A quiet complaint register is not a resident's answer

The complaint account made some streets highly visible and left others almost absent. Our proposed research programme initially followed that pattern, concentrating field attention where contacts were most frequent. It seemed responsive: the company would listen where people were already asking to be heard. The difficulty was that the study was also intended to help us understand the wider service. If we observed only the places most visible in the contact system, we would learn little about whether quieter streets were well served, unable to use the channel or simply unwilling to raise a concern through it.

The frame contained sixty streets divided into three predefined contact groups of twenty: higher-contact, lower-contact and no recorded contacts during the review period. These were descriptions of the records, not classifications of resident satisfaction. I proposed twelve street reviews, selecting four at random within each group. That design gave each part of the frame a place in the planned work. It did not make every household equally represented or settle how the eventual findings should be analysed. It gave us a clearer observation question than a study whose coverage merely reproduced the complaint register.

The channel is part of the observation

A recorded contact depends on more than the underlying service experience. Someone must know the route exists, be able to use it, decide that reporting is worthwhile and have their concern recorded in a usable form. Different streets may have different relationships with that process. Silence therefore cannot be read as a satisfactory-service response without further evidence. The point is not to distrust every complaint statistic. It is to understand the process that produces it, so the company does not confuse a measure of expressed concerns with a complete account of the concerns that exist.

The higher-contact streets still mattered. Their residents had raised issues, and those issues required the appropriate response. A broader research frame should never become an excuse to dismiss active concerns as merely an artefact of reporting behaviour. I wanted both obligations recognised: act through the established service processes and examine the coverage of the evidence used for wider conclusions. The company could do one without pretending it had accomplished the other. That distinction made the discussion more demanding, but it avoided an unhelpful choice between listening to visible complaints and questioning what the complaint account left unseen.

Our three groups were useful for selecting the planned reviews because they described the particular gap we wanted to investigate. They were not permanent labels for the streets or claims about the people living there. Contact levels could change over time, and a street could contain residents with very different experiences. Any eventual account would need the period, frame and selection method preserved. I did not want the categories to become a shorthand that followed a neighbourhood into unrelated decisions, particularly when the original purpose was to challenge the assumption that absence from a channel meant absence of a problem.

Research coverage and service response are different decisions

The service sponsor worried that distributing research effort would reduce the depth available in the streets already asking for help. That was a real trade-off. A finite programme cannot examine every location equally and immediately. We agreed to distinguish the evidence programme from the response to current operational complaints. Urgent concerns continued with their established owners. The data and research function proposed where to learn about the service; it did not decide municipal priorities, remove operating obligations or instruct frontline colleagues to withhold attention until a study had produced a result.

The programme received ₹76 lakh authority for the defined work. Its cost covered more than a list of locations: access arrangements, field preparation, the conduct of reviews and the handling of the resulting evidence all required resources. At the stated cutoff, ₹61 lakh had been spent and ₹15 lakh remained unspent. That remainder was not a saving against a completed equivalent scope. Some planned work was still unfinished. Keeping the financial and field accounts together prevented us from presenting the smaller expenditure as though it demonstrated a more efficient delivery of the entire intended programme.

Selecting four streets within each group made the design transparent, but selection alone could not ensure a complete or unbiased result. The way visits were conducted, who was reached and what remained unavailable would matter. A random selection step does not make every later source of missingness disappear. I wanted those conditions considered before the eventual interpretation was drafted. Qualified field colleagues owned the practical method, and any broader inference would need to reflect the actual achieved coverage. We did not promise that twelve planned street reviews would provide an all-purpose measure of satisfaction across seven towns.

An unfinished visit must stay in the account

Ten reviews were complete at the reporting cutoff. Two selected streets remained postponed because access arrangements had not been finished. The easiest presentation would have replaced them with locations that could be reached quickly and described the programme as complete. We kept the original selections and their status visible instead. A substitution can be appropriate under a designed method, but it should not happen quietly merely to satisfy a reporting date. The reason a location is hard to reach may itself matter to the coverage question the programme was meant to examine.

The incomplete work did not support a finished comparison across the three groups or a population estimate for the sixty-street frame. We did not invent an answer for the unvisited streets, nor assume their experience matched the ten completed reviews. That left management with a less conclusive account at the cutoff. It also gave the next review an honest starting point: which work was done, which remained, what had been spent and why the current evidence could not yet answer the broader question. Unfinished research can still inform planning if its status is preserved instead of disguised as a final result.

I would now examine access to the reporting channel when defining a service-evidence programme, before the contact counts become the default map of where the company should learn. That does not mean assuming quieter places have worse service. It means refusing to fill an absence of evidence with the most reassuring explanation. The frame should make room for the possibility that the channel sees different parts of the community differently. The resulting study should also be willing to find that some quiet streets are well served, rather than treating the original concern about invisibility as a conclusion it must confirm.

For a board, the broader question is whose experience enters management's account and whose experience is inferred from silence. Directors can ask how reporting access was examined, how unfinished observations were handled and whether a research programme has been confused with the operational response it is meant to inform. They do not need to select the streets themselves. They need assurance that the evidence process can reveal something beyond what the existing channel already makes visible. Our twelve planned reviews, ten completed visits and two postponements describe that effort honestly, with further work still needed before the company can claim the answer it set out to find.