Frequently asked questions about thinking, epistemology, and cognitive tools. 9738 answers
Treating input-output specification as obvious and therefore not worth writing down. The failure is not imprecise specification — it is absent specification. You know what the input 'should be' and you know what 'done' looks like, but you have never made either explicit. The result is that every.
Define clearly what goes into each workflow and what comes out. Without precise input-output specification, you cannot chain workflows, automate steps, or diagnose failures.
Where one person or system passes work to another is where errors are most likely.
Where one person or system passes work to another is where errors are most likely.
Where one person or system passes work to another is where errors are most likely.
Where one person or system passes work to another is where errors are most likely.
Where one person or system passes work to another is where errors are most likely.
Where one person or system passes work to another is where errors are most likely.
Choose a workflow that involves at least one handoff — work passing from you to another person, from one tool to another, or from your present self to your future self (a project you set down and pick up later). Map every handoff point in that workflow. For each handoff, answer four questions: (1).
The most common failure is assuming that the handoff is the other party's problem — that if the recipient does not understand, they should have asked better questions. This reverses accountability. The sender is responsible for ensuring that the handoff contains sufficient context for the receiver.
Where one person or system passes work to another is where errors are most likely.
You cannot improve a workflow you do not measure. Track cycle time, throughput, error rate, and energy cost — but track them lightly, because invasive measurement distorts the very process you are trying to understand.
You cannot improve a workflow you do not measure. Track cycle time, throughput, error rate, and energy cost — but track them lightly, because invasive measurement distorts the very process you are trying to understand.
You cannot improve a workflow you do not measure. Track cycle time, throughput, error rate, and energy cost — but track them lightly, because invasive measurement distorts the very process you are trying to understand.
You cannot improve a workflow you do not measure. Track cycle time, throughput, error rate, and energy cost — but track them lightly, because invasive measurement distorts the very process you are trying to understand.
You cannot improve a workflow you do not measure. Track cycle time, throughput, error rate, and energy cost — but track them lightly, because invasive measurement distorts the very process you are trying to understand.
Pick one workflow you execute at least weekly. For the next three executions, record four numbers: (1) cycle time — wall-clock minutes from start to finish, (2) touch time — minutes you were actively working versus waiting, (3) error count — how many times you had to redo, correct, or recover from.
Measuring so many things that the measurement itself becomes a workflow burden. You install time trackers, build dashboards, tag every task — and then spend more time maintaining the measurement system than improving the workflows it was supposed to illuminate. The opposite failure is equally.
You cannot improve a workflow you do not measure. Track cycle time, throughput, error rate, and energy cost — but track them lightly, because invasive measurement distorts the very process you are trying to understand.
After each execution look for one thing to improve in the workflow.
After each execution look for one thing to improve in the workflow.
After each execution look for one thing to improve in the workflow.
After each execution look for one thing to improve in the workflow.
After each execution look for one thing to improve in the workflow.