Dixie Iron Works

After Installing Finite Capacity Scheduling Software, Dixie's Scheduler Asked For More To Do

by Gerard Danos
General Manager and Vice President of Operations
DIXIE IRON WORKS

Down the road from Corpus Christi, in Alice, Texas, Dixie Iron Works is an $8 million machining shop that serves the oil field industry. This is the story of how it grew its profits by 300% with guidance from Galt Management Services, a consulting firm, and by using Resonance, a finite scheduling software from Thru-Put Technologies.

The oil field is a 24 hours per day times seven days per week industry driven by the high day-rate costs of rigs. High rates mean vendors such as Dixie Iron Works can command premium pricing for quickly manufacturing and fixing parts that have broken. However, this is a Catch-22. Premium pricing on the part of a vendor goes hand-in-hand with a demand by the customer for premium performance. The rig operators need quick turn-around and delivery at the time it is promised. If we can't do it, our customers will find those that can.

At Dixie, we specialize in gas compressor and engine components. We're a CNC and re-manufacturing job shop that both creates new specialty parts and makes replacements for broken parts. We also manufacture our own line of high pressure — 15,000 psi — oil field plug valves that compete with the Big Names. We also provide CAD and CAM engineering capabilities to our customers.

The Old Way

For some time, we've tracked jobs and people to learn how long it takes to manufacture a certain part so we could price it accordingly and establish some type of schedule. We were using the RealTrak shop control system and still employ it for order entry.

Our inventory manager would supply us with replenishment orders biweekly. We would then pull the data in from our system and combine it with his needs to create a week's schedule. The problems we had are almost self-evident.

We were working for 24/7 customers needing extremely quick turn-around with a weekly schedule. In 1993, we were barely breaking even. With only a 3 percent operating profit, our president, Joe Merritt, started looking for answers. After reading "The Goal," he became a disciple of the Theory of Constraints and called in Galt Management Services of Dallas to help us.

Problem Analysis

We first undertook a problem analysis to learn what we were doing wrong. Our problems were not unique; they were those faced by most job shops. The first identified was one of which we were painfully aware. Our due date performance was poor.

Indeed, only one out of 20 jobs was being completed on time, even though our people were working hard and conscientiously. Such performance kept us in continued hot water with our customers and our only saving grace was that many of our competitors weren't doing any better. A few were and, of course, our customers were finding them. So, this was an area in which we absolutely had to improve.

Another one of our problems, it was quickly pointed out, was that we took on any and all work. Being good ol' Texas boys, if Farmer Brown's plow broke, we'd fix it. After all, it was a neighborly thing to do and those couple bucks we'd make would help pay the phone bill. Why not help Farmer Brown and pick-up some spare change?

Well, the reason is simple. Critical employees and equipment might be tied up on a nickel-and-dime plow part while a well-into-5-figure oil field engine component is waiting down the line. As a happy Farmer Brown hooks up his repaired plow, a very unhappy rig operator is screaming for his engine part. Yet, according to cost accounting, that Farmer Brown deal was good, extra money. We had the time to do it ... the oil field part did get delivered and paid for ... what's the problem? The problem is that Farmer Brown will definitely come back to us but that rig operator won't! The cost accountants don't have a column for lost business.

Lastly, we had very weak pricing policies based on traditional cost accounting. The above scenario is just one example. Classic cost accounting taught us that if ABC Company needed five Part Xs, we ought to make ten, putting five into inventory awaiting the next time ABC needed a Part X. After all, amortizing set-up and adding in run time, doubling the number of parts brings the cost per part lower.

Of course, it also grew our cost of inventory — we only turned inventory two times per year! — and meant, again, that our customer's well-into-5-figure oil field engine component was waiting in line. So, as those five Part Xs, with their reduced cost of manufacturing, got happily added to inventory, an unhappy rig operator was screaming for his engine part. He didn't care that the next time he needed a Part X, we'd have it in inventory. All he cared about was that his rig was down and idle ... now.

Management of Constraints

It was emphasized to us that all resources are finite. For instance, you can have five pounds of peanut butter, four pounds of jelly, several loaves of bread but only one little plastic knife. All these resources are finite. But, it doesn't take a genius to figure out that the most important finite resource we have for making sandwiches quickly is that one little knife. It's critical.

To exploit and control their critical resources and generate high returns on per unit of capacity, successful shops take this reality into account throughout their planning and scheduling process. It is too late to recover after jobs are released to the shop floor based on wrong assumptions.

Starting in the summer of 1993, working hand-in-hand with Galt Management, we began to attack our problems. We held a series of in-house seminars and asked all employees to read "The Goal." We began to select our critical control points, those areas in our operation that constrained our output. Then we figured out how to exploit them, i.e., make the best use of those resources.

Not only did the theory work, it bore fruit. Our due date hits were up to 30 percent. Inventory turns increased to almost six and profit margins grew 67 percent to five percent.

Computerizing Scheduling

Once we hit a plateau with the manual process, we decided to search for a software system we could use to schedule our shop. Such a system needed to be based on constraint theory concepts, run on Windows, and offer extreme flexibility.

In early 1995, we met with Thru-Put Technologies to learn about their Resonance™ finite capacity scheduling software. We met with other manufacturing system vendors as well. We immediately reaffirmed why we wanted to eliminate simulation systems. Used as a means to predict the impact of alternative schedules, using simulation for actual scheduling is expensive. If you subscribe to the realities of manufacturing — finite capacity dependencies amongst various resources, statistical fluctuations and process variability — it is almost impossible to use simulation as a guide.

Scheduling, on the other hand, optimizes various production factors and makes tactical decisions regarding overtime, batch sizes, material allocation and other variables. We had to have our scheduling system comprehend these realities with a simple model. That's also why we excused complicated, data intensive models like MRP and ERP systems.

Finally, we selected Resonance and began working with Thru-Put in April. Upon installation, the first thing we learned was that much of our data was dirty. When using it manually, we could rationalize our way around it. Once computerized, we were told one plus one can only equal two.

For instance, line workers were supposed to punch in the quantity or piece count of the job being worked into our electronic shop control system. However, for some time, we'd been lax on this procedure. Since this information drove work-in-process, our WIP picture was quite cloudy.

However, because Resonance is based on Theory of Constraints, it operates on an 80:20 rule. Using the drill down features in software, we quickly identified the data problems and got it right in a short period of time.

We now had the capability to quickly and accurately access where and when bottlenecks would occur and began using the software's drill down features to highlight what orders would be affected.

No longer did our scheduler have to intervene, solving conflicts job by job. The Resonance software, instead, recalculates a rescheduling solution for our bottlenecks. It suggests optimal overtime, allows alternate routings, minimizes change-overs and shows the impact on throughput and delivery dates with a couple clicks of a mouse. We can schedule our shop within 15 to 30 minutes.

Here's a scenario, some version of which happens all the time. Our largest customer is directly across the street from us and often gets very large international orders that require fast turnarounds. Recently, he brought over a large P.O. for us that had a due date that was really tight.

Based on the machining we had, the software advised us we could handle the job but only to the detriment of all the other customer orders now in house. Based on all product being due on this specific date, we had to refuse the order.

However, we asked him how many he truly needed to satisfy his immediate needs. Using the software, we could schedule in his immediate needs without harming any other customer order and produce the rest of his order still in time to meet his longer-term needs. This is a very common example of how finite capacity scheduling allows our customers and ourselves to have our cakes and eat them too.

We are now prioritizing production using a number of criteria with a goal of maximizing throughput. For instance, the software recognizes situations where saving set-ups improves output and minimizes change-overs by scheduling similar parts back to back.

As earlier mentioned, our inventory control manager puts in his make-to-stock orders biweekly. Our shop lead time is longer than two weeks. As often as not, we'll receive orders for two of Part A one week, five of Part A the next week and eleven Part A's the week after.

Keeping track, the software suggests we run 18 Part A's for delivery at the time the two original parts are needed. Instantly, we just saved two set-ups. Prior to Resonance, we were often unaware of what was in the order pipeline.

Results

At Dixie Iron Works, we have institutionalized the Theory of Constraints concept. We search for bottlenecks, identify them, exploit them, and start the process over again. We use our Resonance software as a scheduling tool. Directly as a result of having such software, we even canceled an order for a very expensive machine and instead ordered a less costly one.

This happened very early after installation. We had already ordered the expensive machine knowing very well we had a bottleneck in its destined area. Luckily, the lead time on that machine was several months. Once we cleaned up our data, we learned that we also had an interacting bottleneck with the ordered machine's area. Indeed, a less expensive machine put in this second bottleneck's area would solve most of the constraint problems in the original bottleneck area. The difference in cost easily paid for the software.

What-Ifs let us translate our business projections into various resource requirements. We can explore such top level options for long term decisions as what type of equipment we need to purchase or the kinds of process improvements we should employ.

We are able to provide due date quotations and assure ourselves we can back them up. After only four months of running Resonance, we had more than doubled our due date performance to 65 percent from 30 percent.

We also more than doubled our inventory turns to 12 per year. Most significantly, operating profit increased four-times over two years ago when we operated in our old mode, and 2.4-times over using constraint-based scheduling manually.

By automating scheduling, we improved on-time delivery, throughput and ease of shop management. Our scheduler literally asked us for more to do.

For more information about Thru-Put Technologies visit their website at www.thru-put.com