Lean tools

Process flow analysis

In French: Analyse de déroulement

Process flow analysis records, in order, every step a product or a file goes through, classifying it by nature — operation, transport, inspection, delay, storage — to bring out what contributes nothing.

The record and its five natures

An object is followed — a part, a file, a patient — from start to finish, and each step is noted with its nature, its duration and the distance travelled where relevant.

The operation transforms: it is the only nature that creates value. Transport moves. Inspection checks. Delay does nothing. Storage keeps while awaiting a decision.

The record is made on the shop floor, actually following the object, not in a meeting room from what everyone believes. That is what distinguishes process flow analysis from a flowchart reconstructed from memory — and the gap between the two is systematically instructive.

What the count reveals

The first figure that strikes is the proportion of operations in the total. On an administrative process, it is not rare for them to represent less than one step in five — the rest is movement, checking and waiting.

The second is the ratio between actual working time and elapsed time. A file handled in forty minutes of real work can take eleven days to cross the department. That ratio, often below one per cent, points to where the lead time is decided — and it is almost never the processing speed.

The third is the total distance travelled. Adding up the metres from one station to the next regularly produces kilometres per day nobody had any idea of, and makes concrete a re-layout that was being discussed in theory.

What is done with it next

The record improves nothing by itself: it provides the list to which the ECRS grid is applied, operation by operation. It is the natural sequence of the two tools, and one without the other remains incomplete.

It also feeds the layout decision: repeated transports between two distant stations can be read immediately on a trace of the route overlaid on the floor plan — which the spaghetti diagram puts into pictures.

The common mistakes

Recording a theoretical flow. The process described in the procedures and the one that takes place always differ, and it is precisely the gap that holds the potential. A record made in a meeting produces a document that is correct and useless.

Following only one object. A standard file crosses the department differently from an incomplete one, and the special cases often represent most of the load. Two or three contrasting journeys are worth more than a single carefully chosen one.

On the shop floor — Hospital — patient pathway

The record of an outpatient consultation pathway counted thirty-four steps for eight minutes of actual medical time. The eleven waits added up to two hours forty. None of the solutions considered until then — increasing consultation time, adding a practitioner — addressed the waits.

The other tools illustrated by this trade: healthcare and life sciences.

Frequently asked questions

What is the difference from value stream mapping?

Process flow analysis follows an object through the detail of the physical steps, at the scale of a department or a workshop. Value stream mapping takes a higher view over a whole flow, adding the information flows and the stocks between processes. The first prepares a station project, the second a flow strategy.

How long does a record take?

From a few hours for a station to several days for a pathway crossing several departments. The real constraint is elapsed time: following a file that takes two weeks to come out requires two weeks, or a reconstruction from existing timestamps.

Does every step have to be timed?

The operations yes, with reasonable precision. Waits are dated rather than timed — the arrival time and the resumption time are enough, and it is often the only practicable measurement on a long process.

Does the record disturb what it observes?

Yes, a little, and it must be said: someone who knows they are being observed works differently. Clearly announcing the object of the record — understanding the pathway, not judging people — reduces the effect; observing several times makes it disappear.