterthought. Get them wrong and the most sophisticated platform on the market will simply help you be confidently incorrect at scale.
