"We can't act yet. The data isn't good enough." I hear it in most boardrooms. It sounds like prudence. Usually it is the wrong sentence, and it costs the organisation a quarter while everyone waits for a standard the decision in front of them never required.
The right sentence is shorter: good enough for what?
Good is not a property of data
There is no such thing as good data. There is only data good enough for a purpose. Goodness is not a quality sitting inside a dataset waiting to be measured, the way temperature sits inside a room. It is a relationship between the data and the question you are asking of it. Change the question and the same dataset moves from excellent to useless without a single value changing.
Most "is our data good enough?" conversations are unanswerable as posed, because they have dropped the only term that settles them: what is it for. Once you accept that, the next move is to notice that purpose comes in a small number of archetypes, and each sets a different bar.
- To improve something you control, you are looking for direction, not proof. Messier, coarser and slower is fine. The bar is "good enough to act."
- To sell it, or stand a claim or a price on it, the data has to be relied upon by someone who did not produce it. The bar rises to "good enough to depend on."
- To defend a position, to discharge an obligation or survive a challenge from an auditor, a regulator or the other side's lawyers, the bar is different again: "good enough to withstand challenge." Not freshness or volume, but provenance and an unbroken chain.
Every director already lives this distinction, even without naming it. Your monthly management accounts are good enough to run the business. They are nowhere near good enough to file as statutory accounts. Same ledger, two bars. And nobody waits for the audited numbers to make an operating decision, because the wait would cost them the year.
Change the question and the same dataset moves from excellent to useless without a single value changing.
Construction makes the point three times over
In construction the same data exists at three altitudes at once, and the gap between them is where careers and companies are made and lost.
At the coalface there is the project cost-value reconciliation, the CVR or CVC. Monthly, granular, produced by the commercial team. It carries forecast final cost, anticipated recoveries, variations not yet agreed, claims not yet certified. It is full of judgement, and it is meant to be: its job is to tell the project team where margin is leaking and what to chase this month. It is good enough to steer the job. It was never built to be audited.
Those CVRs roll up into the business management accounts: the aggregate position across the portfolio, cash, overhead recovery, the picture the board runs the company on. Directional, monthly, unaudited. Good enough to run the business.
And those, in turn, must resolve into the statutory accounts: signed off, filed, able to survive the auditor and Companies House. Good enough to withstand challenge.
One set of underlying transactions. Three completely different bars. Each layer is entirely fit for its own purpose and reckless if promoted to the next. A CVR is the right instrument to manage a job and the wrong instrument to reassure a lender. The skill is not making all three "good." It is knowing which bar you are standing at.
When bar discipline fails
When that judgement fails, the failure is spectacular. The hardest lessons in our sector, Carillion the most public of them [1][2], were not, at root, failures of data quality. They were failures of bar discipline: forward-looking, judgement-laden project optimism treated as if it were realised, provable, audited value. Steering-grade numbers waved through at the defend bar. The data was not lying. It was being asked to do a job it was never good enough for.
So, can we afford to wait?
This is where "we need better data first" usually goes wrong. The wait is only worth it when the cost of being wrong is greater than the cost of delay. That is a question about purpose, not about data. To improve something you control, the cost of being slightly wrong is small and recoverable, and the cost of delay is real, so you act now and refine as you go. To defend a position, the cost of being wrong is a challenge you lose, so the wait to get it provable is justified.
Which means the reflex to wait for better data is, more often than not, a misclassification. It is a steering decision that has quietly been handed a courtroom standard. You are polishing a CVR to the evidential bar of a statutory account, for a decision that only ever needed the CVR. The cure is not more data. It is naming the purpose, finding the bar that purpose actually sets, and acting at it.
The misclassification runs both ways, and the other direction is the dangerous one. Treating a forecast as a fact, a CVR as an audited position, a working estimate as proof, is how organisations walk confidently off a cliff.
This is one cut of a more structured way of reading data: against its purpose, ending not in a maturity score or a dashboard but in a decision, whether to act, wait, not rely on it, or not hold it. But the part worth taking away does not need the apparatus. Before your organisation delays a decision for better data, ask three questions in order. What is this data for? What bar does that purpose actually set? And has anyone moved the bar, up or down, without telling you?
Most of the data paralysis I see dissolves at the first question. And most of the data disasters I have watched would have been caught at the third.
Synaform's Data service and the Playground start from the same place: connecting and modelling the data you already hold so that each decision is read against what it is for, and turned into insight and foresight.
What is the last decision your organisation delayed for data it never actually needed? And, the harder one: what is the last it waved through on data that was never good enough for the bar it was standing at?
Sources
- [1] House of Commons Business, Energy and Industrial Strategy Committee and Work and Pensions Committee, joint parliamentary inquiry into the collapse of Carillion, 2018.
- [2] Financial Reporting Council, investigation into the accounting and audit of Carillion.