Vision de Jeff Liker sur le "Lean IT"

Voici le point de vue de Jeff Liker, auteur de « The Toyota Way » et co-auteur de « Toyota Product Development System » sur le Lean Management et les Systèmes d’Informations.

Son analyse concerne surtout le rôle de support que peut jouer la fonction informatique dans un processus de production industriel :

If we think of the material and information flow diagram then IT is dealing with the information flow. When we physically transform a process by moving things around we are acting on the material flow. When we transform the information presented to help make decisions we are dealing with IT whether it is in the form of an empty space on the floor, a card, or something from a computer. The concept of value stream mapping is to design the material and information flow intentionally based on defined principles to achieve a clear business purpose.

One principle is that one piece flow is the ideal. Another is that the value of people is in doing kaizen which requires information at the right time, right place, and right amount to facilitate kaizen. When we think of things this way then it is natural to want to shape your IT to fit your principles and process. In other words if the IT department comes to us with their new whiz-bank linear optimization scheduling package that will tell us exactly where and when to move every thing we could ask whether it:

A) is moving us further toward one piece flow

B) it is supporting people doing kaizen.

If the answer is no to either question we could reject it as anti-lean or we could ask if we can use it more effectively. We might do this by asking why it fails to support people or one-piece flow and whether we can change something to enable this effectively.

At one office furniture maker they made hundreds of thousands of end items and had a policy of build to ship complete orders of all the pieces of furniture someone wanted. The old system was MRP building individual pieces in batches and trying to mix and match to find what the customer wanted. They developed with a third party an optimization package that would look at all the orders coming in, look at their capacity, look at the parts they had in the warehouse, and build a schedule to optimize throughput and minimize the lead time for whole orders.

If sales committed to a builder that all their furniture would be at a certain site on a certain day that could be specified in the software as a constraint. They organized around work cells by product type and each cell had a computer screen. They worked hard to customize the screen design so it was very easy to see the current order being worked on and what was up next. Parts coming from the warehouse were pulled one order in advance and staged in sequence. For the parts coming from the warehouse they developed a parallel card-based system so you could visually see if the things that were supposed to be there were in fact there and whether you would be short on some parts. If there was going to be a problem making a delivery on time it would get flagged very quickly so the team could problem solve. At the front end the system told the sales person what they could promise based on plant capacity and parts availability.

The lean cells, visual kanban, and optimization system worked wonderfully and this plant had almost perfect on time delivery and very little inventory. The point was to envision the material and information flow and then develop the right IT system that supported it in the right way.

Then continue to kaizen the IT system.


One Comment

  1. […] la  lignée de notre précédent article sur le point de vue de Jeff Liker, voici le point de vue de Michael Ballé, Chercheur Associé à l'Ecole Nationale Supérieure des […]

Leave A Comment