Chapter 6, Strategy, Coordination, and Planning
Plant managers often have to make difficult production decisions quickly. In this excerpt, find out how to base supply chain management decisions on strategic models for success and learn what questions a manager should ask when faced with supply chain challenges.
In Pursuit of the Perfect Plant: A Business and Technical Guide, Ch. 6
Table of contents:
Sales and operations planning for manufacturing
Manufacturing plant information management
Information integration in manufacturing: Process control and planning software
Manufacturing supply chain strategy: KPI and execution in process industries
Manufacturing operations management and planning based on strategic models
When the Rubber Hits the Road: Decision-Making Based on Strategic Models
"Suppose you have to decide to shut down a line or defer maintenance 30 days," Moulton continued. "The less time it takes for corporate to make a decision, the better off the person making the call in the plant will be. Often, these are seat-of-the-pants, intuitive decisions. For instance, say a pump goes down on New Year's Eve. The plant foreman calls a maintenance guy, who looks at the spare pump's record. If it looks good, he makes an instant decision not to call out a crew that he knows will cost triple time. If he can't make that decision, it rolls up one level, and it keeps on rolling up until someone finds the economics that will enable a good decision. Absent the data, the plant manager will go to the floor and say, 'Damn it, you keep this place running, whatever it costs.' He knows what his risk-reward is. He's going to get an 'Atta boy' if he gets the thing fixed that night, regardless if it costs him $10,000 to call out guys from the party. He also knows that if it stays broken, he's going to get canned."
"But there's got to be an alternate approach, one that's a bit more sane," Bala said.
"The alternative is arming yourself with more facts and figures," Moulton said. "The shift foreman needs to know, for instance, that every time he loses capacity because of an ailing pump, it's going to cost $10.00 a barrel. Therefore, he always needs to know the condition of his spare pump. If he's not willing to risk that $10-a-barrel shutdown, he also needs to know it's only going to cost $3,000 to call out the repair crew to make things right. In a more sophisticated plant, better data enables you to reach the same decision with greater confi dence. You know that your mean-time-between-failure for this kind of pump is three years. This pump is only a year old, so there's no reason to panic."
"Let's now take a look at our best questions," Moulton said. "Assuming you have an hour to walk into a plant and fi re away at anyone, here's a list of questions to ask to get a good sense of how well their strategy and planning are coordinated with the realities of the plant. The answers provided by managers and executives at a plant should indicate how well the strategy, coordination, and planning are working together."
Moulton wrote on the board:
1.) Is this plant run as part of an overall supply chain?
"Do you make what you sell," Moulton said, "or do you sell what you make? Are you the original manufacturer, like an auto-maker, or are you reconstituting a commodity, with a market on either side of the plant? Refi ning would be a good example of the latter. Generally, this would be a question to ask senior management before you went down to the fl oor. It helps determine whether you're looking at a discrete factory or a process factory."
2.) Are production planning and maintenance planning done together?
3.) When you make scheduling decisions, do you know the impact on the other guy's KPI?
4.) Do you implement any mechanism to systematically and comprehensively address these issues and make decisions?
5.) Does top management understand the risk or degree of control of a planned or unplanned shutdown?
6.) If there is an unscheduled breakdown or shutdown, can you opportunistically schedule other maintenance, or are you just coping? To what degree are you able to look at that as being 'in control' or 'out of control?'
7.) How reliable are your metrics for managing the plant?
"I have a question about number six," Mulcahy said. "What sort of answer might suggest that you need to contend with certain risks in the way you run the plant?"
"The worst answer is a blank stare," Moulton said. "The best answer would be that they'd established an acceptable level of risk and know exactly what the plant can tolerate—say, an 8% risk of breakdown. For example, they might have spreadsheets and hard data and some concrete information on which they can base an estimate. That's defi nitely better than nothing. If they've been running a production environment and gathering data about how long things will run for before they need to be shut down and repaired— a.k.a. preventative maintenance—they'll be able to provide an estimate of the failure-free interval and an assessment of a guaranteed failure point. Somewhere between the guaranteed failure point and the failure-free interval there's a curve that starts to change. As you cruise through the failure-free interval, your risk starts to increase.
"In other words, if you had 24 hours between the failure-free interval and the guaranteed failure point, then when you're at 12 hours over the failurefree interval is your risk of a breakdown now at 50%? Maybe, maybe not. It depends on the piece of equipment. But if you can assess the amount of risk in that little interval, and then add up the risk over all the pieces of equipment, you can get an assessment of the overall risk in the plant."
"OK, so the overall risk is the average risk of all the equipment in the whole plant. That is easy," Bala said. "In sum: the more data you have, the better decisions you can make."
"I have more best questions," Moulton said, returning to the board.
More Best Questions
8.) Does the plant compare its planning program prediction with what's actually happening in the plant?
9.) Do you know what S&OP is?
"I want to tell you a story about question number nine," Moulton said. "As you know, I am an advocate of S&OP, which is why I went to a seminar on strategy and vision as preparation for giving this presentation. The presenter asked the room, 'How many of you have heard of S&OP?' Amazingly, only 50% of the participants raised their hands. Then the presenter asked, 'How many of you have a defi ned S&OP process?' Guess how many people raised their hands then? Just one! These were sophisticated companies in attendance, too."
10.) Do you have a visible leader of your supply chain?
"This question was one of many from an article I read," Moulton said. "It asked a number of best questions. Its basic proposition was this: if you can't answer these questions as 'yes,' then you probably are the problem. I suggest you all read it."
11.) Is there visible support from the executive level?
"How would you know whether the support was visible?" Mulcahy asked. "You would ask the question more specifi cally," Moulton said. "'Do my CFO and CEO have visible knowledge and interest in the outcome of that meeting?'
This spawns further questions, such as…"
12.) How well does the meeting reflect reality?
13.) Can you use the meeting to adjust to either market or operational situations?
14.) Are the results understood and incorporated into the perfect plant?
"One way to know that the results are not being used productively," Moulton said, "is to ask the manager if he is continuously juggling calls from multiple sales people, production management, and senior management asking whether they need to change grades or insert a given order. If so, there are people who don't understand what the manager is being measured against. There isn't a clear distinction of roles."
15.) How many of your maintenance decisions are made with good data, tested data, or quality information, and in how many of them do you let the economics do the talking?
"I think we all know the most likely response to this question," Moulton said.
Perfect Plant Playbook
"To close this afternoon," Moulton said, "I'll take you through some quick pointers on how to get started with optimizing strategy, planning, and coordination at the perfect plant. You can make a real dent in a strategically challenged organization armed with little more than a pencil and spiral notebook.
"Let's start with simple supply-demand balancing. It's amazing how much it can enhance your overall understanding of how to improve things. It just gets people thinking in a more collaborative manner and forces the issue.
"And before you worry too much about software, it's critical to account for your plant's S&OP. If nobody seems to know what it is, then either you don't need it, because operations are so simple, or you need it badly. If your planners are creating 90-day targets but you only get a feedback loop from the plant once a month, you need an intermediary enforcer, or group of enforcers and mediators, to ensure that the information fl ows between the two camps. Without the intermediary, valuable information is just leaking away.
"Suppose you do want to create a responsive, tactical planning system with real-time updates to the process representations in the model. It may involve the scary but necessary step of ditching your current planning tool. A familiar system that's inadequate now will only limit the performance of the integrated system.
"You also need to examine the structure of your process. Before installing any software, you may need to reorganize your people. In fact, if you don't do this first, the stage may be set for a big software failure. Many plants are run from the top-down, where the plan is imposed on schedule and processes. If processes deviate from the plan, the process supervisors or plant managers have to explain why. This results in poorly defined targets for advanced process control systems when it comes time to automate. If you look at strategy from a bottom-up approach, where the plan is tuned to the process, the processes are still subject to targets and specifications. But you can update and re-run the plan as many times as necessary to balance it with the processes. Once you figure out if that works, then you can think about the software you need.
"I'll leave you with 11 critical success factors," Moulton said, and wrote more on the board, "that I cribbed from a lovely book called Planning, Scheduling and Control Integration in the Process Industries, by C. Edward Bodington. "In order to implement a successful integration of planning, scheduling, and control, you need…"
The Critical Success Factors
- A realistic and well-documented business process model and material flow diagram of what you want and expect to achieve, the timing, and the activities and ownership identified.
- A Process Information System for real time data collection and analysis of process operations that is blueprinted and configured to support the expected business process model.
- Data communications networks for easy, standardized transfer of information between elements.
- A central, rationalized database for storage and dissemination of all input and results.
- Model-based data reconciliation both for the overall plant, and for the individual processes.
- Advanced Process Control, including process optimization.
- A formal computerized scheduling system to manage process interactions and inventories, to capture and report production and shipments against plan into a form useful to the advanced process control systems.
- A responsive tactical planning system with real time updates to the process representations. A plan redeveloped as often as necessary to maintain coordination between customer requirements and the processes.
- Expert System applications imbedded in all the other elements to interpret data and results and improve the efficiency of data analysis.
- Process models tuned in real time so that planning predictions for current operation agree with actual performance.
- An organizational structure that rewards cooperation and teamwork.
"While this advice was gained from experience in the process industries, it's not hard to generalize these lessons," said Moulton.
"It is clear these points come from experience," said Mulcahy. "This was a fascinating session, Peter. Thanks for your effort."
Read other excerpts and download more sample chapters from our manufacturing ERP bookshelf
This was first published in November 2008