The phone number is correct, but the communication may still be wrong

A notification system can report an excellent delivery rate while knowing surprisingly little about the circumstances in which its messages are read. The number is valid, the device receives the message and the transaction is marked complete. Yet the device may be shared. Another person may see the text first, or the recipient may prefer a different channel for a particular kind of communication. These possibilities do not make messaging unusable. They make delivery an incomplete account of whether the communication was appropriate.

At Fervali Hearing Services we reviewed preferences across ninety-six accounts and found thirty-one shared-phone concerns. That finding changed the starting point of our discussion. We had been looking at the system's ability to reach a device. We also needed to understand what a message could reveal and what the person using the service wanted us to do. A phone number was a route, not a full description of the audience at the other end.

The service team initially favoured keeping the familiar reminder wording because callers recognised it and staff could explain it easily. That was a reasonable concern about continuity. I argued that familiarity for our colleagues was not enough if the wording created discomfort in a shared-phone setting. We accepted the work of revising scripts and testing clearer choices rather than treat the existing text as fixed simply because it had been used for some time.

## Delivery does not describe the audience

The ₹22 lakh redesign introduced neutral notifications and clearer handling of preferences. Clinical, service and qualified privacy colleagues considered the responsibilities within their own remit, while IT implemented the agreed behaviour. I did not ask a delivery statistic to answer a legal question or claim that the redesign settled every applicable requirement. The technical team could show what the system sent and how a choice was recorded. Other judgements required the appropriate people and a fuller understanding of the service.

Neutral wording also involved a tradeoff. A shorter, less revealing message might give the recipient less immediate context. That could produce questions for the service team or require another step before an appointment detail was clear. We had to account for that work rather than present the change as a costless improvement. The purpose was to make the communication more suitable for the circumstances identified, while preserving a usable way for the person to obtain the detail they needed.

I found it helpful to separate three questions. Did the system use the recorded channel? Did it apply the chosen communication settings? And were those settings appropriate and current for this person and this type of message? The first two can be examined through system behaviour. The third may require a conversation and professional judgement. Combining all three into a single success measure makes reporting simpler but leaves directors with a less useful understanding of the service.

The reviewed population matters too. Thirty-one concerns among ninety-six accounts told us something specific about that review. It was not a basis for declaring that every shared phone presents the same issue or that every other account is free of concern. People organise communication in different ways. A business should avoid replacing one unsupported assumption with another merely because the new assumption sounds more cautious. The review should help improve choices, not classify people permanently by a device arrangement.

There is also a difference between recording a preference and making it usable by colleagues. If a choice is visible only in one screen, a separate calling process may miss it. If the wording is ambiguous, two staff members may act differently while both believe they have followed it. Technology design therefore needs a service account of how the preference will be used across relevant interactions. The objective is not to collect another field; it is to support a decision that someone must make.

## A preference is a continuing conversation

Circumstances change. A person may obtain another phone, share a device for a period or prefer different communication while travelling. A preference process needs a sensible route for revision. That does not mean asking the same questions at every interaction without purpose. It means avoiding a design in which an old entry acquires permanent authority merely because nobody knows how to revisit it. The service should be able to explain how a person changes a choice and what happens after that request.

Failed delivery also needs careful interpretation. A message that does not arrive is a service problem to investigate, but it does not automatically justify using any alternative contact available in the record. The appropriate next step depends on the service, the communication and the person's circumstances. I want that decision to be designed with the relevant service and privacy owners rather than left to a technician's improvisation or an automatic rule whose consequences nobody has examined.

For a board, the useful discussion is not about approving individual reminder text. Directors can ask how management knows the channel is suitable, how preferences are maintained and what happens when the process fails. They can also ask which results are measured and what those measures cannot establish. A high delivery rate may remain useful for operational monitoring. It becomes misleading only when it is allowed to stand for a broader claim about appropriate communication that the underlying evidence does not support.

The work did not make every future message unproblematic. It gave us a better response to the concerns identified and a clearer basis for reviewing the process. I consider that a credible outcome. We changed wording, clarified choices and accepted some additional service effort, while keeping the scope of the claim bounded. An improvement does not need a universal promise to justify its cost. It needs a clear account of the problem addressed and the responsibility that continues after implementation.

I now pay closer attention when a technology report describes a person through a technical endpoint. An email address, phone number or portal account may be necessary to deliver a service, but it is not the whole context in which that service is experienced. The CIO's contribution is to connect the technical account with the decisions other people must make, and to preserve the questions that a successful transmission cannot answer. That is where a communication system becomes part of responsible service design rather than merely a reliable way to send text.