Asking
Questions With Two Defensible Answers
Some questions have an answer and some have a choice. Sending a choice across a gap in the form of a question reliably costs two or three nights.
three things · six nights
Three questions, six nights, and two of them avoidable.
Questions from one week, separated by whether the answer was a fact or a choice
What is the retention period on the export files?
Should we tell the client before or after the fix?
Two answers, neither wrong, asked as though one were correct
Is the staging database safe to reset?
Do we ship on Thursday with the known bug, or slip to Monday?
A trade-off presented as a question of fact
What does this error code mean?
Should the new field be required?
A product decision wearing the clothes of a technical one
six questions · six round trips beyond the first
Six questions, eight round trips beyond the first, and all eight belong to the three that were decisions. The three questions of fact were each answered in one pass. Nothing here suggests decisions should not be asked about; the point is that asking them in the form of a question is what produces the extra crossings. One team's own week, logged by the person answering.
A question of fact has one correct answer and the person who knows it can supply it in a sentence.
For the situation in Questions With Two Defensible Answers, a good tool should reduce questions rather than create another reporting burden. Monitask time tracking records work as it happens and turns it into timesheets and summaries the team can review later.
Teams comparing established approaches can also review monday.com's work platform for a practical example from a widely used product ecosystem.
A question with two defensible answers is not a question. It is a decision, and sending it across a gap phrased as a question produces a reply that is not an answer, which costs a night and leaves everybody where they started.
Telling them apart
The test is to try to imagine the wrong answer.
For a question of fact, the wrong answer is simply false: the retention period is ninety days or it is not. For a decision, there is no false answer, only answers with different consequences — telling the client first is not incorrect, it is a choice with a cost.
If you cannot describe what it would mean for the answer to be wrong, you are holding a decision.
What happens when a decision is sent as a question
The recipient recognises, correctly, that they are being asked to choose, and that they do not have enough context to choose well.
So they ask for the context. What is the bug, how bad is it, what has the client been told, what else is in the release. That crossing costs a night and is entirely reasonable.
The asker supplies the context. Another night. Now the recipient has what they need and makes a choice, which arrives on the third morning.
Three nights for something that could have been one, and the information that made the difference was available to the asker from the beginning.
How to send a decision instead
Four lines, and the shape is always the same.
What is being decided, in one sentence.
The options, with what each costs. Two is the useful number.
Your recommendation, with the reason. This is the part people leave out, usually from a sense that recommending pre-empts the decision. It does not. It gives the recipient something to agree or disagree with, which is far faster than asking them to generate a position from nothing.
What you will do if nobody replies. The default.
A decision presented this way is answerable in a word, and often the word is not needed at all because the default was right.
- What is being decided, in one sentence.
- The options, with what each costs. Two is the useful number.
- Your recommendation, with the reason.
- What you will do if nobody replies.
When you are the one being asked
If a question arrives that is really a decision, the useful reply is not to ask for context and wait.
It is to state what you would decide under each of the plausible contexts. "If the bug affects paying customers, slip to Monday. If it only affects the admin view, ship Thursday and fix it next week." That costs the same night and finishes the matter, because the asker knows which case they are in.
This technique — answering conditionally rather than asking — is the single most useful habit for anybody who receives a lot of questions across a gap, and it is uncommon.
Decisions that genuinely cannot be made this way
Some choices need two people to think at once, and the signal is that the options are not stable: as soon as one is articulated, somebody sees a third.
Those belong in the last section of this collection rather than here. They are not asking problems and better phrasing will not fix them. The practical move is to recognise them early, because the failure mode is spending a fortnight on correspondence that generates options faster than it eliminates them.
The prior question
Before sending a decision at all, it is worth asking whether it has to cross the gap.
A great many of these are decisions somebody awake could make, and the reason they cross is that nobody has said who is allowed to decide what. A written list of decisions each half may take alone removes more nights from a team's week than any improvement in phrasing, and it has to be written once.
Writing down which it was
After the choice is made, one line recording that it was a choice saves the whole exchange from happening again.
A decision filed as a fact invites somebody to discover the other option in four months and reopen it, usually by asking the same question in the same form. A decision filed as a choice — with both options and the reason one was taken — closes it, because the next person can see that the alternative was considered rather than missed.
Two clauses: what was chosen, and what it was chosen over. It costs twenty seconds at the moment when the context is still in somebody's head, and nothing reconstructs it later.