Automating a workflow you have not mapped is how people spend money to make the same mess run faster.
This template maps one workflow on one page. The output is not a diagram for a presentation. It is a list of steps, who does each one, how long each waits, and which ones a customer would notice if they vanished.
How to use it
Pick one workflow. The narrower the better: not "sales", but "quote requests that arrive by email". Map it as it actually runs today, not as it is supposed to run. The gap between those two is the finding.
Fill it in by watching, not by remembering. Watching takes an hour and catches the steps people have stopped noticing because they do them without thinking.
The template
Workflow:
Starts when:
Ends when:
Who touches it, in order:
Steps
| # | Step | Who does it | How long it takes | How long it waits | Would a customer notice if this vanished? | |---|---|---|---|---|---| | 1 | | | | | | | 2 | | | | | | | 3 | | | | | |
How long it takes means active work. How long it waits means it is sitting in an inbox or on a desk. The second column is usually larger than the first, and it is the one nobody has ever measured.
The handoffs
Every point where the work moves from one person or system to another. Each handoff is a chance for the work to be dropped, and most of the elapsed time in a workflow lives in these gaps.
| From | To | What is handed over | How (email, system, verbally) | |---|---|---|---| | | | | |
The steps that exist for no reason
List anything here that nobody could justify when asked. Do not delete them yet. The point of this list is that it is usually longer than expected, and it is the cheapest place to start.
1. 2.
What is done twice
Any information captured more than once. Every duplicate capture is a chance for the two copies to disagree, and the disagreement is what produces the rework.
| Information | Captured first in | Captured again in | Which one is trusted? | |---|---|---|---| | | | | |
Reading the result
Three signals, in order of how much they are worth:
The longest wait, not the longest step. A workflow where a customer waits two days for a quote is not fixed by making the quoting step faster. It is fixed by removing the wait.
The most handoffs. Four people touching one request means four chances to lose it. Collapsing to two is usually worth more than any tool.
The duplicate captures. These are the clearest automation candidates, because the fix is mechanical and the benefit is immediate.
Map before you automate. The map costs an hour. Automating the wrong step costs the project.