Step 1 of 8
FLOW · measure first
Find — collect the candidates
F is collection, not selection. The step ends with a list, not a winner. That sounds like a formality, and it is the reason most improvement projects end up working on the wrong process: prioritise immediately and you prioritise by volume of complaint, because no figure exists yet to sort by.
What is being sought is not the worst process but the complete field. A candidate missing from F cannot win in L — and expensive processes tend to stand out because they run often, not because they annoy.
- Answers the question
- Which processes are worth looking at at all?
- What the step produces
- A list of five to fifteen processes, each with a trigger, an output and a rough annual volume. No ranking, no assessment.
How the step is carried out
Fix the scope
One department, one order type, one quarter. Say “all our processes” and you get a list nobody works through. Write the scope down, or it will drift while the interviews are still running.
Read three trails at once
Backlog: where do items sit, where do people chase, where are the reminders? Manual handling: where does a person move data from one system into another? Complaint: what is complained about, by whom, how often? Each trail surfaces different processes; two trails alone give a lopsided field.
Ask, do not infer
Talk to the people who run the process, not to the people who can describe it. Three questions are enough to start: what held you up most this week, what do you regularly type in a second time, and what do you wait longest for?
Record each candidate in one sentence
Trigger, output, rough annual volume. Nothing more. Mapping step sequences here means spending days on processes that will drop out in L anyway.
Do not assess
The urge to start ranking peaks at the end of this step — the impressions are fresh and one of them was the loudest. Ranking happens in L, in euros.
Finished when
- Every candidate has a trigger and an output, both in one sentence.
- Every candidate has a rough annual volume. Estimating is allowed, guessing is not — somebody has to stand behind the order of magnitude.
- At least one person who actually runs the process has confirmed it.
- Duplicates are merged: two departments describing the same workflow under different names are one candidate.
The typical mistake
Starting with the loudest process
The loudest complaint marks the most irritating process, not the most expensive one. Irritation comes from interruptions and chasing; cost comes from volume. A task that annoys twelve times a year weighs less than one that costs five minutes two thousand times a year — and nobody complains about five minutes. In F both are worth the same: one entry on the list.
What you can compute yourself
The bottleneck calculator asks this step’s questions in a structured way and sorts the answers. It does not replace the conversations; it makes sure none of the three trails is forgotten.
The step, worked through
Every sample analysis starts from a process picked out of a field like this. Its “situation” section names the trigger, the volume and the reason it was selected — the three things F produces per candidate.
The analyses are constructed models, not client projects. Their figures show what a measurement would look like — they are not results achieved for anyone.
- Sample analysisAN-2026-08
The sales lead who shows up as two boxes
A packaging printer measures the path from incoming order to order confirmation. The bottleneck is the credit check. The second finding is a role two steps share, and it only becomes an action once the first one is dealt with.
- Sample analysisAN-2026-07
The loudest process in the building is not the most expensive
An engineering consultancy wants to digitalise leave requests. The measurement agrees with the complaint and still not with the investment. The process is slow, but cheap.
- Sample analysisAN-2026-06
The €500 approval limit nobody has touched since
A plastics processor has every purchase above €500 signed by the managing director. The limit is old, the prices are not. Four out of five orders pass through a gate that opens twice a week.
Questions
Five to fifteen. Below five the scope was drawn too narrowly and the selection in L means nothing. Above fifteen, L costs more than its result is worth, because every candidate gets its own calculation there.
No. Process mining assumes the process already runs in a system and leaves traces there. The processes that cost the most tend to run in spreadsheets, mailboxes and people’s heads. For this step, conversations are the more reliable source.
That is a finding, not a problem. Two accounts of the same workflow that do not match usually point to a media break or unclear ownership — both come back in R and E. In F it is noted and the candidate stays one candidate.
If you want to know what your process costs: measure it.
Start with the free diagnosis or download FlowVisual. If you want to speak to someone afterwards, we're reachable.
Remote · fixed price · result in euros