Category Archives: Supply Chain Concepts

Inventory Drivers – Order Quantity

Inventory driverWelcome back to the ongoing series about Inventory Drivers.  Changing gears slightly I would like to move from accuracy to Order Quantity.   There are two aspects to Order Quantity – Minimum and Maximum.  There are also two different operational groups involved – Purchasing and Production.  All of these affect inventory levels both directly and indirectly plus you also need to consider whether you are looking at individual materials or aggregate levels.  The primary effect of order quantity on inventory levels is reflected in the standard formula for Average Inventory levels of individual items which is:

Average Inventory = ½ the lot size + safety stockInventory Driver Order Qty

The immediate and direct effect of this is that the larger the lot size the larger the average inventory.  Thus it would appear to make sense to reduce lot sizes as this would automatically reduce average inventory levels.  The problem, of course, is that there may be downsides to reducing the lot size.  If reducing the lot size increases your cost by more than the value saved by reducing inventory then this is a bad idea.  An easy example of this would be if you were receiving a liquid product by tank truck.  Reducing the lot size of each order will reduce the average inventory but the cost of shipping will be the same for both lot sizes.  Since the cost of transportation per unit is simple the cost divided by the number of units (let’s say Kilograms in this case) then the cost per kilogram goes up as the number of kilograms per load is reduced.  (I specified a tank trailer as you probably cannot replace the reduced material quantity with a second material which you could do if the material was in drums in a standard truck.)  The point is that the minimum and maximum levels you currently use are probably a reflection of some other constraining factors.  Even if you originally calculated the ideal quantity using the Economic Order Quantity (EOQ) formula it was then rounded to reflect some other factor or factors such as truck size, packaging quantity, takt time…  Does that mean that EOQ is a waste of time?  Absolutely not, but you must recognize that it is only a starting point.  Rounding from the EOQ is easy because when you graph the cost pattern you get a “bathtub” pattern not a bell curve.  In a bathtub pattern (or curve) you have high points at either end with a long flat lower area in the middle. (It looks like a cut-out of a bathtub silhouette, hence the name.)  Since the EOQ point will be in the middle of this flat zone, rounding the number has very little effect on the ultimate cost per unit.  What we need to do in order to reduce inventory levels is to slide the whole curve down.  In order to do that effectively you must first understand what other factors affect the size.  If it is an internal quantity it may not be that difficult to determine what other factors are driving the lot size (mixer size, takt time,,,).  By modifying the appropriate factors you then modify the EOQ which gives you a different range for rounding.  For purchased material this may be a little more difficult as you may not have access to understand, or the ability to change, the cost structure at your suppliers organization.  In addition you also have a set of logistics based cost factors that may also be affecting the final order quantity.

One important consideration about order quantity is that you may not have to order in multiples of the order quantity.  It is possible that the quantity represents a minimum and that once you pass that quantity you can move to a smaller multiplier or even to a lot size of 1 (order exactly what you need).  Again, understanding the drivers that affect the lot size will define the flexibility in the lot size

Bottom line, this is an exercise in strategy versus tactics.  Management instructions (strategy) for inventory reductions will probably be expressed as an aggregate (eg: reduce inventory by 20%) but proper inventory management (tactics) will require you to plan your reductions at the individual material level.  You need to identify those materials that can be reduced, that have a large enough quantity to achieve your aggregate goals and the drivers that affect the inventory levels (in this case, order quantities) before you create a plan to satisfy the strategy.  Failure to do this will almost certainly result in the inventory quickly growing back to its original quantity.  There are reasons for the inventory on hand and if you do not understand and control those reasons you cannot control the inventory levels properly.

Enjoy thinking about this topic.  Talk to your friends and co-workers about their experience and thoughts on this topic, especially what it means for your organization.   I would love to hear back on your (and their) thoughts.flag

Inventory Drivers – Forecast Accuracy

Inventory driverWelcome back to the ongoing series about Inventory Drivers.  Since the last two blogs talked about accuracy I thought I would continue the trend and talk about Forecast Accuracy.  I realize most people think this is an oxymoron since the guiding principle in most forecast programs is that the forecast is always wrong.  Fortunately we really don’t care that it is wrong.  What we really care about is “how wrong is it?”  This is a question that most forecasters get really passionate about because even though we know perfection is not possible, deep down we are convinced that if we can just get enough information and a fast enough program we can get it right.  We KNOW we can’t but we still want to.  Actually, from a company point of view, accuracy should be a matter of Return on Investment.  If we spend more money on increasing accuracy than we get back in cost saving from that increased accuracy then we are wasting the company’s money.  The important thing to keep in mind here is that an accurate forecast is not required for proper inventory control.  That is why we have things like safety stock.  Admittedly, the less accurate the forecast is, the more safety stock that is required but that is all just a matter of applying the right formula and having a clear definition of just how much risk the organization is willing to accept.  That does not mean that a forecast is not important.  It is critically important for a number of reasons from inventory planning, to financial planning, to production planning…  Notice a trend there?  Forecasts are important to the PLANNING process not the input / output control.  Plans, just like forecasts, are inherently unstable.  To misquote a little bit – No plan survives contact with reality.  Once you have a plan it is necessary to monitor how close reality is to the plan and adjust accordingly.  How often and quickly you do this will dramatically influence how successful you are at controlling your inventory.  Both planning and forecasting are about RISK CONTROL and it is the need for lower risk that leads to the higher inventory levels.  Increased forecast accuracy (to a point) leads to better plans, higher customer service, lower risk and eventually to lower inventory levels.forecast accuracy

One other quick thought before we leave this – WHO in your organization is responsible for generating the forecast.  I would suggest this is the responsibility of the sales department.  They do not need to do all the calculations and details of the forecast process but they are the people in your organization that are dealing with your customers decision makers and therefore have the best chance of finding about about potential changes before they happen.  A forecast built solely on historical data will never catch potential game changing issues until after the fact which always leaves you reacting to everyone else.  It is like driving using only the rear view mirror.  You are OK so long as the road is straight but as soon as there is a curve in the road you end up in the ditch.  Those curves represent changes from historical norms such as changes to customer orders.  Your sales force is in the best position to identify the “curves” before they put your processes “in the ditch”.  The problem of course, is that most sales people are not “fond” of forecasting as that is not their primary job (which is to sell product).  To counteract that I have always given them three instructions about forecasting:

–   Sort the customer or product by forecast accuracy and then start with the worst one and move down the list.  The more accurate ones are already good and working on them will not help as much as working on the worst one

–   Do not spend more than three hours per month doing the forecast.  They may not get through every forecast but hopefully the ones they do will be further down the list next month so rotation will eventually bring each customer to the set that get touched. This is sort of like continuous improvement.  Over the long run you can get amazing results.

–   If you must guess, then guess high.  It is generally less expensive to have a little extra stock than to run out and short a customer.  While saving money is always a good point the sales force will be more in favour with the point of not shorting a customer.

Enjoy thinking and talking to your friends and co-workers about this, especially what it means for your organization.   I would love to hear back on your (and their) thoughts on this topic

flag.