When your work depends on other people
The short answer
Before asking what is most important today, ask what is blocked, and by what. The second question is cheaper and it changes the answer to the first.
Work waiting on someone else does not move however high you rank it. Meanwhile the thing that would unblock it is usually small — a message, a decision, a file — and it stays invisible because nobody writes down two-minute tasks that belong to other people.
Send the unblocking messages first, before the day fills. Then work on what is actually yours. This is the single highest-return habit in any job with dependencies, and it takes about ten minutes.
The name for the worst version
When a low-priority item holds something a high-priority item needs, the important work is stuck behind the trivial work. The formal term is priority inversion, and the size of the blocker tells you nothing about the size of what it is blocking.
The famous case is Mars Pathfinder, 1997. The rover kept resetting on the surface of Mars because a low-priority meteorological task held a lock that a high-priority bus-management task was waiting on. The high-priority task was not deprioritised and was not slow. It was blocked by something that did not matter. Engineers identified the inversion and patched the scheduler remotely.
The office version is a quarter's work waiting on a two-line approval nobody has sent. Once you know it has a name, you start looking for it — which is most of the value.
A ten-minute morning routine
- List what is waiting on someone. Be specific about who and what.
- Send every unblocking message now. Short, one ask each, with a date. "Can you confirm the pricing by Wednesday? I can't send the deck without it."
- Give each one a check-back date, in writing, somewhere you will see it. An unanswered message is not a resolved dependency, and the failure mode of this whole approach is assuming that asking is the same as being answered.
- Then choose today's work from what is genuinely yours.
Step 3 is the one people skip, and it is the one that keeps a dependency from silently becoming a crisis.
How to ask so that it works
One ask per message. Two asks in one message reliably get one answer, and you will not control which.
Say what it blocks. "This is holding up the client deck" converts your request from an interruption into information about their own priorities. Most people respond to that; very few respond to urgency alone.
Give a date, not "ASAP". ASAP has no position in anyone's queue, and work with no date never reaches the front.
Make the answer cheap. Offer the default: "Unless you object by Thursday I'll go with option B." This converts a task on their list into a non-task, and it is the fastest way to unblock anything.
What to do while you wait
Do not keep it open. A blocked item still in progress is still counted against everything else. Park it properly — written down, with what it is waiting on — so it is out of the working set rather than half-live.
Leave yourself the note. What was pending, what you had already decided. When the answer arrives a week later, that note is the difference between resuming in two minutes and re-deriving the whole thing.
Do not build on the guess. Continuing on an assumed answer is usually more expensive than waiting, because the rework is invisible until the real answer arrives and contradicts it.
The honest limit
Some dependencies do not move. The other team is genuinely overloaded, or the decision is genuinely not being made.
At that point the useful thing is not a better message. It is telling whoever owns the outcome that it will be late and why, early enough for them to do something. That is a conversation rather than a technique, and delaying it never makes it smaller.
Where to go next
- Why your important work never gets done — starvation, priority inversion, and aging
- Why deadlines work, and what they cost — why a date beats "ASAP"
- How to prioritise your tasks — the three conditions
Using this in Decide One
Decide One records a line as done, partial, or blocked — three states, because they are three different things and collapsing them loses the one that needs somebody else.
A line you did not finish is unfinished. It is not marked against you, and a day you did not open is not counted.
Open Decide One — free to use, with no account. Nothing you write leaves your device.
Decide One is a priority instrument. Bring the day into view, choose a method, give the first thing real time.