MOST OF THE inventory-control systems in place are predicated on old theory and facts — many of which no longer apply or are arithmetically wrong today, according to Ron Slee, president of RJ Slee & Associates.
Slee says that the lead time — how long it takes to get a product from suppliers after an order is placed — was between four and eight weeks in the 1960s and is now between a day and a week.
“It should have changed dramatically what everybody looks at in the way of inventory control,” he says. “If you turn your inventory six times, how many days of sales does that mean you have on hand? 60. If your stock-order turnaround is two weeks — 15 days — what are you doing with the other 45 days of inventory?”
He says one of the major functions of inventory control is stock ordering, which involves when to order and how much to order.
He says when to order is the order point, or minimum, and how much to order is the order quantity, sometimes called an EOQ.
When-to-order considerations consist of: lead time (how long it takes to get the stock order back to you); the demand during that lead time; safety stock (the 45 days); and service-level considerations (whether you want 50% service to your customers or 95% service to your customers).
He breaks down lead times into these groups:
-
Forecasting systems
Order process lag time. “It's when you know you need to order until when you do order. If you place a stock order every day, the lag time is a day.”
-
Order review time. “How many of you print out a suggested order and somebody looks at it before you send it to the factory? Why do you have the computer system at all? You put into your computer system the rules for inventory management and rules to calculate order points. True? But we're smarter. We know more than the computer, so we don't review the things. Does that make sense?”
-
Placing the order.
-
Supplier processing time and shipping and transportation time. “These two are the only two pieces over which you do not have complete control. From the time your supplier gets the order at their place of business until the time on the shipping dock that it's ready to go, your supplier owns that time. From the time the trucking company picks it up until it's delivered to you, you don't have control.”
-
Physical receiving and placing time. “My rule is, all stock orders received today will be put away today and records updated today before we go home. The four rules are: find every part every customer is looking for every day on same day as they ordered it; ship every order I receive the same day it's received; try to ship today every emergency order I receive today; and put away today every stock order I received today. And in all cases, before I go home. More of you can do that than you think.”
Slee says lead-time demand requires forecasting systems.
“From the early days of batch processing back in 1950s, we've had the need to calculate what the demand would be while we're fussing around with all paperwork,” he says. “In the old days, you'd have a piece of paper go to the data processing department into a key-punch room staffed by a bunch of ladies who would pound away at their data-entry equipment, create a card, run the card back through computer, print a report, take the card and report and input document, send them back to you, and then you verified it. It'd take forever.
“We had to start creating forecasting systems. We started first with Kardex. It was labor intensive. We had to see all of the activity on the card before going to pick parts. From this we could see if we were going to run out of stock soon. We would tag the part at the time of sale for review at order time.
“Computers come along and replaced that in the ‘60s, and we had to have a forecasting model. Normal distribution is the bell-shaped curve, the most basic form of statistics on the planet. It's saying the bell, the curve, is going to align down the middle and the parts on the top and bottom are going to be equal in number. So you're going to have the same number of parts that are fast movers as you do that are slow movers. Does that fit your business? No. So why the heck would we use it? Because it was all we had.”
Along came Robert Goodall Brown with Inventory Management Performance & Control Techniques (IMPACT) — the software package that IBM pushed in the ‘60s that was a very sophisticated forecasting statistical package.
“It still had the bell curve, but it acknowledged the differences we'd have off that medium,” Slee says.
In the ‘70s, it moved to the Poisson Distribution System, which is normally used for traffic lights.
“This statistical package, a better fit to our industry, represents a demand pattern that is more slow-moving parts than fast-moving parts,” he says. “Instead of being the bell, where you have the medium in the middle, this one had everything skewed on the left-hand side — saying that a large proportion of your business is slow-moving parts and a small proportion is fast-moving parts. Does that fit your business?
“I should never have a back order on a fast-moving part. That's the part my customer expects me to have. Yet every single one of you, whether you review stock orders or not, has a back order of fast-moving parts every day of the week. This doesn't make sense.
“What do you have to do to make sure you have the part that sells 25 times a year or higher? Stock it. Enough to make sure you keep your job. Is a 60-day supply enough? It's too much. But I'm safe at 60 days. Now, if you had a back order on a fast-moving part — and I want you to have a 60-day supply — what does that mean is going to happen to your inventory of fast-moving parts? It's going to go up. If you had a back order, that means you didn't have a 60-day supply. If I want you to have a 60 day supply such that you don't have to back order, your inventory's going to have to go up.
“Between 35% and 50% of the part numbers you carry have not met your stocking criteria. Why do you have the parts there? The manufacturers make recommendations to us as to what we should stock based on products in our marketplace. It's an initial stock list that's a recommended spare-parts lists. That's cool. But you ultimately make the decision.
“Thirty-five to 50% of the parts you have in inventory in dollars haven't met your stocking criteria; 35% to 50% of the part numbers you have haven't met your stocking criteria. That means 35% to 50% of the locations you have in your warehouse are unnecessary. That means every time you pick an order, you're going one-third to a half more distance than you need to. So you must be making a lot of money. You're spending a lot of labor, a lot of money on heating, lights, insurance, taxes for the facility, a lot on racking, shelving, and cabinets — whatever you need for storage. What should we do about it? If you have a part that hasn't met your stocking criteria and it's not a requirement of your manufacturer and it's under $10, why not just throw it away?”
Slee says the Minimum/Maximum Approach determines that a minimum quantity of sales be kept as the order point. For example, 90 days of sales supply will be the order point.
“It doesn't need to be any fancier than that,” he says. “How do I determine how many days of sales I want to have on my shelf? First you have to find out what your lead time is. If it's two weeks or less, have an order point that's a month.
“In the financial world, there is a term called Return On Capital Employed (ROCE). I want you to think about that in terms of gross profit. If your gross profit is 40% and your turnover is six, multiply that and I have a 240% return on capital at the gross-profit level. If I take the turnover to 12 from six, I can take the gross profit from 40% to 20%, and end up with the same return on capital. Now I don't want to have all my parts go down in price, but I can take some of them down dramatically in price — the ones I get my butt whipped on all the time because I'm too high in price. It's all because I'm managing my inventory properly. You think you'd get more business if you took your fastest-moving parts down by 20%? Yep.”
Slee says order formula codes allow parts departments to be more precise. The approach used here is to “segment” the inventory along demand and price lines, and set up ordering rules within each segment.
In addressing how much to order, he says the Kerr-Norton formula balances the cost of placing orders with the cost of carrying parts.
“Using the computer costs virtually nothing per part,” he says. “So why are we doing what we've always done before? So how much should I order if my order point is going to be based on time? When I hit an order point, how much should I order? I want to order exactly what I sold — no more, no less. You can set your order point at 30 days and order back every day or every week what you sold in the previous day or previous week, and the performance you get out of your inventory will be better than what you are getting today.
Inventory turnover
“Wal-Mart, what is their inventory turnover? The inventory it has in its distribution centers and stores is owned by the supplier. What's their turnover? Infinite. They have to pay when they sell the part or shirt or whatever. They scan everything, and that scan transaction goes to the manufacturer as it happens. The manufacturer turns around and sends an invoice. Wal-Mart has 30 days to pay. They have your money for 30 days before they have to pay. Is there any wonder they have low prices? It's not rocket science.
“We can do the same thing in our business. It's just that we're not dealing with shirts. We're dealing with steel and other things. Why in the world am I satisfied with four or five or six turns? Why am I not aware how many fast-moving parts I didn't have last week? I want to know how many days you go in a row where you find every single part that every single customer is looking for every day? Is that a worthwhile goal?”
He says the traditional definition of turnover is the last 12 months sales at cost divided by the average inventory value at cost.
“So if I sold $12 million at cost last year and had an average inventory value of $1 million every month, my turnover is $12 million divided by $1 million — 12 times,” he says. “The problem with that formula is, the last 12 months' sales include sales that did not come out of your inventory. So the turnover you're measuring is overstated. That's where true turnover comes in: I want to have the last 12 months sales out of the stock, divided by my average inventory.”
Slee says stock-order efficiency is an inventory-control measure gauging the percentage of dollars placed on order with suppliers that are on stock orders.
“So if 70% of the orders you placed to suppliers are on stock orders and 30% are emergency orders, 70% of your turnover is true turnover,” he says. “Then I have a much more real measure of my investment.”
He believes the biggest source of error on physical records is the count itself.
“If that's true — and I believe it — when should I count to minimize the probability of error? When it's at zero,” he says. “The second-most important time to count is when you receive.
“I want you to pull a page from IBM's customer-service revenue that says if there's a problem, you fill the sky with planes, which means you get everybody and their dog involved in the problem. So if I have a part today that I can't find and I call you and tell you I can't find it, tomorrow my boss will call you and say, ‘Hey, I know Ron couldn't find that part for you yesterday. I just wanted you to be aware we're all looking at it.’”