Activity is not yet an outcome
In a group, the number of messages tells us how much was said. It does not explain on its own what needs choosing or what happens next. From a product perspective, a more useful question is what a member can find when returning a week later: the original concern, the reasons considered and the status of the choice. This is our editorial lens on Deferendum, not a claim about user research or measured community outcomes.
A thread from need to choice
Deferendum connects a concern, arguments and a vote under the group’s rules. A proposal does not automatically become a decision. The design implication is interesting: the interface needs to show the stage of the process and which conditions remain unmet. In a hypothetical shared-space example, the question is more than which activity members prefer. It also concerns the need being addressed, the options considered and the information that might change the choice.
The work comes after the vote
In the product, proposed responsibility can be accepted or declined; progress and completion confirmation remain connected to the adopted decision. People do the implementation. As a product principle, this invites a clear distinction between an intention, a commitment and completed work. A check mark that is too easy to obtain could make a process appear finished before the group understands what happened. The clarity of the status deserves assessment separately from the appearance of the screen.
AI can help with understanding
Deferendum describes a private ChatGPT pilot tested with fictional data: six read-only tools, without voting or automatic saving. There is no public installation. The editorial interest is the difference between explaining and deciding. A useful summary should retain uncomfortable objections, identify the information consulted and distinguish proposed wording from the adopted question. These are evaluation criteria, not guaranteed generation results.
Measure continuity as well as participation
A proposed evaluation would examine whether members can explain the choice and locate the next step without help. Other useful questions include how many decisions have accepted responsibility, where updates stall and what a newcomer understands from the available history. We assign no values to these indicators. They offer a way to test the product’s direction. Deferendum is a case of coordination and collective decision-making; its usefulness should be judged in that context rather than through the digital twin label. This analysis does not demonstrate product adoption; it offers concrete questions for a future evaluation with real people and situations.