The supply chain software market has gone through some tumultuous times. The promise of advanced planning and scheduling (APS), “push the button and your problems go away”, turned out to be empty. As appealing as it sounded, and as excellent as the underlying optimizing technology is, it does not actually solve the problems in most supply chains.
Why doesn’t it work?
- Traditional planning and execution need to merge, execution needs to be based on frequent, virtually instantaneous, re-planning. The data needs to be updated frequently so systems can see the current supply/demand picture.
- Not only does the system need to include data from the entire supply chain – from distribution through to contract manufacturers, it needs to understand the supply/demand relationship between all the locations. This entire supply chain “model” can then change supply/demand and corresponding availability relationships dynamically as inevitable and frequent changes occur.
- Business processes (presumably including their supporting systems) need to be connected so that decisions are not made in operational silos.
But there’s one other key aspect of every supply chain: people. Automated systems and processes are great for handling routine events. They can even be pretty good at handling anticipated changes. They do not work well when something totally unexpected happens. This is where people come in!
People need to understand what has happened, how it impacts the operational plan, and what to do about it. Invariably, people must look at data in ways that are different from those used for routine operations. Ideally, the people need to simulate potential solutions to see if they are likely to solve the problem or to create new, worse ones. Furthermore, supply chain decisions are almost never made by a single person. People throughout the supply chain need to understand what has happened, what might happen, and what might be done about it. Given the frequency of change, it is not enough to run a bunch of reports. An infinite number or reports would be required to cover all the different ways people might want to see information. Rather, the participants need to choose and tailor the data they see as they look at it. Then, they want to change data to create one or more possible solutions. Once a course of action has been decided upon, then those actions need to be communicated throughout the supply chain so that all the parties can execute the ever-changing plan.
TrackBack
• Digg This
• Add to del.icio.us
Tags: Supply chain
Posted in Supply chain management
You can leave a response, or trackback from your own site.
Some years ago I was fortunate to visit Fedex in Memphis and was invited to see their operations center (housed in a 747-proof bunker which was formerly the global reservations center for Hilton Hotels). I should also mention that at that time, Fedex had some 5,000 people working in IT in Memphis. The Operations Center was stuffed full with high tech equipment, a quiet buzz of machines and people studying banks of screens. In addition they had one of the most advanced meteorological departments to ensure optimal scheduling of planes in and out of airports to avoid adverse weather conditions.
However, what really impressed me were the two guys on ladders moving magnetic blocks representing Fedex aircraft around on a huge whiteboard. It seemed out of place, and with all the IT resources hadn’t they found a technical solution? Right, despite all the computer power, they had not found a better way of scheduling their fleet of aircraft …only the human brain could handle all the variables associated with unexpected events impacting the fleet around the world.
Indeed, there is a chasm between planning and execution in many of today’s supply chains. My experience has been that operational planning happens either weekly or monthly and then execution is what it is minute by minute. It’s only when the planners sit down to create the next iteration that adjustments are made for for actually happened; even that isn’t always the case. Sometimes planners just continue on with their original case ignoring reality for a time.
So, I like the concept of execution based replanning in almost real-time. I think that is the ultimate solution. Has anyone mastered this approach?
Great question: “has anyone mastered this approach?”
I think the most honest response is “they are working on it”. Kinaxis has a product, RapidResponse, that does precisely that – execution based replanning in almost real-time. It also allows for unlimited simulations of real and potential scenarios, based on complete supply chain data. It provides visibility and collaboration amongst thousands of users.
Kinaxis’ customers include some well-known names including Honeywell Aerospace, Toshiba, Jabil, RIM, and many, many more. Have they “mastered” their supply chain? I suspect they would say not quite. Mastery is hard to attain when there is so much change going on. Are they coping, obtaining significant benefit from the approach? Absolutely!
While on the one hand it is important to close the gap between planning and execution, it is even more to make applications involved in different aspects of planning to speak with each other. Think about a Forecasting Process that spews out numbers that have no reflection of supply network resource constraints or a allocation strategy for distributing limited supply to multiple customer groups. Looping in the planning process is very important to make sure that planning applications are a true reflection of your supply chain strategy. Execution applications are even more downstream and closing the gap becomes very effective when the short, medium and long-term planning applications are aligned correctly. (Read the blog http://infosysblogs.com/supply-chain/2008/10/the_closedloop_supply_chain_st.html)
Blog reader “Jeff” writes:
To introduce myself, I am a Forecasting, Inventory, and Planning Manager at a company that is in an extremely volatile market. Our industry has taken quite a beating over the past 6 months. We have been chasing “the falling knife” as we say when it comes to trying to forecast where the markets are going.
I have recently read the article “Why You Need to Re-evaluate Your Approach to Supply Chain Planning” and was both excited and concerned. I was excited about some of the items the article talks about that we have already implemented. One of those is developing a tool that will allow us to “break free” of the limitations of our traditional MRP system. I have designed a system that recreates the MRP system but also allows “What If” and also allows the user to make changes to the plan on the fly. The updating of this system takes mere minutes to do, and is much more user friendly than the ERP system that we use.
The concern that I have is that our system relies greatly on the accuracy of forecast. We do not use any special tools for forecasting. I have in our system a forecast override so that we can see “what if” when the forecast falls short or is too great. Our business is very seasonal and we do not know which customers are going be in business from one day to the next. Therefore we base all of our forecast on market trends and sales history. Looking at where we were and how far up or down we are going to be. I have been in my position for almost 1 year and have learned that it is not if my forecast is right or wrong. It is how wrong I am. The break down of the data to do the forecast is done in my head. We have no set formulas to determine what it is going to be. We forecast in weekly buckets and look at how the market has performed in the previous weeks. Our longest lead time items come from Europe. They have a 2 month leadtime and they like to have a 1 year forecast. We give them this but make sure they understand that this is a guess at best. We have a hard time figuring out what we will be doing this December, let along next December.
Does there exist a method of forecasting that can keep up with these dramatic changes?
I appreciate the article and am 100% in agreement with it. I have been in Manufacturing for 5 years and I have come to appreciate the concept that you have to be able to be flexible and change on a whim. Change is good.
Thank you.
Jeff,
Thanks for your comments and your question. There are lots of approaches to forecasting and there are certainly companies that are reasonably happy with their forecasting tools. However, we are seeing that investing in better forecasting rapidly results in diminishing returns. As you say, it is very difficult to predict the future while looking backwards (at history). In any case, is past performance a very good predictor of the future – especially in the current global business situation?
It certainly makes sense to get as much input as you can from as close to your customer as possible – but there are practical limits to this. In addition, it makes a lot of sense to improve your ability to respond to inevitable changes, as they occur. An interesting revelation from the mathematics of statistics is that the error in a forecast increases with the square root of the number of random elements. Put another way, the forecast error over a month (4 weeks) is twice the error over a single week. Thus, you can cut your forecast error in half simply by being able to define and respond to changes every week rather than once a month.
Jeff further replied:
You have affirmed that the process that we have switched to is in fact the one the makes the most sense.
We implemented this process back in July and are testing it with our highest dollar finished goods from Europe. I attest to the statement that forecast error decreases with the frequency in which we respond to changes. We try our best to keep getting input from our sales staff, but as you say, there are practical limits to this. Before July, we were adjusting forecast and ordering once a month and would not catch the changes in time. Now that we look at it once a week, we are able to see these fluctuations and identify and ask the right questions to our sales staff. Our “Panic Orders” have disappeared and our excess stock is on the decline. Looking out into the coming weeks, we can see there is an order pattern developing for each sku as we have set a min and max based on the forecast.
Currently our greatest limitation is our supplier’s capacity. They are our sister company and cannot slow down fast enough to meet the decline in our economy nor our better grasp on ordering. Therefore, we have no choice but to order to their line rate rather than just what we need. The way we are dealing with this is ordering what we need in all models and then putting whatever excess we have to order in our fastest moving stock. We figure it is better to keep extra stock in something that is 60% of our business and we know will sell. The only thing our supplier does not get is, at some point, it has to catch back up.
Again, thank you for your response. It gives me a good feeling to know that my boss and I are not the only ones who think this way. To me, it just seems like common sense.