How Demand Driven MRP enables Lean Manufacturing to become practical in complex environments

Demand Driven MRP eliminates the component, parts and material shortage problems that create so much conflict between the “Pull” philosophy of Lean manufacturing, and the “Push” necessity of coping with long lead times and multi-level Bills of Material

(Or: how a piece of 50-year-old logic embedded in their modern multi-million dollar ERP systems is preventing major corporations from gaining a Return on Investment from the additional millions they are spending to implement Lean Manufacturing)

Background

In 1981, the first presentation in North America of the details of a technology labeled “Zero Inventory” introduced a way of working that was in play in Japan, with a level of success that was evident in the automobile marketplace, but which was utterly alien to Western manufacturers in almost every way.

It focused on a philosophy of operating with very low inventories, and introduced a mechanism designed to help work flow quickly and smoothly through the resources of a plant; a simple, elegant mechanism called kanban.

Soon afterward, the philosophy became re-branded as “Just in Time.” And it expanded beyond just the kanban mechanisms, because it was clear that a lot more than just the kanban technique was necessary to achieve the type of results that the Japanese originators were achieving.

Set up reduction (usually via single-minute exchange of die, or SMED) was introduced. Poka-Yoke (mistake-proofing) was discussed, even if little was actually done in terms of application. Quality improvement became a major topic, first with Deming’s work in Japan gaining some publicity and then with the emergence of Total Quality Management, or TQM. Concepts of reducing variability began to get an audience. Statistical Process Control no longer caused managers to make the sign of the cross and run in the direction of “away.”

But the Lean revolution really began with the publication of “The Machine that Changed the World” by Womack, Jones and Roos in 1990, focusing on the automobile industry and highlighting the differences between Toyota and others.

“Lean Thinking” in 1996 went far beyond this when it took the concepts in use in Toyota and presented them as formal principles, applicable to any organization, with a supporting cast of tools and techniques  including the Just In Time elements of kanban, continuous improvement, kaizen, poka-yoke, set-up reduction, load leveling, continuous flow, and much more. A whole new language came into play, with the focus on the elimination of “muda,” or waste; with value stream mapping, the application of takt times, the implementation of the 5S techniques and more.

Today, manufacturers are spending millions of dollars in their efforts to implement Lean Manufacturing, and while many companies are simply paying lip-service, others are deeply committed – and the multiple year commitment shows in terms of budget, personnel dedicated to the Lean implementation, extensive education and training for hundreds, even thousands of workers, consultants being flown in from Japan to lead “kaizen events” (especially in the early days), and of course … meaningful changes in policies, procedures, measurements, systems, processes and behavior at many levels of the organization.

“Pull” versus “Push” Production

Soon after Zero Inventory made its debut, in order to characterize the differences between traditional manufacturing practices and Zero Inventory, the phrases “Pull system” and “Push system” were also introduced.

The Just In Time approach was to make only what was needed, only when it was needed; and to implement systems that explicitly made it impossible to produce parts and products that were not needed.  The market therefore “pulled” work into and through the resources of the plant. It was a “Pull system.”

In contrast, the Western way was to focus on machine and people efficiencies, and to maximize efficiencies it made sense to push work into the plant to keep people and machines busy … regardless of whether the parts or products were actually needed yet, regardless of whether the quantity was more than was needed.  This is the “Push system” philosophy, called by some “Push and promote” because the end result is often a sales force being compelled to sell what was produced, rather than having a factory producing what was selling.

When Lean theory is stymied by frustrating shortages

Unfortunately, all the investment, all the education and training in all the techniques, all the commitment and dedication can be rendered mostly useless if the plant’s circumstances are such that it will inevitably encounter chronic shortages of materials, purchased parts, fabricated and manufactured components, and finished goods.

Even a single shortage of a single $1 component can be enough to shut the flow of work down.  And when the practices are typically to operate with as Lean an inventory as possible … stock-outs and shortages that are INEVITABLE due to a combination of predictable circumstances are made even worse by operating to the low-inventory rules.

What’s surprising is that no-one seems to be addressing this systematically and successfully; companies  are left to try to resolve the problem on their own, almost as if they are pioneers in facing the issue.

When stock-outs and shortages might be (almost) inevitable

In some production environments, Lean pull is easy to establish. Shortages are rare.

But in some challenging and complex environments, an unfortunate combination of technologies and realities  make it almost inevitable that there will be availability problems.

The combination arises when companies using MRP software (MRP, or Material Requirements Planning, is a module in ERP systems and implemented in approximately 70% of all ERP implementations) are in environments where:

  • There are highly repetitive builds (product or process)
  • There are deep and complex Bills of Material …  and necessarily so
  • Longer and more complicated routings create lead time problems and almost guarantee that schedules are quickly invalid
  • The same materials or parts or components have multiple where-used situations
  • The market will reward you for offering a reduced lead time

The reason the problems are inevitable in these environments

Inside even the most modern ERP package, the actual logic of the MRP (Material Requirements Planning) module was developed in the 1950’s, programmed in the 1960’s, extensively commercialized in the 70’s and onwards … and it hasn’t changed to any meaningful degree since it as codified more than 50 years ago.

This may not sound too bad; the fundamental approach of MRP, exploding independent demand for finished products through Bills of Materials, netting these demands against on-hand stock, planned and in-progress work and purchase orders, is absolutely proven and genuinely solid in terms of the math and cold logic.

This logic has never been needed more than it is today, where things change so quickly that the ability to recalculate dependent demand requirements  is essential.

However MRP is also more dangerous than ever before.  The problem is that in many environments, especially the ones we’ve highlighted above, the assumptions under which that logic converts the arithmetic into an action plan are no longer valid.

For example, in the 70’s and early 80’s, when the MRP Crusade was to manufacturing what the Lean revolution is today, the issue of a system being too “nervous” was a common conversation among MRP experts – and companies of any size typically had dozens of MRP experts in-house.

A “nervous” MRP system is one where any kind of twitch in independent demand causes tsunami waves throughout the whole system. For example, this twitch could be a change in forecast; or in received customer orders; or an inventory adjustment; or a change in the Bill of Material; or  …etc.

If you Google “Nervous MRP” you’ll find many references to the topic – and they’re all from the 70’s and 80’s. Today the causes of the nervousness are far more prevalent, but no-one even recognizes this problem as even having a label. All people see are the shortages and there is insufficient MRP expertise in house to even understand why.

(Instead of scores of MRP-trained people in-house, with some genuine experts, many manufacturers today don’t have a single true MRP expert on staff. And most ERP vendors don’t, either. An ex-VP of a MASSIVE ERP company, a name you would recognize, suggests there wasn’t even one person in that corporation during their time as a VP that could hold a meaningful MRP-based conversation.)

To further accelerate and amplify system nervousness – today we are faced with shorter product lives, more complex products, new technologies developing all the time, new competitors emerging from all over the world, longer lead times being imposed as a result of off-shoring and outsourcing parts production and raw materials suppliers, higher variability in rates of supply, increased volatility in rates of demand, shortened customer tolerance time, and more.

These factors are of a scale that was not even imagined when MRP logic was developed, and when MRP “nervousness” was being debated (without any feasible solution, incidentally).

So we have a strange paradox; we’ve never needed MRP more; and it has never been more dangerous.

And as a result, when we do combine the MRP approach with our current challenging environment, the inevitable outcome is chronic shortages of materials, purchased parts, and manufactured components resulting in unhappy customers and increased costs.

How the combination creates problems for Lean

In some environments the sheer volume of planner actions that follow any significant change can overwhelm the planning department. Some changes don’t get made; and there are shortages. Others can’t be made … there is simply no meaningful rsponse to the new situation. And so parts that should be made … aren’t. There are shortages.

In other situations, the sheer volume of changes might not be unmanageable from a planner’s perspective, but the shop floor simply cannot execute effectively in the face of constant, numerous reschedules and the corresponding priority changes.  It isn’t hard for reschedules to move towards shop floor chaos; and the end result is, again, shortages.

The solution for Lean?

It’s the same solution that works for a stand-alone ERP/MRP system, and for a Theory of Constraints implementation. The key is to buffer the system at strategic points so that the “cascade” of (re)planning stops dead when it hits the buffered part; the system is effectively decoupled when we can achieve this. Th effect is to “dampen” the nervousness of the system, to bring stability, reduce the flood of reschedule messages, and buy Planners more time. Along with a tactic for managing non-stocked, critical lead time parts and materials, the result is a substantial improvement in availability and a massive reduction in stock-outs.

There are many other benefits of the Buffering; lead times can be compressed, overall inventory can be reduced, priorities become clear and highly visible and manageable, inventory actually adjusts dynamically to “right size” itself as demands increase or decrease, and as trends emerge and patterns such as seasonality develop. Execution tools become relevant.

The components of the solution are the 5 elements of Demand Driven MRP (DDMRP):

1. Strategic Inventory Positioning.

Choose the points in the Bill of Material and Routing (or, in the whole supply chain) where Buffering provides the outcomes we’re looking for, taking into consideration;

  • Customer tolerance time
  • Variable rate of demand
  • Variable rate of supply
  • Inventory flexibility and the product structure
  • More!

2. Buffer Profiles and Buffer Level determination

Set parameters, taking into account such issues as:

  • Seasonality
  • Order cycle
  • Supply variability
  • Demand variability
  • Fixed and Cumulative Lead time
  • Average daily usage
  • Minimum and maximum Order quantities, and any order multiple

3. Dynamic Buffers

Dynamic Buffer management, including

  • Adjust traits to reflect changes in reality
  • Monitor Buffer Status & recommended Buffer adjustments

4. Demand Driven Planning

This includes, for example:

  • The Buffer size composition set in Step 2., with buffer sizes dynamically re-sized based on real demand and variability.
  • Replenishing buffers as parts are consumed and the inventory level moves into the rebuild zones.
  • Demand being generated and passed down through the Bills of materials, level by level … but with DDMRP buffers decoupling the cascade of re-planning.
  • Component part orders being highlighted when they are out of sync.
  • Lead time calculations that recognize the standard parent lead time but also the cumulative lead time for parts that are not buffered but which form the longest leg of the Bill of material.
  • Buffers that reveal the degree of penetration of orders … this penetration or buffer encroachment becomes a priority management technique that is far superior to due-date priority management.
  • An Order Spike horizon needs to look out into the future for any maverick order spikes … large, anomalous sales orders, such as those often associated with promotions, or trade shows.  The system can plan for these.
  • Every order … including make-to-stock orders … has a formal due date, but Buffer penetration percentages add precision for priority management.

5. Highly visible and collaborative execution.

With highly visible priorities via the color-coded Buffers, and Buffer penetration percentages for more refined priority management, the foundation is laid for the Execution of the plan.

  • Alerts warn of zone threats based on on-hand inventory and projected on-hand
  • Alerts warn of lack of synchronization between demands and supplies
  • Back-office integration with email and other routine tools simplify communication

Launching the appropriate purchase orders and work orders is only the start; these orders now have to be maintained in sync with the inevitable changes that will take place inside the execution horizon.

With DDMRP, planners no longer have to respond to every message that is off by even a single day; the system provides intelligence about those parts that are truly at risk of inventory problems.

Be Sociable, Share!