August 1996 € Volume 6 € Number 8

What If ...

By Tom Wallace

"Hey Lee, what if we pick up two more Target regions out west, can we start shipping their new-store orders next month?"

"Hey Lee, can we ship the Ford order on the 16th rather than the 23rd?"

"Hey Lee, if that order for GE gets delayed two months waiting for engineering, what can we pull up to keep the plant busy?"

Sound familiar? If you've ever had a job like Lee's -- in planning, production control, master scheduling -- you probably know how she feels. How in the world can Lee respond effectively to those kinds of questions? Answer: with great difficulty unless ... she has the right kind of tools.

With yesterday's informational tools, Lee's answers would probably be limited to responses like, "I don't know," "Maybe," and "We'll give it a try." Notice: "No" is not in this list. And that points out a fundamental problem that experienced production control people know only too well: Without the facts, it's impossible to say "no."

Prior to MRP II, people in planning and production control jobs -- the Lee's of this world -- didn't have a chance. They didn't have the facts and thus they couldn't say no. Most of the time, they said yes, and as a result, three other customer orders fell through the crack.

With MRP II, simulation -- the "what-if" capability -- became possible (but not practical, as we'll see in a moment). The theory was great; it went like this:

1. Copy all of your MRP II-related data onto a separate set of disks, to be used for simulation.

2. Play "what-if" by making the requested changes to the simulation files.

3. Run the master schedule, capacity planning, and/or MRP systems using the simulation files.

4. Evaluate the output and respond -- with facts -- to the person who asked the question.

This was a breakthrough. For the first time ever, people could play "what-if" with meaningful results. Prior attempts at simulation came up dry, because they couldn't capture all of the complex variables present in a manufacturing company. There are too many variables, too many resources, too many customer orders. Invariably the person doing the simulation had to make a number of "simplifying assumptions." Well, one would quickly simplify one's way out of anything that provided actionable answers to the questions being posed.

So MRP II filled the void. It offered many advantages: the same data, being processed by the same program logic, displayed by the same screens and reports, being used by the same people. This was a quantum leap over the simulation programs with their simplifying assumptions, their different logic, their different displays, etc.

MRP II made simulation possible, but for many companies it wasn't practical. The reason? Time. It simply took too darned long to get answers to the what-if questions. With MPS/MRP/CRP run times ranging from a number of hours to several days, MRP was usually run on the weekends. Is it practical to wait for two or four or six days to give an answer to someone who wants to know when their order can ship, or can it be shipped earlier, or what happens if we don't get the order? Is it practical to wait for another whole week to rerun the simulation if the first pass didn't give you a satisfactory solution? In a word, no.

And for that reason, simulation has been one of the most underutilized of all of the MRP II tools. But not for long. Thanks to some very effective software, it's possible to run your MRP simulations in far less time than several hours to several days. Would you believe several minutes? Well, it's being done, by more and more companies every month.

How, you ask, is such a thing possible? Well, conceptually it's very simple: All the relevant data is loaded from disk into main memory. Then MRP/MRP II is executed at the speed of light. No more disk seeks, reads, writes, latency, etc. MRP runs were historically disk bound. The solution: Get rid of the disk because that's the constraint. I suspect Eli Goldratt would be proud.

Today, companies like FastMAN Software, Red Pepper Software, Enterprise Planning Systems and others offer these packages, often referred to as "MRP accelerators." They're quite good, and if you get one, you'll be in fine company. I spoke at a FastMAN user group meeting about a year ago, and some of the companies represented in the room were Hewlett-Packard, Allison Transmission, Intel, Lexmark, the Trane Company and Microsoft.

With a tool like this, it's possible for people like Lee and her fellow professionals in the planning and scheduling field to readily answer questions like "Can we take the new big order?" "Line four just went down for 10 days, what's going to slip?" "Do we have parts committed to stock orders that we can shift to take care of _______ (fill in the blank: Ford, GE, Wal-Mart, HP, Boeing, Office Depot, or whomever).

These kinds of packages are very user-friendly and highly graphical. Some of them make effective use of artificial intelligence, and can show the dollar impacts of competing alternatives, resource utilization, etc.

Bear with me for just a moment, while I engage in what might seem to be some blue sky. Imagine you've implemented sales and operations planning in your company. Your top management team is meeting once per month to authorize sales, production, and inventory plans that will harmonize demand and supply. One of the executives -- the president, or perhaps the CFO -- raises questions: If we can pull up the new product launch by six weeks, we're sure we can beat the competition to the market. Can we do it? If so, what other products might be affected?

You're projecting the S&OP display for this product family onto a large screen from your PC, which contains the MRP II database for the entire company. You ask for a brief time-out while you run the MRP II simulation using your MRP-accelerator software. Within no more than several minutes, you have answers: Plan A is feasible but will affect product X; Plan B is feasible but will affect products Y and Z; and Plan C will have the least total cost but will cause serious stockouts across much of the product line.

The phrase "top management war room" comes to mind. I predict that within about five years this type of capability will be widespread: sales and operations planning linked with simulation software running at the speed of light -- supporting major demand/supply decisions with facts, not guesses -- in a top management setting. This is not blue sky. It's coming. What an exciting time to be in this field!

Tom Wallace is an independent consultant based in Cincinnati. He is the author of Customer Driven Strategy: Winning Through Operational Excellence (1992) and editor/author of The Instant Access Guide to World Class Manufacturing (1994). Tom is co-director and a Distinguished Fellow of the Ohio State University's Center for Excellence in Manufacturing Management.

For more information about this article, input the number 11 in the appropriate place on the Reader Service Form

Click here to return to the table of contents.