In 30 seconds

Before mapping the steps, pin down two things: what starts the workflow, and what counts as a finished, good result. Everything in between is easier to map once the ends are fixed.

Why this comes up

Maps sprawl when the boundaries are vague. Without a clear start, you keep adding earlier steps. Without a clear end, you keep adding later ones. Soon you are mapping the whole company.

Fixing the input and output turns an endless process into a bounded workflow you can actually reason about.

Find the input

The input is what triggers the work and the information it arrives with. Ask:

  • What event starts this? An email, a form, a scheduled time, a status change?
  • What information comes with it, and what is missing?
  • How reliable is that information when it arrives?

Be specific. "A request comes in" is too vague. "A customer submits the returns form with an order number and a reason" tells you what you actually have to work with.

Define the output

The output is what a good result looks like. Ask:

  • What has to exist or be true for this to be done?
  • Who receives the result, and what do they need it to contain?
  • How would you know it was done badly?

A useful output definition is observable. "The ticket is created in the right queue with the customer, issue and priority filled in" is checkable. "The request is handled" is not.

A small example

For a supplier invoice workflow:

  • Input: an invoice arrives by email, usually as a PDF, sometimes with a purchase order number.
  • Output: the invoice is either approved for payment with a matched purchase order, or returned to the supplier with a clear reason.

Now the middle of the map has a job: get from that input to one of those two outputs.

What this does not solve

Clear inputs and outputs do not guarantee the input is reliable. A workflow can start with information that is often wrong or incomplete. That is a real finding, and it usually points to work needed in the digitize stage.

Good enough to move on when

  • The trigger is a specific, observable event.
  • You know what information arrives with it.
  • You can describe a good result in a checkable way.
  • You know who receives the output.