An empty field is not a statement from the customer

## Ask what the value would claim to know

A missing value is often described as a technical obstacle. A comparison cannot run, a form cannot be submitted or a report contains an inconvenient gap. The apparent remedy is to fill the field using a default or an estimate. Sometimes an analytical method can appropriately use an explicit assumption. The difficulty begins when that assumption enters the operational account without its status and becomes indistinguishable from something a customer actually declared. The column may become easier to process while the business becomes less able to explain what it knows.

Our fictional stewardship review concerned annual turnover in 2,400 small-business files. Four hundred and twenty had no recorded declaration, ninety explicitly declared zero and 1,890 stated a positive amount. A preparation rule proposed assigning a typical amount to the missing entries. That might have satisfied a numeric requirement in the downstream workflow. It would not have supplied a statement from those businesses. I wanted the review to begin with the claim represented by the field, rather than with the inconvenience created by its absence. A value's usefulness depends partly on how it came to exist.

The ninety zero declarations needed their own treatment. Zero is a substantive recorded answer; missing is the absence of that answer in the account we reviewed. Neither automatically establishes the truth of the underlying business circumstances. A declaration may need further professional examination, and a missing entry may reflect an intake problem rather than a refusal to provide material. Collapsing the states would remove questions that the responsible owners still needed to answer. Preserving them did not require the data function to decide an underwriting outcome or accuse anyone of making a false statement.

This is why I distinguish analytical assumptions from operational assertions. A specialist may choose a method that uses an estimate for a stated purpose, with its limitations examined. That does not authorise writing the estimate back as the customer's declaration. The separation should survive movement between systems and reports, because a later user may not have heard the original explanation. A data leader should ask whether the status of the value travels with it, who can change that status and what a business decision-maker will reasonably infer when reading the field months later.

## Provide a route for unresolved records

Keeping a gap visible is only part of the responsibility. If the workflow has no way to handle unresolved material, colleagues will face pressure to bypass the distinction. A principled data rule that leaves the operating team unable to proceed can encourage informal workarounds. I wanted underwriting owners involved in defining an assisted route and in determining what further evidence they needed. They retained the actual professional decision. My function's contribution was to make the record's state clear enough that the decision would not rest on an invented declaration.

We spent ₹48 lakh on source review and confirmation work. Of the 420 missing entries, 312 were subsequently confirmed through the agreed route, while 108 remained unresolved. Those figures describe the outcome of that exercise. They do not mean that every positive amount was independently verified or that the business could assume the remaining population had the same characteristics. The unresolved entries stayed visible. The confirmed values retained their later confirmation status rather than appearing as though they had been present from the beginning. The history mattered to anyone examining how the account developed.

The delivery lead's preference for a complete numeric field was not frivolous. A uniform field simplifies comparisons and reduces exception handling. The revised approach accepted delay, continuing stewardship and a more demanding explanation. Those costs belong in the case, because otherwise the decision reads as an obvious correction with no operational consequence. We chose to incur them because the alternative obscured a distinction needed by the business. The resulting dataset was less mechanically convenient and more faithful to the statements available. That is a trade-off a leadership account should be able to explain plainly.

I also want unresolved status to have an owner without becoming an accusation. A missing declaration can persist for several reasons, and a review should avoid treating every absence as the same kind of failure. The responsible team needs a practical next step, a way to record the response and an escalation route where appropriate. Those arrangements require the actual business context. They cannot be settled by a universal rule that every blank means zero, every blank means average or every blank should prevent service. The data account should preserve the question for the people authorised to answer it.

## Make completeness answerable to the business meaning

Completeness is a useful measure when its definition is clear. It can reveal intake failures, broken transfers or fields the business genuinely needs. It becomes less useful when the incentive is merely to remove blanks. A populated field with an unsupported meaning can be worse than an acknowledged gap because it invites a confident decision. I therefore ask what completion means for the particular attribute: supplied by whom, derived by which method, confirmed through what route and suitable for which use. The answer may differ between attributes even within the same customer file.

The board does not need to approve every data-status convention. It does need to understand whether a management account distinguishes material assumptions from evidence and whether the people making decisions can see that distinction. A CDO can illustrate the issue through a small concrete case, then explain the wider stewardship responsibility. The question is not whether the organisation has eliminated every unknown. It is whether unknowns are visible in consequential decisions and whether management has a credible route for addressing them. That is an oversight question rather than an instruction to rewrite a particular field.

I would improve the intake specification before repeating this work. The intended meaning of the turnover declaration, the handling of absence and the method of recording later confirmation should be considered when the workflow is designed. Discovering those questions during a downstream analytical exercise is more expensive and leaves more room for an undocumented convention to become established practice. The ₹48 lakh review was necessary in the fictional case, but its necessity should not become a reason to celebrate remedial work without examining why the original arrangement left the meaning uncertain.

An adapted executive profile should retain this level of specificity. Explain the field, the population, the proposed shortcut and the authority that remained elsewhere. Do not turn confirmations into a claim of perfect accuracy or a data correction into a measured improvement in insured losses. The story is valuable because it shows how an apparently small convention can change what the institution believes a customer said. Leadership is demonstrated by preserving that distinction, funding the work needed to resolve part of it and leaving the remaining uncertainty visible enough for someone else to make an informed decision.