A regional manager asked a sensible question when we proposed reducing the detail in our employee-listening report: how could the local team act on a problem it could not locate? I did not want to dismiss that concern as an appetite for surveillance. Enterprise averages can conceal local working conditions, and a general statement about communication may leave everyone agreeing that someone else should improve it. The difficulty was deciding what level of detail would support action without turning a colleague's response into an exercise in identifying the person who supplied it.
Our 2025 exercise received 534 responses from 780 invited employees. The proposed report combined service point, role group and comments, sometimes for groups with fewer than five respondents. No names appeared in those tables. That fact alone did not settle the exposure. A manager might know that only one person in a particular role had recently raised an issue, travelled to a location or experienced a change in responsibilities. Several ordinary details could become identifying when read together by someone familiar with the team.
A name-free table can still be revealing
The listening provider described the material as confidential. I wanted to understand what that meant in the actual workflow: who could see the underlying responses, what managers would receive and what employees had been told. Confidentiality is not a decorative word to place above a chart. It describes choices about collection, access and use that need to remain consistent as a report moves through the organisation. A useful label should prompt those questions rather than replace them. I also did not want to promise complete anonymity that the process could not establish.
There was a genuine tension. A broader cluster might make the source less apparent while weakening the connection to a particular working arrangement. A very narrow breakdown could show precisely where a concern arose while exposing a colleague to speculation. The choice could not be reduced to a universal claim that more detail is better or that every local result must be suppressed. We needed to consider the decision, the population and the context already available to the people receiving the report. Context changed the meaning of the same table.
We suppressed the small-group breakdowns, removed identifying details from shared comments and limited access to the underlying records. A ₹9 lakh external-analysis budget paid for the revised work. That was an incurred organisational choice, not a free improvement in presentation. Some managers received less detail than they had requested. I accepted that consequence because the added granularity did not always provide an equivalent benefit for the action they proposed. It would have been misleading to describe the revised report as delivering every original analytical preference without a trade-off.
The decision also required attention to how managers talked about the material. A report can omit a narrow table while a meeting recreates it through questions about who was absent, who used a certain phrase or who had complained before. We explained that the purpose was to examine working conditions and management responses, not to reconstruct individual participation. That explanation did not prove that every future conversation would be handled well. It gave the organisation a clear expectation against which conduct and follow-up could be considered.
Decide what the detail is meant to change
Lalita Zevnathi, our COO, supported the revised approach on the condition that it still led somewhere useful. She did not want privacy concerns to become a reason for issuing a broad report and doing nothing. We retained enterprise and sufficiently broad cluster themes, assigned management responses and explained the intended follow-up to employees. Where a concern required a more specific conversation, the response could be a facilitated discussion of the work rather than an attempt to identify the survey respondent. The route to action did not have to reproduce the route by which the original comment had arrived.
This distinction is easy to lose when a survey becomes a management score. A leader may concentrate on whether the result makes the team look good and whether an unfavourable comment can be explained away. The organisation then risks spending more energy interpreting the respondent than addressing the condition. I prefer to ask what decision the material can responsibly inform. It may support examining workload, clarifying a role or reviewing a process. It may not establish why every person feels as they do or who should be blamed for an aggregate result.
There is also a limit to what a listening exercise should carry. A confidential grievance, a safeguarding concern or an allegation requiring a formal process may need a different route and properly qualified handling. A survey report should not be presented as a substitute for all those responsibilities. My role is to make the purpose and routes clear, with appropriate professional owners involved. I would not promise an employee that placing a concern in a questionnaire necessarily starts or completes the process their circumstances require. Clarity about the channel is part of respecting the person using it.
Keep nonresponse and uncertainty in the account
The 246 people who did not respond remained outside the respondent account. We could report that 534 responses had been received; we could not describe those responses as the opinion of all 780 employees. Nor could we infer a single reason for nonparticipation. Some may have lacked time, some may have seen little relevance and others may have had concerns about the process. The numbers did not resolve those possibilities. Treating silence as either satisfaction or distrust would have supplied an explanation the exercise had not established.
The revised reporting approach likewise did not prove an increase in trust. We had made a choice to reduce unnecessary exposure and preserve a practical follow-up. Whether employees experienced the wider organisation differently required further evidence and sustained conduct. I wanted that limit in the management account because a well-intentioned process change can become self-congratulation very quickly. A leader who announces that confidentiality has been achieved may stop asking the questions that made the improvement necessary. A more modest account leaves room to learn where the arrangement still fails people.
For a board, the useful discussion concerns the quality of management's listening and response, not access to every underlying comment. Directors can ask whether the reporting population is clear, whether the use of the material matches the explanation given to employees and whether action follows without unnecessary identification. My experience would inform those questions, but it would not make me the recipient of every individual case. Listening earns its place in governance when it helps the company understand and improve the work. Detail is valuable to the extent that it serves that purpose responsibly.