Learning Center

Traditional MRP was developed in the 1960’s and put into computer code in the 1970’s. It essentially hasn’t changed. What has changed? The conditions in which we work and the environment in which we live. Customer expectations have certainly changed. And the complexity and length of our supply chains have certainly changed. The business expectations (lower inventory investment) have also changed.

Think of it this way. If you live 5 miles from your workplace, you might ride to work on a bicycle. If you move and now live 50 miles from your workplace, you need to revise your thinking. Either you need to leave much earlier or you need a different mode of transportation.

Yes, the equations are still valid. And they still require the same degree of precision in the calculations.

For example: When our customer lead times were less than our supplier lead times, this was manageable. But now, I have a customer that wants finished product next week and my lead time for the components is six months? I can no longer wait for the customer order before I purchase my components. I need to predict when I will need the components and be prepared by having them on the way or in stock.

The problem comes in two parts.

1. It’s impossible to accurately predict my usage that far into the future.
2. When I correct those wrong predictions, MRP must recalculate the current quantities
   and timings due to the precision of MRP calculations. This changes the requirements of
   every other part in the bill of materials.

Forcing precision into an imprecise situation is counterproductive and the harder we try, the more counterproductive it becomes.

Precision itself is not the enemy. It’s the length of and the steps involved in the precision that causes trouble. Why do most of us prefer one direct flight vs two stops when making travel arrangements? Because every stop is a potential for things to go wrong. One domestic flight
with no checked luggage vs. three segments with an international portion. It’s easy to say which one is more predictable in terms of timing. So, precision works, but the longer the timeframe, the more likely precision will fail. That timeframe is determined by each business (there is no
“one-size-fits-all”). In between the precision, we need something to absorb the variability so that precision can work.

The intent of safety stock is to protect against variations in demand and supply (paraphrased from APICS Dictionary). Sounds great. Useful even. But the application of this principle breaks down quickly. Safety stock is never intended to be used. If safety stock is breached, immediately the demand goes to past due, making it an emergency (which is what we’re trying to prevent by having safety stock). Safety stock is just a new zero, in terms of the application.

Safety stock is not a planning tool. It’s like a last chance, “wake-up call” for inventory management. It’s like an alarm you set for 6am and then hit snooze every three minutes until 6:30am.

An inventory buffer is a planning tool. We make our decisions using the inventory buffer always, not just when we’ve dropped below a “dangerous” level. Every day we review our situation (net flow position), and we compare that result with the planning buffer. If it’s above a certain amount, we are fine. If it’s below that amount, we place an order. If we’re far below that amount, we expedite.

With the inventory buffer, we develop a warning system. A percentage-based ranking that tells us our current risk.

With safety stock, we are either OK or NOT OK. Because safety stock is our last line of defense, not a tool for decision making.

Great question. Your home kitchen and meal planning/preparation/consumption provides an excellent example of inventory management, production planning, production, delivery and customer service. You have the ingredients (bill of materials), recipe (work instructions), Directions (Production routing), Grocery shopping list (Purchasing), Dinnertime (customer request date), Dining location (logistics and delivery), Stove, oven, mixer (equipment and maintenance), and several other specific similarities. Inventory management is similar too. You keep some ingredients all the time because you use them in many different meals or dishes. And some items you only buy when you need them, because they are very seldom used. And for items you use often, like eggs, you subconsciously set a “safety stock” of six eggs (for example). When you get to 6 eggs, you make a note that you need eggs and review whether or not the 6 eggs you have will last until the next time you drop by the store. This works great when you have one refrigerator and you are on-site. But if you have 100 refrigerators and you only have visual access to three of them, it’s a different story 

First, we find the strategic locations to place inventory buffers. The purpose is to absorb variability, such that the precision that is between the buffers is allowed to work.

Second, we determine how much inventory should be placed in these strategic locations. And we focus on maintaining these inventory levels.

Third, our initial work in creating the buffers uses data that can change. We need those changes to automatically implement and adjust the buffers dynamically. We cannot wait a week or a month to adapt our decision-making tools.

Fourth, we need to follow the output from the system. We’ve put our work in up front. Now, the resulting suggestions and plans are appropriate. We still look for outliers, but we don’t have to question every suggestion.

Fifth, we create a simple system of showing the relative urgency or priority of the parts or the schedules involved. This means that anyone can clearly see what should be done first or next.

1. There is a type of lead time of which traditional MRP isn’t aware. The lead time from buffer to buffer. We call this the Decoupled Lead Time.
2. The Net Flow Equation defines the Net Flow Position which is one part of our decision-making process.
3. The Decoupled Explosion (of the bill of materials) significantly calms the chaos from the impact of changes to order quantity and timing. Traditional MRP explosions mean every time a customer changes an order, absolutely every component related to that order also changes. The Decoupled Explosion only changes the items until the next buffer point.
4. The Relative Priority. We now have a way to compare the priority of components to each other and to their own buffer. Even though the quantities may vary, a component that has 20% of its buffer remaining is high priority and a component that has 30% of its buffer remaining is still high priority but less priority than the first component. While a component that has 95% of its buffer remaining is just starting to raise nervousness.

The first step to implementing Demand Driven MRP is to learn the methodology. There is not a secret component to DDMRP. We believe that you should learn the methodology first, then run a pilot program to test the potential, and then, purchase software as needed.

You can implement for a segment of your business, or you can implement an entire organization. If you are struggling with having too much of the wrong, too little of the right and overall, too much inventory, then DDMRP can help you.

The good news is that results start to appear in a relatively short time. It varies with your business, but generally improvements are shown in three to nine months.

Take the Demand Driven Planner class as a first step. Develop a spreadsheet to fine tune the methodology for your business and your application. The class is a 16 hour workshop, which Become Demand Driven offers over 2 full days in person or 4 half days virtually.