Say no with a question

It is Wednesday of a two-week sprint. Someone asks whether you can also add bulk resend, since you are in that code anyway. It is three days of work and you have four days left.

You have three available responses and two of them are bad.

Yes is bad because you cannot do it, and everyone finds that out on Friday week, when it is too late for anybody to make a different plan.

No is bad for a subtler reason, and it is the reason this chapter exists. No is a claim about priority, and priority is not yours. When you say it, you are asserting an authority you do not have, in a conversation where somebody else does have it. They will notice. They will go and get the decision reversed, usually within a day, and now you are both overruled and marked as the person who says no.

You cannot refuse, so stop trying

The instinct is to look for a better-worded refusal: firmer, softer, with an explanation attached. None of that fixes it, because the problem is not tone. The problem is that you are answering a question that was never yours to answer.

What is yours is the information nobody else in the conversation has. You know what is in the sprint, what it costs, and what happens to it if this goes in. That is the scarce thing, and handing it over intact is more useful than any verdict you could offer.

So do not decline. Convert.

The move

Turn the no into a question that hands the trade-off back to the person who owns the priority.

I should say plainly that this is craft rather than evidence. Nothing measured exists on developers declining work, on being labelled a blocker, or on whether question-framed refusals protect your standing. The negotiation literature people usually reach for here has no empirical test behind it either. What follows is an argument and a script, offered as such.

THE THREE ANSWERS

no    "Can't, we're full."
        a claim about priority.
        overruled by Thursday. remembered for months.

yes   "Sure, we'll fit it in."
        true on Wednesday. false on Friday week.
        discovered by everyone at once, too late to replan.

ask   "That's about three days, and we've got four left.
       I can fit it if the reconciliation view moves to
       next sprint. Do you want me to make that swap,
       or keep the current plan and queue this?"

        a priced trade-off, handed to the person
        whose job it is to make it.

The third answer contains no refusal at all. It contains a cost, a consequence, and two options. If they choose the swap, you were never blocking anything. If they choose to queue it, they refused it, which means it stays refused when somebody asks again next week.

That is the part worth noticing. A no that you issued has to be defended every time it is questioned. A no that they chose does not.

Making it work in practice

Price it before you reply. The question only functions if it carries a real number. That’s about three days is what makes it a trade-off rather than a complaint. A vague that would be quite big invites them to decide it is not.

Name what moves, specifically. Not something would have to give. The reconciliation view, by name, moving to next sprint. People cannot trade against an abstraction, and if you make them imagine the cost themselves they will imagine it smaller.

Offer two options, not one. One option is an ultimatum wearing a question mark. Two is a decision.

And then actually accept the answer, including when it is the swap you did not want. This move only survives if the question was genuine. Ask it twice while visibly hoping for one answer and people stop treating it as a question.

What it costs

It puts real work on them, and some people will find that annoying, particularly those who were hoping you would simply absorb it. That is the trade you are making, and it is the right one, but do not pretend it is free.

It can read as procedural if you use it for everything. Small requests should just get done. Save this for things that displace other things, or you become the person who produces a trade-off matrix when asked for a favour.

And sometimes you do have to say no outright. If the request is unsafe, breaks a legal obligation, or ships customer data somewhere it should not go, that is not a priority question and you should not dress it up as one. Those are the cases where you have not just the standing but the duty to refuse, and the rest of this book is largely about earning enough credibility that the refusal lands when you use it.

Try this week

Next time you are about to type we can’t, stop and write three things first: what it costs, what it would displace, and two options.

Then send this, adjusted for your situation:

That’s about three days. We’ve got four left in the sprint, so I can fit it if the reconciliation view moves to next sprint. Want me to make that swap, or queue this for the next one?

Then wait, and do whatever they say. You have not refused anything, you have not agreed to anything impossible, and the decision is sitting with the person whose decision it was.