The TOCICO conference this year started with deep-dive workshops on Sunday, much like the old knowledge "upgrades" that used to happen. I sat in on the Demand Driven: A Practical Workshop on The Strategy of Flow by Debra Smith - she's one of the long time experts on this topic from her experience creating the concepts of DDMRP and implementing with many clients.
Demand Driven MRP is one of those topics I have wanted to learn more about, as it is a variant of the classic Drum-Buffer-Rope that Alex Rogo implemented in his plant in The Goal. Clearly, if I want to learn more, I need to read more (and there are plenty of materials out there, starting with the DDMRP Institute and the books). The workshop wasn't intended to be a full how-to-implement DDMRP, instead an overview of the flow and lots of relevant tangent conversations about key elements that Debra Smith sets as critical to such an implementation. I am certainly not an expert now that I've sat through this workshop (and a few others in the past.)
Smith makes it pretty clear that she does not "sell TOC" when doing DDMRP implementations with clients. Sure, she uses many of the familiar TOC concepts, but her focus is on supporting the client's needs. From her perspective that doesn't mean teaching them the thinking processes or running generic simulators (she develops a simulation of the client's operations) or even educating them about the damages of large batch sizes and resource utilization. The "TOC mindset" is meaningless if there isn't a way to act upon it. She does use the thinking processes in her preparations and debriefs with the client, but rather than pushing them through the process, she creates them herself and uses the results for communication. The right thinking and behaviors are a byproduct of having good signals and knowing how to act on them.
Most people know that MRP is broken. Her focus is on developing a solution for the client that gives them the right signals, so that they can react appropriately. (Instead of using the signals they get today, that drive them to act against the system.) A lot of the conversation was around the idea of relevant information - information relevant to the situation at hand. (Goldratt coined the phrase that, "Information is the answer to the question asked.") She talked about varying time horizons where different information is relevant to operational, tactical and strategic decisions. Forecasts, for example, are useless for operational (day-to-day) decisions. Interestingly, she suggests that access to this information can become the constraint to growth for an organization - they either can't get what they need, or they have the wrong information that drives the wrong decisions and behaviors.
DDMRP uses many of the concepts from the concepts I've understood in my other work in TOC. Buffers are placed in strategic locations to decouple the system and bring signals of variability closer to the place where it occurs. Buffers are used as the key signalling mechanism throughout. And buffer statistics are used for ongoing improvement efforts: those things that cause frequent buffer consumption are where there are opportunities to improve the system, drive down variability and improve flow. One of the differences that Smith called out in connection with the DDMRP buffers is that they are calculated, based on the information for the point being buffers - each buffer zone "means something" mathematically. (In standard DBR, the buffers are 1/3, 1/3, 1/3 based on a rule of thumb.) The other difference appears to be in the number of buffers - Smith talks about placing control points (and thus buffers) anywhere you need to understand the signals.
I think I have a better feel for the DDMRP concept. If I am to go much further, it will require investing some time into more study and experience.
And some other comments and callouts from the session:
Smith wrote The Measurement Nightmare in her early thinking about these topics. While she says some of the details are out of date, the overall concept is still there. It's another book that's been on my long list of "TOC books." The core lesson, "Metrics block improvement."
What else blocks improvement? Education (I "know"), Bad measures, Software.
What is a "proof" is it logical? It is defined as "Evidence or argument that compels the mind to accept an assertion as true." But does this mean that it IS true? Or that I am finally compelled to accept the assertion?
F. Donaldson Brown is the father of management accounting and is one of those people we don't hear about much. But he was one of the keys behind General Motors' turnaround (strongly linked to Sloan).
The first law of manufacturing, as coined by George Plossly in Orlicky’s Material Requirements Planning, "All benefits (ROI) will be directly related to speed of FLOW of materials and information." Flow comes first.
As visibility improves, variability decreases. (And as visibility declines, variability increases.) Visibility is defined as relevant information for decision making.
There are a wave of DDMRP implementations in France. It's almost become a national phenomenon there.
- There are ~35 software vendors out there that have been DDMRP certified. Interestingly, none of them are software that we traditionally hear about at TOCICO.