This forces companies to then rely on increased safety stock buffer points to reduce lead times to customers, or to avoid missed delivery dates, a key customer metric. All this adds cost or reduces customer satisfaction, leading to a deteriorating competitive position and a reduced bottom line.

The question is, how can this situation be improved? The answer is: forecasting should be viewed as an important tool in meeting corporate goals for growth and profitability, not just as a budget exercise. In order to leverage forecasting as a vital asset to the enterprise, following issues need to be addressed:

1. Forecasts are a dynamic variable, so they can change significantly over even a short period of time. This means the forecast process needs to be much more frequent than annually, preferably monthly. While a baseline is needed (budget) for financial accountability, the demand picture needs to adjust to reality.
2. Forecasts should cover your longest lead time items, in order to properly anticipate demand and not be caught short. This means forecasts need to roll forward, covering the full length of the demand horizon at any point in time.
3. In order to support a more frequent forecasting cycle and improve accuracy, the forecast must be streamlined and easy to use. A forecast which takes a month to prepare is already out of date by the time it is released. This requires good data to base assumptions on, and a tool which can quickly and accurately generate forecasts for analysis, preferably with the ability to quickly compare various scenarios in order to determine the optimal one for the current environment.
4. The forecasting process must be viewed as integral and important tool in the overall functions of the company. The people generating the forecast must be aware of its importance to the strategic interests of the enterprise.

I have heard some people comment that, “Our forecast is always wrong, so we need to look at better safety stock management tools.” While safety stock is an important tool to buffer against unanticipated demand fluctuations, a better strategy would be to invest in instituting a continuous forecast process as a competitive tool.

We are currently going through another tough phase in the business cycle, with consumer demand softening and business spending becoming very conservative. This slowdown is causing even more volatility in the market place, which will require even better and more frequent analysis in order to able to respond to the fluctuations taking place in demand. Having an accurate, regularly updated, long range forecast to provide guidance to the supply chain is more important than ever. A tool which can quickly and easily supply such a forecast is a must, as is a process to provide feedback on forecast assumptions to enable corrective action to keep the enterprise on track.

Kerry Zuber, one of my colleagues who is a Lean/Six Sigma black belt, has introduced me to two great terms that are crucial to understanding the value and key capabilities required of effective response management:

• Time to detect
  This is all about knowing sooner that some event has occurred that is creating risk or harm to your organization or supply chain.  This may be that a customer has changed their mind about an order, or a supplier has de-committed on a delivery date or quantity, or that a tsunami has occurred in Japan that has wiped out a large part of world-wide semiconductor manufacturing capacity. But not only knowing about the event sooner, but also knowing about the impact sooner (which lines will go down, which orders will be impacted), and, perhaps as importantly, knowing who is impacted sooner.  Without knowing all of these three things you cannot act.
• Time to correct
  Once you know what went wrong, you need to act quickly to find a solution through compromise across multiple functions, even multiple tiers, often with competing objectives.  The timeliness of resolution is a key measure of the quality of the solution. We still hear of companies that have not fully understood the impact of the Japanese tsunami on their operations and their ability to satisfy demand, let alone put into place a recovery plan. And yet it is often the day-to-day events, such as when a customer changes their mind on an order, where most margin or customer satisfaction is lost because either the response to the customer is too slow or actions are taking with little understanding of the financial and operational impact.

The significance of these two terms is that for the most part the physical supply chain is a constraint in that your ability to change it in the short term is very limited so every minute lost in detecting something ‘wrong’ and, once the event has been detected, every minute used to determine what to do to reduce or eliminate the risk or harm, is a minute taken away from being able to use the physical supply chain in some manner to reduce the risk.  Even in situation where there is next to nothing that can be done to eliminate the risk, if you know sooner you can at least tell your customers who are impacted about the impact sooner so that they have more time to react.

There are three core capabilities required that impact the time to detect and the time to correct:

• Being able to represent several versions of a multi-tier supply chain in a single data model so that there is a single version of the truth, and being able to represent several states – historical, present, and future – of the supply chain in order to determine what of significance has changed that is causing future harm to the organization, as well as several scenarios that represent alternative ways of solving the problems.
• Being able to determine who is impacted by the change and therefore who needs to be alerted, not only in your own function, but in other function within your own organization and people in your customers, contract manufacturers, or component suppliers.
• Being able to bring those people impacted by other the original threat or proposed actions together in order to collaborate on the resolution of a problem by testing several ways of resolving the issue and being able to reach compromise through sharing of the impact of possible courses of action on financial and operational metrics.

These concepts and capabilities can be applied to a broad range of business problems, not just the supply chain.  In the diagram on the left we can substitute other function’s analytics for the supply chain analytics and still retain the core capabilities described above to solve response management problems in a wide range of business activity.  Within manufacturing industries the supply chain is absolutely central to being able to get product to market.  But within manufacturing companies response management issues are not constrained to the supply chain.  Workforce management is an easy one to describe in this context because it is (or rather it should be) tightly coupled to the longer term revenue forecast. If a company anticipates their demand or sale revenue growing they need to ramp up recruitment because of the long lead time in recruiting and training people, but they also need to be very quick in shutting off the recruitment process if they see that their revenue is not growing as anticipated. In other words workforce management also requires the ability to plan for the future based upon anticipated demand, monitor actual performance against anticipated performance, and a quick response to correct the misalignments.

What drives the need for response management is the fundamental fact that we cannot predict the future with a great deal of accuracy and our success or failure is dependent on how effectively we respond to the present, how we manage in the “now”.  What varies across industries, across functions, and across business processes is how far into the future we need to predict what we think is going to happen, and how quickly we can affect change.  But what does not change is that the time it takes us to detect that something of significance has changed and decide on the best course of action determines both the time we have available to carry out the course of action and the feasibility of the course of action, which is of course a key measure of the quality of the decision.

So where do you think the next breakthrough will come from?  Better forecasting or better response management?

Will this problem ever go away? It is argued that if you have brand loyalty then the risk of perishable demand and worried investors is low. This argument holds for Apple where customers are willing to wait. Does the same hold true for BMW when their entire supply of 5 Series was consumed in one month in 2010? Some people just need to get a car. Brand loyalty may not be as influential. Which companies plan for limited supply versus the risk of excess inventory? The article talks about the Kentucky bourbon called Rip Van Winkle. Have you heard of it? Probably not as it is usually hidden behind the counter of the liquor store. They would rather keep production low than risk having inventory. Other companies with short life cycles may do the same. They must get their product to market as soon as possible but cannot risk the bottom line impact of scrapping product when demand is not meeting forecast. There is also the case where limited supply is not planned and can have serious consequences. Canadian company Lululemon faced shortages when their apparel line exceeded expectations and were forced to pay premium freight to accelerate supply. Margin erosion is often a result of demand exceeding supply.

So what does this really tell us? For the most part, forecasts are inaccurate. It has been proven that improving forecast accuracy results in higher customer service with the same inventory or the same service level with less inventory. How do you improve forecast accuracy?  Companies are finding more innovative ways to address this. Many are recognizing that improving demand response will reduce the cost and error of forecast error. Improved collaboration with trading partners; customers and suppliers also improves forecast accuracy. Duncan Klett has written an interesting white paper http://www.kinaxis.com/campaign/demand-planning-reduce-risk-and-impact on Demand Planning where he is really talking about the value of response management in the demand planning arena. He statistically proves that by focusing on customers with high demand variability, reducing cycle time (more frequent demand updates) will improve service and reduce inventory. Collaboration is another component where the sharing of up-to-date forecast information between trusted partners results in improved accuracy and reduced latency.

What are your thoughts?

Lora Cecere has been writing about forecasting, principally within the CPG industry for many years. She has worked in industry, for a software vendor, and most recently as a highly respected analyst. In a recent blog Lora states that

Mean Absolute Percentage Error (MAPE) for a one month lag was 31 percent + 12 percent.  Data eight years ago for the same companies was an average of 36 percent + 10 percent MAPE.

This made me sit up and listen.  Especially when she went on to quote from her research while at AMR Research that

Based on AMR Research correlations, a six percent forecast improvement could improve the perfect order by 10 percent and deliver a 10-15 percent reduction in inventory.

In other words, there is a lot of benefit to getting the forecast right.  But a range of highly respected CPG companies cannot do better than 31 percent MAPE, with a range of 19 percent to 43 percent?  That caught my attention.  Mostly because I am more familiar with the High-Tech/Electronics industry which has much shorter product life cycles than CPG and therefore more volatile or variable demand patterns. Of course it is difficult to be precise with industry classifications. Does Consumer Electronics fall into CPG, High-Tech/Electronics, or both?  However we slice it, things like cell phones, tablets, cameras, etc have shorter product life cycles, greater seasonal variations in demand, and greater demand variability than do nearly all categories of CPG such as soap, washing powder, etc.  In Consumer Electronics, and more generally High-Tech/Electronics I hear from companies that they seldom get their forecast accuracy, as measured by MAPE, above 50 percent, which is consistent with my observations about the characteristic differences with CPG.  Higher demand variability/volatility would imply a lower forecast accuracy.

Before anyone jumps down my throat, especially Lora, let my state unequivocally that everyone MUST forecast and that all companies should be demand driven.  But …

But where is the discussion about how best to satisfy the missing 31 percent demand in the case of CPG and 50 percent in the case of High-Tech/Electronics?  Where is the discussion about the profitable response to the demand that is not anticipated? I feel as we are only having half the conversation.  The half about forecasting.  But if the best we can do is improve forecast accuracy from 64 percent to 69 percent over eight years in an industry segment with relative stable demand, I think we should be talking about supply chain agility and responsiveness.  What amazes me is that since the early 1990’s we have been applying optimization engines, typically Linear Programming (LP), to the supply side.  Ignoring for the moment the inherent issue of using linear models to represent highly non-linear systems, if you are basing your optimizations on inputs that are best 69 percent correct, are you not focusing on the wrong problem?  Should you not be focusing on systems that enable you to detect true demand early and determine the best way to satisfy the unanticipated demand using the competing requirements of profitability and customer service?  Of course you will need a supply chain that can execute in an agile and responsive manner consistent with your decision.

Here is the rub: All our resources are limited. Time. Cash. People. So in this zero-sum game, where are you going to apply your energies?  Spending eight years to improve the forecast by five percent, or working on the manner in which you satisfy the unanticipated demand in the most timely and profitable manner?

While trying to address a real issue that has huge financial impact, I believe Max’s suggested cure is worse than the original symptom because of the distortion to the forecast. Getting a forecast that is as accurate as possible is vitally important and we should not be adding distortions. I also think that using a ‘periods of cover’ inventory policy instead of a fixed quantity SS is a very good way of providing a more dynamic SS. In commenting on the use of ‘periods of cover’ I state that:

The policy may be to keep 3 weeks of demand in stock because that is the supply lead time to the stocking point.  If the forecasted weekly demand is 10, 20, 20, 30, 10, 10 then the SS would be 50, 70, 60, 50 over the next 4 week.  In other words the safety stock fluctuates with the anticipated demand.

I always find it difficult to separate out the topics of inventory policy and postponement. In fact I don’t see how one can separate out these topics because they are so tightly related. In addition, postponement is often thought of in terms of manufacturing or assembly postponement, but it is equally valid to think of inventory or distribution postponement.  Your ability to postpone either manufacture or distribution of FG determined by the markets order-to-delivery (OTD) lead time expectations and your supply chains supply lead time and agility.  If the market expects a 1 day OTD lead time and your supply lead time is three weeks, there is very little opportunity for postponement, and SS must be kept as FG at the most forward stocking location.  If on the other hand the OTD lead time expectation is one week, SS could be kept as FG at the factory.  If the OTD lead time expectation is five weeks, then the SS can be kept as raw materials before manufacturing or assembly.

But the discussion above is fairly simplistic and describes principles. Hewlett Packard (HP) adopted an approach, much of it pioneered by Hau Lee and Corey Billington in the 1990’s, that, in my opinion, makes them a leader in inventory management.  Currently they may not have the ‘cool’ products that Apple has, but they have some great supply chain ideas. They focus a lot on both demand segmentation and product segmentation. For product segmentation they use the terms ‘No Touch’ (completely outsourced), ‘Low Touch’ (sub-assemblies outsourced), and ‘High Touch’ (mostly insourced). Actually they have five classifications in total. While demand volatility is one of the factors considered in outsourcing decisions, outsourcing has more to do with IP and core competencies.

In a webinar titled “Using Segmentation Strategies for Better Demand and Supply Balancing in the Mid-Market” (registration required) Jeff Range of March Networks discusses a very similar concept. March uses the categories of ‘Runners’, ‘Repeaters’, and ‘Rogues’ to segment products, and the categories ‘Fixed’, ‘Flex’, and ‘Forecast’ to segment demand within each of these categories. Supply chain agility, or postponement, and inventory policy decisions are made for each of these intersections. This is very different from most approaches which tend to set inventory policy by product group only.

While slightly different, the diagram below, from a presentation made at Stanford University, shows how HP segments demand. For the ‘Lo scenario’ they put in place long term minimum cost contracts based upon regular and guaranteed quantities and a guaranteed, but longer, lead time. For the ‘Base scenario’ they put in place shorter term contracts that balance flexibility and price, definitely at a higher unit cost than the ‘Lo scenario’ and also possibly a more variable, but shorter, lead time. The ‘Hi scenario’ is essentially a spot buy, greatest flexibility (uncommitted) but also greatest cost and possibly the most variable lead time, which can also be offset with higher cost through the use of ‘expedited’ transport to reduce lead time even further.

What is hidden in all of this is postponement.  The presentation is title Procurement Risk Management (PRM). For the ‘No Touch’ products clearly procurement means of FG, for ‘Low Touch’ it means for sub-assemblies, and for ‘High Touch’ it means for components. Of course I am making broad generalizations and the decisions are more subtle than I describe. But for all these product classifications there is still the question of the difference between supply lead time and order-to-delivery lead time.  Postponement can only be used as long at the supply lead time from the point of postponement is shorter than the expected order-to-delivery lead time. If the expected order-to-delivery lead time is a lot shorter than the supply lead time, then the safety stock buffer has to be in FG. If the supply lead time is shorter than the order-to-delivery lead time, then the safety stock buffer has to be kept in the raw materials or components.

To get back to Max’s question, how should you calculate safety stock? Should you even have safety stock, or should this be dealt with in the manner that HP does? First of all I think phrasing the question in terms of FG doesn’t get to the heart of the issue because it all depends on the level of postponement. So let’s generalize the question to ask if this if forecast consumption is not more sensible than the use of safety stock at the point of postponement.

The most common way to calculate safety stock and the reorder point to satisfy a certain customer service level expectation, where Lead Time is the supply lead time to the buffer, is as follows:

1. Z: NORMSINV(Service level) , for example Z=1.64 for a 95% service level
2. Safety stock: {Z*SQRT(Avg. Lead Time * Standard Deviation of Demand^2 + Avg. Demand^2 * Standard Deviation of Lead Time^2)}
3. Re-order Point (ROP): Average Lead Time*Average Demand + Z*SQRT(Avg. Lead Time * Standard Deviation of Demand^2 + Avg. Demand^2 * Standard Deviation of Lead Time^2)

In other words, if demand is fairly stable (relative to volume) and the supply lead time is fairly stable, very little safety stock is required. This is where HP was really innovative. For the ‘Lo scenario’ they were able to reduce inventory throughout the supply chain by putting in long term contracts that reduced variability of supply and for the demand for which they could assume zero variability, therefore effectively reducing safety stock to zero, but taking a bit of a hit on ROP since their average supply lead time was a bit longer. On the other hand for the ‘base scenario’ they accepted greater variability but focused instead on reducing supply lead time, which reduces both safety stock and ROP. Of course this strategy works best for mature products with low demand variability relative to volume. For early stage NPI nearly all demand will be ‘Uncommitted.’  As far as I am aware HP still put in place long term contracts based upon their estimates of total available market and expectations of market share.

In summary, I think Max is bringing up the right question. But I think the answer lies in deep analysis of both demand and products to determine differentiated inventory policies – which includes postponement, SS, and ROP – to best arrive at what Hau Lee calls a Triple-A supply chain: Agile, Adaptable, and Aligned.

What do you think? Is this too complex? Should we instead just add some quantity to the demand as Max suggests? It would be great to hear from practitioners on how they are solving this issue.

Typically, a forecast is developed first by applying statistical model(s) to sales history with the result referred to as the ‘statistical’ forecast. The technical statistical forecast is then adjusted by various other fundamental factors such as product life cycle phase, seasonality, or marketing or sales promotions, usually as part of the Sales and Operations Planning (S&OP) process. The resulting forecast is typically known as the ‘consensus’ version of the forecast. However, related to planning, and to account for potential upsides in sales and to maximize customer service, some buffer or uplift may be added to the forecast. Moreover, this uplift can be driven in the planning system by being added to the forecast or as safety stock.

From my perspective, the basic difference between adding a buffer quantity to the forecast versus driving safety stock in the planning system is that a buffer quantity in the forecast could be consumed by actual customer orders if the buffer is realized by actual customer orders. On the other hand, safety stock (depending on the approach utilized) will plan to keep a certain level of inventory regardless of customer order levels. Conceptually, I am wondering what the best approach is for this. To summarize it seems that there are three options:

1. Forecast should inherently account for potential upsides in customer demand based on the statistics used and the fundamental factors or adjustments applied.
2. Additional ‘buffer’ should be added to the forecast – advantage is that the forecast can be consumed by actual sales orders and the disadvantage is that this buffer needs to be separately maintained.
3. Safety stock should be planned in addition to the forecast – advantage is that the buffer does not need to maintained and can be calculated by the planning system, whereas the primary disadvantage is that additional or excess inventory may result.

Intuitively, to me it seems that the option 1 above should be used – let the forecast drive demand in the planning system and not attempt to plan any buffer or safety stock. However, as we know, the first rule regarding forecast is that they will be wrong. Therefore, to complement, a way to get early signals that there will be customer demand greater than the forecast and have a system to rapidly respond to potential shortages in supply needs to be present.
Do you have any experience or insights into this? Please let me know your thoughts.

Note: I posted an excerpt of this piece on the Supply Chain Expert Community – Join the discussion!

The benefits of increased forecast accuracy include:

• Increased customer satisfaction due to reduced stock outs
• Reduced write-off and obsolescence costs due to unsold inventory
• Lower inventory carrying costs, as less buffer stock is required to cover missed forecasts

Significant benefits can be realized in a company’s top and bottom line results if a more accurate forecast generation process can be realized. One of the cardinal rules of forecasting is that a forecast becomes more inaccurate as it goes further out in time. Another is that a forecast becomes more inaccurate as the granularity is increased (ie. It is easier to get an accurate forecast at the business unit revenue level than it is to forecast the number of units of each SKU that make up the sales for that business unit).

This second rule can be better addressed by focusing not an individual SKUs sales history or possible future trends, but by looking at the ‘group’ this SKU(s) fall into and analyzing its history and trends. The attributes shared by these SKUs constitute their forecast group. The advantage of grouping forecast SKUs are many:

• A lot of market research is available at a higher level in the marketplace (at the channel product category level versus an SKU level). Market share analysis must be done at this level
• Aggregating SKUs reduces the month to month demand fluctuations seen at the SKU level, allowing for better data analysis for trend forecasting (reduces the noise)
• Revenue forecasting generally occurs at the product group level or higher. Since revenue projections are the most common driver of forecast models, a better alignment between forecasts at the SKU level and revenue forecasts can be achieved if there is a clearly understood link between the two
• A better alignment of revenue and SKU forecasts allows the business to more clearly drive production (also known as the sales versus production forecast). A poor alignment between the two can lead to missed sales numbers, increased inventory carrying costs, or stock outs.

The breakdown of the revenue forecast into a realistic SKU level forecast can be complex, and can vary greatly among different business units and markets. While these different processes can lead to a more accurate forecast based on local market conditions, it makes it very difficult to get a true picture of the alignment between the global revenue forecast and the individual unit forecasts, as the methods to get from one to the other can vary greatly.

This leads us to the recognition for the need of a formal, structured process for generating SKU unit forecasts from revenue forecasts that can generate a better forecast for the business as a whole. In order to allow the process to accommodate the differences between markets and regions, it must be have the functionality to allow for market specific processes, while maintaining a common underlying process that can tie back to the revenue forecast. The best way to achieve this goal is to use region and market independent product attributes, as they can ‘filter out’ the variations that can occur due to different business cultures. This also describes the best product attributes to select, as they should reinforce the underlying commonality of the SKUs being forecast. Market specific processes can then be applied to fine tune the forecast for a specific market or region.

In order to more accurately predict future demand at the SKU level, a tool is required can process large amounts of historical data and knowledge based assumptions quickly. This allows the analysts to perform what if simulations and determine the effects of various assumptions, as well as incorporate any trends that are evident in the marketplace. This leads to another “must have” feature: the ability for the analyst to easily and quickly incorporate trend simulations into the numbers. By using aggregated attributes on multiple levels and presenting the numbers in a percentage of total view, the analyst can quickly modify the splits at each attribute level to reflect known assumptions or model future trends, while keeping the totals within the revenue guidelines (the sum of the percentages must always equal 100). As well, by modifying percentages versus actual numbers at the lower attribute level, the overall revenue forecast can be kept consistent throughout the forecast process. The revised percentage splits can then be applied to the revenue numbers to break them down into a SKU level forecast that ties back to the revenue in an understandable way.

In summary, it is desirable to have a close correlation between the revenue forecast and the SKU level forecast. The best way to get there is to use product attributes, historical patterns, and the analyst’s expert knowledge to break the revenue forecast down to an accurate SKU level forecast. A tool that can support an underlying, formalized process with enough flexibility to allow for regional and market differences will deliver the optimum solution.  This tool must be intuitive to use and allow the analyst to quickly generate a valid forecast that supports the consensus revenue projections.

Ah, enough of that….There is a great article by David M. Katz in CFO magazine about demand planning leading up to the holiday season. With all the talk of a double-dip recession, CFO’s are asking themselves do they ramp up for growth or hunker down and wait out another drought? 

Ultimately, the question is – how and what do you forecast?  There are risks on both sides if you get it wrong – excess inventory or missed sales opportunities.  As such, at times like these where there is little to no predictability, experts are now saying that the ability to respond and react to demand is more critical than the ability to forecast/plan it.  Can I hear a hallelujah!

As the article points out…

In these volatile times, it’s hard for companies to get a reliable read on what to expect from their customers and how to deal with rapid shifts in demand. Here are four steps that supply-chain experts say are essential to coping with a fast-changing economic landscape:

1.Ditch the long-range forecasts.

2. Avoid gut feeling.

3. Go granular.

4. Launch an S.O.P.  

Any other steps that should be included?

Given the assumption that a plan is crucial, together with the realization that the plan will not be accurate, we are led to the conclusion that we need a stable plan, but be able to adjust the plan as and when needed. We need to be able to adjust the plan only when significant enough factors warrant a change to the plan, and with enough lead time and stakeholder buy-in to execute properly.  To restate, I believe that the following are important:

1. Plan Accuracy and Stability – The MRP plan needs to be stable enough to enable effective execution but we need to be able to detect exceptions that are significant enough to warrant a change
2. Responding to Change – The capability to effectively respond to required changes needs to be in place

How do we effectively accomplish the above?

Plan Accuracy and Stability

• First, the plan needs to start with an effective Sales and Operations (S&OP) process.  The more robust the S&OP process, the better that the high level plan will be.   
• We need to be able to detect or sense the need for changes, and once needed changes are detected, we need to be able to discern which are significant enough to warrant a change to the plan. 
• We also need to be able to prioritize these since there may be more than we can deal with. 

The key is that potential problems such as material shortages and late customer orders need to be detected for the future.  Obviously, once late orders or shortages have occurred, they are easy to detect (maybe even by way of angry calls from customers or buyers getting urgent messages from production regarding shortages.)

Referencing a recent blog post by Kerry Zuber, “Driving performance improvement through exception management“, Kerry states that in some organizations, there can be as many as 30,000 action messages generated by an MRP regeneration.  This exemplifies the complexity of the MRP plan and sheer volumne of exceptions in many organizations.  The organization cannot work all of these recommended actions, but which ones are the right ones to work?  Which ones signal that something in the higher level plan needs to be adjusted?  A second level, automated process needs to be in place in this type of environment to prioritize actions and also alert the responsible parties. 

The capability is required to detect what future demand will be late due to the mis-alignment of supply schedules,  issues with capacity in the supply chain and other issues.  If the future state/impact cannot be detected, then adjustments or contingencies cannot be put in place to avoid or mitigate them.  And as mentioned, we also need to be able to determine which of the detected changes require action and by who.

Responding To Change

Once changes are detected, we need a process to effectively implement these changes. This involves two key process and system capabilities:  simulation and collaboration.

We need to be able to simulate what-if scenarios to determine how best to deal with the change.  For example, it is difficult to calculate what the impact of a supplier changing commitments on a PO schedule will be in many environments without being able to simulate what that change in the commitment does to the overall plan.  In developing a response to the change, we need to be able to simulate multiple action alternatives and assess how well they will solve the problem and also whether they are achievable.

These simulations cannot be done in a silo.  Any significant change needs to be collaborated on with the extended supply chain.  Collaboration is certainly required with other internal organizations and potentially with affected external suppliers.

I realize that the above is very high level and probably over simplified, but I believe the general concepts are necessary in a complex manufacturing environment to optimize planning.  Without an optimized plan, execution cannot be effectively and efficiently accomplished and we have to resort to a lot of brute force exercises, including expediting.  Even the best of plans needs to be monitored for required adjustments and we need to have effective processes and systems in place for responding to these changes.

Has your organization implemented a process for responding to change?

Stephen Mills (who responded to the LinkedIN version of the post) talked about using what we know about past relationships and other key variables that may be in the future to determine what sales, demand and production should be. It’s the “shocks” to the forecast that can’t be built into the model. How you respond to the shocks will determine the impact of an inaccurate forecast. Running scenarios to determine the impact of “future shocks” to your replenishment times, inventory policies and customer relationships, etc. all play a factor on how accurate the forecast needs to be. Stephen also supplied the best quote related to forecast accuracy,

“Forecasts are either wrong or lucky.”

Stephen points out that a robust end to end supply chain will ensure that an inaccurate forecast doesn’t mean bad luck for the business. It’s only one piece of the puzzle.

Another respondent also pointed out that forecast accuracy was only one piece of the equation. This response also talked about forecast communication. Communication between functional partners on everything from market trends, process improvements and “shocks” are discussed in a timely manner so adjustments can be made in time to improve the business. Some relevant examples included the case where demand for a product may be unexpectedly soft, so marketing may shift promotions to help on the sales and supply chain side. Finance would also be in the loop so they could adjust their balance sheets.

I believe overall respondents agreed that the need for the forecast to be accurate is a function of such factors as the cumulative lead time, safety stock policies and flex capacity. Continuous improvement activities around lead times and quality will take some of the burden off those responsible for developing the forecast. Operational excellence and the ability to respond to “shocks” are a competitive advantage when unexpected demand opportunities present themselves. One response pointed out that this introduces an element of “time” to this issue of forecast accuracy. How good the forecast needs to be will be dependent on the range of the forecast and service level policies, especially on critical lead time items.

As pointed out, forecast should not be left on its own but accompanied by all the background and risk information so demand plans can be set, supply rationalized and plans easily re-evaluated if the “shocks” hit. This was only a small sample of some great insights from the blog responses. Thanks to all those who participate in these discussions. It really is worth it!