A sharper image does not settle where a feature is

The image in our proposal was clear enough to encourage confidence before anyone examined the performance wording beside it. I endorsed three-centimetre ground sample distance as though it established three-centimetre positional accuracy. Those are different properties. One described the spacing represented in the imagery; the other required evidence about position against an appropriate reference. I knew that distinction in principle. In the proposal meeting, I allowed the readily communicated number to do work it could not support. The resulting problem was an executive commitment, not simply a terminology error for a specialist to correct after the sale.

The commercial team's attraction to the number was understandable. Clients need to compare propositions, and a concise specification gives them something tangible. A long qualification can appear to weaken a good product or make a supplier sound evasive. Yet a concise claim remains useful only when its meaning is stable. If the seller means image detail and the buyer hears positional performance, both parties can agree to the same sentence while agreeing to different services. The CTO has to notice that ambiguity before a successful sales conversation makes it harder to correct. The difficulty is not solved by moving technical explanations to an appendix nobody uses.

The property the picture could not prove

Arun Jezmavi, who heads our qualified survey practice, asked for the comparison behind the positional claim. Twenty-four independent checkpoints, withheld from the processing fit, were examined against the practice's predefined criterion. Seventeen results were within that criterion; seven were outside. This examination did not substantiate the statement I had endorsed. It also did not establish a universal quality score for our whole service or prove that all imagery from the programme was useless. The result had a defined purpose and scope. It showed that the proposed positional commitment lacked the evidence it needed for this decision.

The appropriate response began with the claim, not with finding another attractive statistic. We withdrew the unsupported positional statement and held release for the relevant acceptance decision. A sponsor suggested emphasising the image specification while investigation continued. That would have been reasonable if the scope and supported use were actually changed and clearly understood. It was not reasonable to retain the original implication under a more prominent visual description. The client needed to know what had changed in the proposition, because the intended use might depend on the property we had not established. Rewording alone could not make the original promise valid.

There is a wider temptation to treat every technical product as though one specification can summarise its fitness. A spatial product may have detailed imagery, incomplete ground observation, a particular positional result and limits on the uses of its sources. Improving one property need not resolve the others. A more detailed image can make a positional mistake easier to see without correcting it. That is why the receiving use matters. The evidence required for one decision may be inadequate for another, even when both users are looking at the same impressive representation. The service must explain its supported purpose without suggesting that every possible use has been examined.

An examination with a defined purpose

I did not ask Arun to change the criterion until the result looked acceptable. The qualified practice owned the method and interpretation appropriate to the project. Technology supported preservation of the comparison material and investigation of the processing. These responsibilities were connected but distinct. A CTO can be accountable for making a defensible commitment without personally replacing every specialist judgement behind it. In this case, accountability meant admitting that my approval had bypassed the evidence question, funding the correction and ensuring that commercial colleagues understood why the acceptance result could not be negotiated into a different meaning.

The independence of the checkpoints also mattered to the story we told. Evidence used to shape a result does not automatically provide an independent examination of that result. That observation is a reason to ask how a comparison was constructed, not a universal recipe for the number or arrangement of checks on every project. Qualified practitioners must determine an appropriate examination for the intended use. An executive should resist borrowing a sample count from a convincing case and treating it as a portable assurance. Our twenty-four points describe what happened in this fictional review; they are not a technical minimum being recommended to another survey.

Nor did seven results outside the criterion tell us, by themselves, exactly which corrective action would work. Investigation and further acquisition were necessary parts of the response, with their own decisions and limitations. We allocated ₹ 68 lakh and retained the release hold. This account offers no corrected accuracy figure because that result has not been established here. It would be easy to make the story more satisfying by adding a final score, but doing so would remove the very distinction the case is meant to preserve: an approved effort to improve the evidence is not the same thing as evidence that the desired performance has been achieved.

The cost belongs beside the corrected claim

The expense made the original error visible in the investment account. I could not describe the additional work as a measured saving, nor assume that every rupee would have been avoided by a different proposal process. What I could say was that the enterprise had committed funds to investigate and obtain further evidence after withdrawing a claim I had supported. That is a more candid account than presenting the programme as an unqualified technical advance. Commercial leaders needed to plan around the held release and reconsider which commitments remained supportable while the practice did its work. The financial and relationship consequences belonged beside the technical finding.

For a board, the useful questions are about the basis of management's promise and the response when that basis changes. Directors need not select survey parameters to ask who accepted the performance statement, what evidence supported it and how the company informed the customer of the revised position. They should also understand what expenditure has been authorised and what decision remains open. A board discussion that becomes a contest over technical vocabulary may miss the central issue. The organisation made a commitment whose meaning exceeded its evidence, and management must show how that gap is being addressed without concealing it in a different measure.

I now want the relevant practice owner involved while the performance language is still being formed. That is earlier than a final document check, when changing a sentence can threaten a timetable already discussed with the client. It also gives commercial colleagues an opportunity to explain the customer's actual use, which may reveal a better proposition than the one number initially chosen to describe it. The objective is not to burden every proposal with every possible uncertainty. It is to ensure that the important claim has an identifiable meaning, an accountable owner and evidence proportionate to the decision the client will make.

The lesson remains personal because knowledge alone did not prevent my mistake. I had worked with spatial systems for years and still let the image's apparent precision lend credibility to a separate assertion. Seniority can make that error more consequential: a CTO's agreement may end a discussion that should have remained open. A useful leadership account therefore includes the unsupported claim, the examination that challenged it, the expenditure and the unresolved result. Our product may eventually support a revised commitment, but that commitment must follow the evidence. It cannot be borrowed from how convincing the picture looks today.