A percentage can acquire authority faster than the explanation of how it was produced. In 2023 a study of 1,250 online volunteers found 54% brand recognition within that sample. The proposed advertising headline described recognition across India. Seven hundred and thirty participants came from metropolitan areas and 520 from other connected locations. The study had not recruited offline audiences. Its result was not fabricated, but the proposed use asked it to speak for a population the recruitment process had not adequately represented.
Precision does not define the audience
The first task was to distinguish the number from the inference. Fifty-four per cent described responses in the study. It did not by itself establish who else would have responded similarly, how non-participants differed or what people without the same means of participation would say. A large-looking sample can make those questions feel secondary, especially when a campaign team needs a concise proof point. Yet the denominator is not merely the number printed below the percentage. It includes the account of how those people came to be in the study and which intended audiences remained outside it.
The distinction was commercially inconvenient. We had spent ₹28 lakh, and the research sponsor wanted a visible result that could support the next campaign. The agency proposed weighting the existing responses. Weighting may be part of a qualified research approach, but it could not be treated as a magic way to create observations from people the process had not reached. I asked the team to preserve what the study actually told us and withdraw the broader headline. The decision did not require calling the original research worthless. It required refusing a use its evidence did not support.
That narrower account still mattered. The volunteers had responded to particular questions in particular circumstances. Their reactions could help us develop hypotheses, compare aspects of the material or decide what needed further enquiry. The problem arose when a useful description of those responses became an assertion about the whole market. Research can serve a valid purpose without supplying every claim a marketing team would like to make. A leader should be able to defend the useful limited purpose while also saying clearly that the more ambitious question remains unanswered.
Do not make the missing respondents disappear
We authorised ₹64 lakh for a separate mixed-mode inquiry, with qualified researchers responsible for design, recruitment and interpretation. That work was still in progress at the December 2023 reporting point; this account makes no assertion about its later findings. We did not invent a final recognition figure or claim that the additional allocation had already improved trust. The new design was an attempt to ask the broader question more responsibly, not an outcome in itself. Adding another mode could improve access while leaving selection, non-response and measurement questions that the research team still needed to examine and explain.
It was equally important not to confuse recognition with trust. Someone may recognise an insurer's name without trusting it, considering its products or understanding its service. A recognition response does not establish a policy purchase or a satisfactory claim experience. Marketing language can slide between those concepts because each sounds like evidence of a stronger brand. The research account must keep the construct being measured explicit. Otherwise a perfectly accurate percentage can become the first step in a chain of increasingly unsupported commercial conclusions.
The original expense remained in the financial account. We did not erase ₹28 lakh because the countrywide headline was withdrawn, nor call the entire amount wasted to justify the new programme. The earlier study had a narrower value and an inappropriate proposed application. Those are different judgements. A useful review asks whether the original brief was clear about its intended decision and whether the commissioning team understood the limits of the method. If the public claim was not discussed until after results arrived, the organisation had created an avoidable conflict between the evidence it purchased and the statement it wanted to publish.
This is why I want marketing and research colleagues to agree the intended use before fieldwork. Is the purpose to explore language, compare propositions among a defined group, describe a population or support a public claim? Those aims may require different designs and different cautions in interpretation. The answer should influence recruitment, questions, reporting and the eventual communication. A research brief that merely asks for a brand score can postpone the most important decision until the organisation has already paid for a result and become emotionally invested in making it useful.
Commission the question before buying the headline
The agency's concern about delay was legitimate. Research can become an indefinite reason to avoid a commercial decision. A better inquiry therefore needs a purpose, timetable and account of what management will do with the result. It may support a campaign, reveal that the proposition needs revision or show that operating work should precede more communication. None of those outcomes should be predetermined by the desire to recover the research expense. The value of the work lies in improving the next decision, which may include deciding not to make the public claim originally imagined.
I also want the reporting format to preserve uncertainty where it affects that decision. A concise executive summary should not bury recruitment conditions so deeply that the headline becomes the only remembered fact. That does not require presenting every methodological detail to every audience. It requires retaining the few distinctions that change the permissible inference: who participated, how they were reached, what was asked and what the result does not establish. Clear boundaries can make research more usable because colleagues know which questions it can answer and which still require another source of evidence.
At board level, the issue is not whether directors can personally validate every survey method. It is whether management has access to appropriate expertise and presents the scope of its evidence honestly when seeking expenditure or making claims about the market. A board should be able to ask why a finding supports the proposed decision and what could make that interpretation wrong. The answer should connect the research design with the commercial inference, rather than rely on the size of the sample or the precision of the percentage as substitutes for that connection.
My lesson from the 1,250-person study is to keep the population beside the number from the first discussion to the final campaign. The volunteers' responses were real evidence about a defined exercise. They did not need to be promoted into a national verdict to have value. Withdrawing the broader headline preserved that value while acknowledging the unanswered question. The more disciplined next step was to commission work suited to the decision, accept that its result was not yet known and leave the public promise no larger than the account the research could responsibly support.