David Zatz

One of the interesting things about forecasting finished goods is how many functions participate in the development of the forecast and are subsequently impacted by the results.  In a typical S&OP and consensus forecasting process, Sales, Operations and Planning collaborate to reach a best forecast to drive the operations and profit forecast for the company.  We struggle to make this process work because each function has a different view toward the development and use of the forecast.  Supply Chain needs the forecast at the SKU level by location, usually in cases.  Sales is forecasting revenue and cases by customer.  Finance is calculating net revenue and profit based on the product mix forecasted usually divided into price category groups.  And Marketing forecasts at the brand level in gross revenue for a mid to long range time frame.  You can put these differences into a chart that looks something like this.

And the ways we each arrive at our forecasted volume is different which makes for a diverse set of numbers to start the process.  The S&OP process is not about the differences between the functions but the need to reach a consensus to find the one number, the best number to drive the operations and forecast profit for the company.  In order for an S&OP process to be successful, those are two of the goals that must be achieved: agreeing on “one number” which is of course a series of numbers, and that it is the best guess call, not biased toward under calling to beat the forecast or over-calling to ensure adequate inventory.

Most forecasting techniques used, especially in the Supply Chain function, use actual shipments from the past to project the future.  The tools to do this have gotten very sophisticated over time and enable us to use huge extracts of data to generate fairly accurate detailed, low level forecasts going forward.

Forecasting in the Marketing function utilizes a very different approach which brings a perspective to the S&OP process that enables the diverse views to reach a consensus and find the best number when compared to statistically derived forecasts.  In Marketing, we start with the actuals from the same month last year, itemize the factors that drive our business, and compare the revenue impact of each of those drivers compared to one year ago.  These drivers include

• Advertising
• Consumer Promotion
• Base Velocity
• New Products
• Merchandising (Trade Promotions)
• Base Price
• Distribution
• Inventory

Some businesses also separate out large or unique customers because that volume and shipment history is stored and treated differently.  For example Wal-Mart forecasts are often isolated because that volume is not part of the Nielsen data and because the volume can be such a large percentage of total shipments.  Other channels like club stores, Dollar stores, Military sales, export, etc. may also be managed and forecasted by a separate part of the business based on the drivers unique to each channel.

To calculate each monthly forecast for the core business, start with the actuals from last year, add and subtract the volume impact of each driver compared to that driver a year ago and arrive at the forecast for each month this year.  The technique to calculating each of these drivers can be simple or complex but clearly are geared toward developing the most accurate forecasts over time.  For example, for advertising, you compare how much you spent last year during each month, to how much you plan to spend this year and multiply that by a calculated return on that advertising investment, or payback to arrive at an incremental volume as a result of that advertising campaign.  And these calculations vary depending upon what product line or flavor is advertised and what media is used (TV, Print, Digital, etc.).

In Orlando at the IBF’s Supply Chain Planning & Forecasting:  Best Practices Conference, I will be talking more about the techniques used by Marketing to generate a quality, mid to long range forecast.

David Zatz
Marketing Forecast Planner
Post Foods

See
David Zatz Speak in Orlando at IBF’s:

$895 USD Only for Conference

Ed White - Bayer Canada

I’m sure that all of us have played the Telephone Game at one time or another. One person starts a message that travels through several people and then we all have a good laugh over how much that message changed by the time it got to the last person. As funny as that can be in a game, think about the consequences if the information is critical in some way. Similarly we have all seen war movies where the captain of a ship issues an order, which is then verbally repeated by an officer, then by another sailor and so on until it gets to whoever does the actual action. Ever wondered why they go through this whole process of repeating when everyone standing on the bridge heard the captain’s original order?

There are two purposes to this; first and foremost it is to avoid the telephone effect. To ensure that the message is heard correctly before it is actioned. If it is repeated incorrectly then there is an immediate indication of a communications failure. The second purpose is to ensure that everyone else in the chain of command understood the order correctly. On a normal day that may not be a critical issue, but when the warship is in action it is very possible that any person in that chain of command could abruptly be the person in charge. That means that everyone needs to clearly understand what is happening and planned no matter what chaos is going on around them.

Obviously, forecasting is rarely a life and death type of situation (though it can be) but it is still critical that all communication be clear and understood. If the wrong information is used in the forecast process or the information is biased in some way by the process then the final result will not be everything it should be. (Think GIGO). The problem is that most people take communications for granted. They assume that everyone understood them or that they understood what they were being told and this is not always correct. If the information used to create the forecast, the S&OP process, the plan, whatever, is not understood and interpreted correctly then the results will be sub-optimal.

I am looking forward to the chance to meet with other IBF Best Practices Conference attendees and sharing thoughts on improved communications and why that is important to the forecasting community.

Ed White
Supply Chain Specialist
Bayer Canada Inc.

See
ED WHITE Speak in Orlando at IBF’s:

$895 USD Only for Conference

Navistar

Why don’t you have these in stock?

Why do you have so many of these in the warehouse?

Why is your fill so low?

Do these questions seem familiar coming from members of your management team or your customers?  Of course they do!  They are constantly asked throughout the year as we all try to balance customer expectations and fiscal responsibility by managing inventory at the appropriate levels.  But with so many variables within the supply chain how can we accomplish this goal?

It is no secret that forecast accuracy is connected to inventory levels; therefore, the more accurate our forecasts the lower our safety stock.  This frees up capital for other parts of the organization.  So how do we improve forecast accuracy?

I strongly believe that building a collaborative environment in our organizations as well as in our supply chains is the best way to achieve the lowest possible forecast error.  By building a cross-functional collaborative environment of all stakeholders we can proactively identify and control the associated variables of both the forecast and the supply chain.  For example, consider a large promotion where product is heavily discounted.  By leveraging the relationships with our suppliers, internal marketing, sales, and distribution network we can proactively work to maintain lean inventory levels while ensuring product availability to our customers.  With this coordination we can obtain information from our sales teams to gauge potential interest in such a promotion, and ensure that the marketing information is available to our customers so that they are aware of the timeline of the promotion.  Plus, we’ll know if our suppliers are capable of supplying the product in a timely manner and that our distribution network has the capacity to receive and ship, ensuring overall customer satisfaction.

At Navistar, we are constantly looking for ways to improve upon collaboration within our supply chain and forecasting efforts.  We have recently implemented a new supply chain system that exchanges information between Navistar and its suppliers helping us to have the right part, at the right place, at the right time!

I look forward to expanding on these concepts and sharing Navistar’s implementation experience with you at the IBF’s Best Practices Conference in October!

Joseph Motta
Forecasting
Navistar Parts Group

See JOSEPH MOTTA Speak in Orlando at IBF’s:

$895 USD Only for Conference

Maria Simos CEO e-forecasting.com

A big treat in attending a conference is not only learning from attendees, but also having the benefit of one-stop shopping when it comes to vendors in the exhibit hall.  These vendors travel across the country, often times lugging huge displays, screens, white papers and swag to meet with current and potential clients and share with them what their software can do to help assist in their planning needs.

What better way to make this trip even more worthwhile than to share some top trends and news from the companies that have made the trek to exhibit this week at the IBF Best Practices Conference?

Top tier sponsors of the show JDA is coming in with major company news.  They are now the largest single company for supply chain planning and optimizations thanks to their recent acquisitions of competing firms Manugistics and i2 as recently as January.  With this synergy, the company now has over 6,000 companies across different industry segments using their software.  Danny Halim, VP of Industry Strategy and Calvin (Cal) Otto, Business Development Manager shared that what makes JDA truly unique is the company’s intimate knowledge  across the entire supply chain. This includes everything from raw materials to the retail space with the consumer experience.  Their company recently announced record first quarter profits, making Q1 the 22nd consecutive profitable quarter for the firm.  A major trend they see is the idea of supply chains competing versus one another rather than individual companies doing so with a convergence of the supply chain.

Smart Software and their Director of Sales Gregory Hartunian shared some impressive news that they have received not their first, but their second National Science Foundation Research Grant (NSF).  Ten years ago they were awarded their first Small Business Innovation Research Grant from NSF to develop a technology called the Smart-Willemain method of forecasting intermittent demand, also known as slow moving demand.  With their second NSF grant, Smart Software will further expand upon the Smart-Willemain method.  With this research completed, they will be the only vendor to offer a ‘next generation’ forecasting solution for slow moving capital goods, like service and spare parts.  Companies use this technology for a variety of applications, Kimberly Clark is using this to track in-house inventory as an example.

Tom Reilly from Autobox shared news of a new joint project with HP which was presented in more detail  Friday.  For this project, they were approached by a Principal Scientist of HP to work and develop a semi-hourly forecast model.  By breaking the day into 48 discrete time periods they are able to better determine precise demand at specific times throughout the day.  This methodology has been used for the last three to four months with application in call centers.  This method also easily translates using a mixed frequency modeling approach for making power estimations for power plants.

Forecast PRO’s Trac has a neat feature which shows how well the model fits with the history.  Bob Leonard gave a brief demonstration showing the archived forecasts over time.  Using this rich forecast archive helps track the accuracy of lead times.  Their software is off-the-shelf and a 5 user system can be implemented for $15-22K.

Boardwalktech Inc will be launching the 3.2 version of their software this June.  The company’s collaborative platform supports concurrent multi-users  down to the cell level using a back end system.  The software is easy to use and can be role based.  The real-time server recognizes who made the last change and makes notations.  Benefits of this system include integration that takes place in weeks not months, it extends the collaboration process, reflects a complete picture of the business and provides greater visibility.

SAS is excited to announce a new forecasting server plug in for SAP.  The plug in, called SAP Advanced Planning and Optimization (APO) links to read and write from live cache.  In other company news, IBF long standing member Mike Gilliland’s intramural basketball team has won the SAS intramural championships the last 2 out of 3 years.  (It’s not always about the forecasts, demand planners also need to have some fun, too.)

John Galt Solutions Inc. has an Atlas Planning Suite which focuses on the consumer-driven supply chain.  The suite allows for use of POS data to help assist in reaching higher levels of forecast accuracy and has over 30 models built in for planning new product launches and promotional events.  Using POS data and forecasting new product supply are also topics that were touched upon during the speed dating session.

Logility has a supply chain management solution called Logility Voyager Solutions which is internet-based.  Given the global nature of their client’s businesses, they have built in multinational support.  The costs and prices are given not only in the currency of the items ‘home market’ but also in local and regional currencies.  With this built into the system, it helps users build rollups to greater levels of detail for their inventory, production and transportation plans worldwide.

RockySoft Corporation has the Inventory Management Suite with Demand Manager and Requirements Planner, aiding clients in reducing inventory.  The suite also includes S&OP and Economic Order Manager (EOM).  With these tools, clients are able to work with the full supply chain to determine forecasts, procurement needs and replenishment quantities. Using this software also allows practitioners to take advantage of price breaks and volume discounts and also use the suite as a support tool to make decisions on a management level for inventory valuation and performance monitoring.   One key feature with the EOM tool is that you can easily compare annual costs of inventory with the annual cost of ordering based on varying volumes.  The suite is easy to use and training on the new system can be done in only four hours.  RockySoft’s applications are comprehensive but not complex.

Another vendor is working to optimize the time it takes to make demand forecasts.  OM Partners USA has  Abhi Patel at the show sharing information on their supply chain planning software.  Their core strength comes with the ability to integrate the forecast with S&OP planning and scheduling.  The company has a variety of suites that peel time down from a 4-week to possibly one or two week cycle.

A lot was learned by walking around and visiting with the vendors during the Best Practices Conference.  At times, and I know this because I have exhibited at a fair number of shows myself, attendees are not necessarily jumping at the chance to come talk to vendors.  Being on the other-other side of things this time around working as an ambassador and live-tweeting and blogging about the event though, I found that the folks exhibiting at the show were just truly excited about the new things their companies are doing.  So many new applications are being developed in this space and it is a real energizing time in the field.  So next time you are at a show, take some time to hear what’s new in the industry.  Visit with the vendors and simply ask, ‘what’s new?’  It just may be the best way to see what’s next.

Maria E. Simos is CEO of e-forecasting.com, an economic research and consulting company based in Durham, NH with clients ranging from media, academics, federal banks, major manufacturers to other consulting firms.  In her role, Ms. Simos works to further develop the reach of e-forecasting’s economic data and reporting capabilities. She also works closely with clients to ensure that they are receiving the important forecasts, economic data and support needed to be successful. She promotes the work of e-forecasting.com and provides economic analysis through her twitter account (@mesimos) and via other social media outlets.  Ms. Simos holds a Master’s Degree in Management from Carnegie Mellon University where she focused her research on management and network analysis. Her research explored social and business networks and their tie in to culture in organizations.  Her undergraduate study was completed at the Tepper School of Business at Carnegie Mellon.

Maria Simos CEO e-forecasting.com

When you come to a conference, all you want is to talk to as many people as you can so you can learn what everyone else is doing and learn from them.  Speaking with one attendee, they shared how in their group, there is only four demand planners, spread across the globe.  The benefit of attending events like the IBF Best Practices Conference is that you are now in a room with hundreds in the same position.  But still, how to talk to as many as possible?  How about some speed dating!  So that’s what about 100 or so attendees chose to do with the late afternoon session at the conference.  We were broken up into several different groups each with a different topic to discuss. We worked our way around the room a few times so that we would have a chance to discuss the topics we were most interested in. Each topic was led by table monitors.  With so many great topics to choose from, it was a hard decision, but  I chose to head  over to the JDA led table first and listened in on the discussion about “Improving forecasting and planning with consumption (POS) and syndicated data” led by Danny Halim.

The group quickly began sharing how they are all ‘trying’ to use consumption data in their demand forecast.  Digging deeper into the reason for the repeated use of the word “trying”, several issues came up regarding the reliability of  POS data.  Several fellow date-ees shared their systems for cleaning the data, or merging several different sources to make it more comprehensive.   Some companies manually merge them together, while others have built complex systems to forecast inventory levels based on POS data provided by major retailers such as Walmart. It is essential to find a system that works to clean the bad data in order to make it usable. The phrase ‘garbage in, garbage out’ was used frequently although I would say these daters were real pros at sharing and I found myself  not  wanting to get up and move on to another table as the discussion was really exciting.  I hope to see some break-out sessions on this topic at future IBF events.

I headed over to the next table where the topic was “what forecasting system works best for you”?  All daters were sharing at first was whether  their organization goes top bottom or bottom up.  Not to be left out, there were also a few working from the middle out.  The demand forecasting world does not discriminate and accepts all creeds! Two daters shared that their group does both (bottom up and top bottom) and then reconcile the forecasts.  Another shared how they begin with a price line item forecast then dollarize it by working with the marketing team and use this to drive the financial forecast.  To do this, they create assumptions upfront on the industry, start with a baseline and use this target for demand planning functions.  Not everyone  shared happy stories of their forecasting process.  One person explained that their sales department does not participate in providing input into the forecast, event though they are the closest to the demand and the customers..  The ideas presented in the Keynote Presentation by Gerry Fay of Avnet EM Velocity came up as ways to combat some of the issues the table faced. Some of these were demand sensing and responding, and command and control.  A few other different approaches presented were forecasting at the SKU level then weighting forecasts by different group functions at certain levels depending on the timing, conflicts with upper management when they do not see what they want in the forecast, forecast ownership and having the sales team provide forecast as a change in trend, rather than level.  Again, when table switching was called for, it was hard to leave but by this point everyone was so warmed up to sharing I looked forward to seeing how the last table would go.

Our last group date was led by Mike Gilliand of SAS and we talked about “new product forecasting”.  Around the table the range of new products spanned  from 5-40%.  Forecasting by analogy is the method most commonly used for new product forecasting.  The focus on the higher levels of uncertainty and risk were brought up, and the strong need to make sure management realizes this as new products roll out.  Also, the importance of tracking past new product forecast reports was part of the discussion as well.  Is your sales team consistently over-shooting? Keep this in mind.  One major takeaway was to make sure and track what assumptions were used when you are making the forecast.  If you carefully track these, it will assist in making the forecast better and help the team in the long run.  The general consensus of the table was that it takes roughly six months, for the most part, to know if a new product is going to succeed before entering it into the standard S&OP process for the organization.

And with that, speed dating was done and all minds were racing.  The level of sharing within the group continued to grow and we all moseyed over to the cocktail reception where the speed dating conversations continued and mini-crab cakes, succulent ripe California strawberries and JDA signature martinis were our award for being such great daters and sharers of demand planning lessons.

Maria E. Simos is CEO of e-forecasting.com, an economic research and consulting company based in Durham, NH with clients ranging from media, academics, federal banks, major manufacturers to other consulting firms.  In her role, Ms. Simos works to further develop the reach of e-forecasting’s economic data and reporting capabilities. She also works closely with clients to ensure that they are receiving the important forecasts, economic data and support needed to be successful. She promotes the work of e-forecasting.com and provides economic analysis through her twitter account (@mesimos) and via other social media outlets.  Ms. Simos holds a Master’s Degree in Management from Carnegie Mellon University where she focused her research on management and network analysis. Her research explored social and business networks and their tie in to culture in organizations.  Her undergraduate study was completed at the Tepper School of Business at Carnegie Mellon.

Jonathon Karelse

Demand planning and forecasting is a process which, because of its intrinsic focus on “error”, too often invites misguided efforts at producing perfection.  Like the alchemists’ quest for gold, the goal is illusory and may cost more than any benefit it yields.  Consider that the result of demand planning is only profitable to the extent that it is actionable – that is, if links above or below in the participant’s supply chain are unable to respond to the data, it might be a purely academic exercise.  Consider further that if the participant’s supply is already near-optimal, the benefit of any marginal incremental gains might be offset by the disproportionate effort required to effect those gains.

The idea of diminishing marginal returns is one which is easily understood as a concept, but too many forecasting and planning practitioners are anathema to their raison d’etre within an organization.  And the mandate for forecasting perfection is often created by senior management; the very group who should be most aware of the cost/benefit of their organization’s activities.  Any case for imperfect forecasting, therefore, must ultimately be underpinned by an incremental benefit.

The process that we undertook at Yokohama Tire (Canada) was as follows:

1. Clearly identify the problem and demarcate the extent to which its resolution is actionable.
2. Estimate the cost/benefit of its resolution and clearly apply a financial metric.
3. Evaluate judgment inputs as a means to that end. By making use of existing business intelligence and ability which may not have been leveraged fully in the process, we precluded the need for additional investment in software, consultants or headcount.

During my upcoming presentation session at IBF’s San Francisco event at the end of April 2010, I will share the means by which we addressed each step of this process and, I hope, allow participants to learn from and avoid some of the mistakes we made along the way.

Of course, we would enjoy hearing your lessons learned on improving forecasting accuracy here.  Looking forward to continuing a dialogue.

Jonathon Karelse
National Marketing Manager
Yokohama Tire

See JONATHON KARELSE from YOKOHAMA TIRE Speak in San Francisco at IBF’s:

$895 USD for Conference Only!

April 28-30, 2010 (3 Days)
San Francisco, California USA

Michael Anderson

John Becker

Monday, 8:05am: your boss calls you into his office and says, “I just got back from the Supply Chain Forecasting & Planning Conference.  We need to start a S&OP process!”  He hands you a 9 inch thick textbook and barks “get it going!!.”

So you dutifully read the textbook, gather your team, collate the data, and have your first S&OP meeting.  It is a resounding and utter . . . failure.  “This data isn’t correct, the sales are off by $1.35!, that’s 0.002%”  “What do I care about OEM capacity, I have a trade show next week!”  “I don’t have time for another meeting, I have work to do!”  Sound familiar?  It does to us, as we lived this nightmare twice.

However, we lived to tell this tale, and currently have a solid but basic S&OP process in place at Fender Musical Instrument Corporation.  Furthermore, we are on the path to implement a full-fledged S&OP process across our organization including subsidiaries and over-seas offices.

In Phoenix at the IBF Conference on Feb 22-23, we will offer some insights into what to do when the best laid plans fail; how we learned from our mistakes, changed tactics, and ultimately got on the right track.  We will discuss how we won over a management team who had lost faith in our data, a sales organization who were indifferent to our efforts, and a marketing team who didn’t realize that S&OP could help them in their effort to create the “Spirit of Rock n’ Roll.”  Please bring questions and your own war stories of successes and failures from the S&OP front.

Of course, we welcome hearing your S&OP challenges here too!

John Becker
VP of Global Supply Chain Planning
Fender Musical Instruments Corporation

Mike Anderson
Manager of Global Demand Forecasting
Fender Musical Instruments Corporation

See JOHN BECKER & MICHAEL ANDERSON from FENDER Speak in Phoenix at IBF’S:

$695 USD for Conference Only!

February 22-23, 2010
Phoenix, Arizona USA

Mike Gilliland

Below details Questions & Answers from IBF’s Webinar “What Management Must Know About Forecasting.”  If you missed it, no worries.  You can view it complimentary by clicking HERE.

1. If a product is not forecastable, what’s the most appropriate step to move the product to become forecastable?

Answer: The most effective way to improve forecast accuracy is to “make the demand forecastable” and a great way to do that is to lower the volatility of demand.  Most of what we do with our organizational policies and practices is to add volatility.  We encourage our customers to buy in spikes, and we encourage our sales people to sell that way.  This is completely contrary to quality management practices, which are all about removing volatility and making everything more stable and predictable.

Review sales and financial practices that are encouraging volatility, and either re-engineer or eliminate them and replace with practices that encourage everything to operate more smoothly.  (Examples of practices that encourage volatility are pricing and promotional activities, and the quarter end “hockey stick” to meet short term revenue goals.)  You should question whether these sorts of practice make sense by contributing to the long term profitability of your business.  If not, pursue ways to reduce volatility and encourage smooth, stable growth.  This will allow you to forecast more accurately and will reduce overall costs, which you can then pass along to your customers.

2. All of this is relative to the base line forecast, correct? What if your items are heavily promotional driven?

Answer: The accuracy of a naïve forecasting model serves as the baseline against which the performance of alternative forecasting methods should be compared.  Thus if the naïve model (say, a moving average) achieves MAPE of 40%, then I want to know how well my statistical model is forecasting, and how well my overall process is forecasting, and compare them to the baseline of 40% MAPE that the naïve model delivered.  This is what I’m talking about as a “baseline.”

This should not be confused with what is commonly called the “baseline” forecast when you try to distinguish baseline demand from promoted demand.  How do you know what demand was baseline and what was due to the promotion?  How do you distinguish the two?  I don’t believe that you can distinguish the baseline demand from promoted demand in a clean or easy or certain manner, so I would suggest not bothering trying to do so.  What matters is “how much total demand is there going to be.”  It isn’t necessary for me to care how much of it is “baseline” and how much of it is due to “promotion” – and I can never know for sure anyway?  Don’t assume you are making your forecasts  more accurate by trying to distinguish the two – you may just be making things more complex.

3. What is FVA?  A tool?  Expert judgment?  Or what?

Answer: Forecast Value Added is a metric, defined as the change in a forecasting performance metric (such as MAPE, forecast accuracy, or bias), that can be attributed to a particular step or participant in the forecasting process.  When a process step or participant makes the forecast more accurate or less biased, they are “adding value.”  FVA is negative when the step or participant is just making the forecast worse.  FVA analysis is the method of reviewing the performance of your process and identifying those non-value adding (or negative-value adding) activities that should be eliminated.  For more information on FVA analysis, see the webinar “Forecast Value Added Analysis: Step-by-Step” or the accompanying white paper.  You are also encouraged to attend the IBF Supply Chain Forecasting conference in Phoenix, February 22-23, 2010, to learn how to do FVA and hear case studies about several organizations (such as Intel) that are using this method.

4. What are the methods used commonly to measure Forecast Accuracy?  (Is MAPE the most common?) And what is a good process to determine forecast accuracy?

Answer: Mean Absolute Percent Error (MAPE) or its variations like Weighted MAPE or Symmetric MAPE seem to be the most popular metrics of forecasting performance.  MAPE has many well known limitations (such as being undefined when the denominator (the Actual demand) is zero), and is not suitable for use with data with a lot of zeroes (intermittent demand).  Also note that with MAPE you can have absolute errors greater than 100%, so you cannot simply define forecast accuracy as 100% – MAPE.

For management reporting I use a “Forecast Accuracy” (FA) metric, defined as:

1 – { Σ | Forecast – Actual |  /  Σ Maximum (Forecast, Actual) }

Note: FA is defined as 100% when both Forecast and Actual are zero.

By using Maximum of Forecast or Actual in the denominator, FA is always scaled between 0 and 100%, so it is very easy for management to understand.  That is why I favor it, even though some professional forecasters are very critical of this metric.

5. What are your perspectives on how do you differentiate volatile demand from uncertain demand?  In my opnion, uncertainty is related to an event and volatility is related to demand fluctuations. Is that right?

Answer: Volatility is expressed by the Coefficient of Variation (CV), which is the standard deviation divided by the mean.  For example, look at the last 52 weeks of sales, and compute the CV of that pattern.  In general, the more volatile (i.e. erratic and variable) the demand, the more difficult it is to forecast accurately.  Recall the Accuracy vs. Volatility scatterplot in the webinar.

Sometimes we can forecast volatile demand quite accurately, where there is structure to the volatile pattern.  You might see this for highly seasonal items, where you can always count on a big spike in demand at a certain time.  (E.g. bunny costumes and egg painting kits before Easter.)  Note: I’m not claiming we can forecast bunny costume or egg painting kits accurately, just using them as an illustration of volatility due to seasonality.

Volatility is measured looking back at what really happened.  If we expect high volatility to continue, we would probably have less confidence or certainty in our future forecasts.  If volatility is very low, we can probably feel more secure (and certain) of our forecasts.

6. Is there any ratio to determine the horizon for the forecast to be measured?  Any industry correlation to lead times?

Answer: Forecasting performance should be reported relative to the supply lead times.  Thus, if it takes 3 months to make changes in your supply, you should measure the accuracy of your forecasts made 3 months in advance. Once inside this lead time, it is ok to continue to make adjustments to the forecast, and many companies even report their forecast accuracy based on a forecast immediately prior to the period being forecast.  (Some companies even allow adjustments to the forecast within the time period (e.g. week or month) being forecast – and then report that as their forecast accuracy.)  However, it is the forecast made at the lead time that really tells you how well (or how poorly) you understand your business.  Don’t congratulate yourself on good forecasts made within the month being forecast!
Regarding forecasting horizon – how far into the future you should forecast – this will vary based on your business needs.  A power company forecasts years (even decades) ahead to know if it will need to make capital investments in new power plants.  For most companies, forecasting 12-18 months ahead is sufficient.  And the forecasting process should always be “rolling,” so that you always maintain that horizon of forecasts ahead of you.

Routinely doing 5-year ahead forecasts if you don’t really need them seems like a silly exercise.  If management insists on forecasting farther ahead than you really need, don’t waste much time doing it.  It is very unlikely you can forecast very accurately that far ahead.  It is much better to keep your organization nimble and able to adapt to however your market changes over time, rather than fool yourself into thinking you can accurately predict that far into the future.

7. How can you do calculate “appropriateness for forecasting” when your time series is too short for out-of-sample testing?

Answer: When there is enough data, out-of-sample testing is a great way to help evaluate and select forecasting models.  Good software, such as SAS Forecast Server, allows you to define and utilize a holdout sample in your model selection process.  Poorly designed software will select a model based solely on “best fit” to recent history, and as was illustrated in the webinar, the best fitting model may be a very poor choice for generating forecasts.

When there is not enough history to use a holdout sample, the appropriateness of a model is based on the judgment, experience, and domain expertise of the forecaster.  In the webinar example, Model 4 fit the history perfectly, but the forecast exploded to huge values which probably weren’t realistic (unless you had domain knowledge that demand would be significantly increasing, you were rolling out to new regions, etc.).  Without any other information, using the mean (Model 1) or a simple trendline (Model 2) seemed to be “most appropriate.”

8. Statistical modeling can be difficult in planning service parts demand. Can you give further input for planning service demand volatility.

Answer: Demand for service parts if often intermittent, with lots of periods of zero demand.  Intermittent demand is difficult to forecast accurately.  Although there are various methods to help you forecast and manage inventory in these situations (see Croston’s method and its variations), you should not have high expectations for accuracy.  It may be easier (and just about as effective) to simply forecast the mean demand each period.

Sometimes there is sufficiently high demand for the service parts that you could use standard time series methods to forecast.  It may be helpful to incorporate known sales of the items requiring the parts, so you can base your forecasts on failure rates.  Thus, if you know 100,000 units of a product were sold, and that 10% require servicing every year, this could help you decide that about 10,000 of the service parts will be needed each year.

One other approach, more applicable to high value machinery (e.g. jet engines, ships, factory production lines), is knowledge of the routine maintenance schedule.  If you sell 1000 jet engines and the maintenance schedule says a part is replaced every 6 months, then you can use this to forecast demand for that part.

9. Do you have examples available of cost of inaccuracy metrics?

Answer: I do not have access to the Cost of Inaccuracy metric used at Yokohama Tire Canada by Jonathon Karelse.  However, Jonathan will be speaking at the IBF’s Demand Planning & Forecasting: Best Practices Conference in San Francisco (April 28-30), so you could follow up with him there.

IBF members have access to a cost of inaccuracy spreadsheet available on their website.  Also, analyst firm AMR has published research (which you could access if you are an AMR subscriber) on the costs of forecast inaccuracy.
Any such cost calculators are based on a number of assumptions which you provide, so be cautious in your use of them and in your interpretation of the results.  Personally, I’m very skeptical of claims such as “Reducing forecast error 1% will reduce your inventory costs x%.” If nobody in your organization trusts your forecasts now, reducing the error by 1% is not going to make anybody more trusting of the forecast, and they won’t change any behavior, so you won’t reduce inventory.  It may take more substantial improvement to reap the cost benefits.

10. Does anyone work in the call or contact center environment for an inbound customer service center?

Answer: These principles can be applied to forecasting demand for services, such as in forecasting needs for call center staffing.  The major difference is the time bucket that is of interest.  Call centers often forecast in 15 or 30 minutes increments (rather than in weeks or months for a manufacturer), to make sure they are sufficiently staffed during peak call periods, and not overstaffed during the low call times.

Michael Gilliland
Product Marketing Manager, SAS
IBF Board of Advisor

See MICHAEL GILLILAND & EMILY RODRIGUEZ from INTEL Speak in Phoenix at IBF’S:

$695 USD for Conference Only!

February 22-23, 2010
Phoenix, Arizona USA

Martin Joseph

I remember the day very clearly when I read Donald J Wheeler’s book “Understanding Variation – The Key to Managing Chaos.” What grabbed my attention was the sterility of the traditional monthly report and, furthermore, the potential for it to be truly misleading.

Let’s consider what is usually provided to managers prior to and, worse, at meetings designed to help them steer the business. The S&OP is one example of such a meeting. Do any of the following describe typical reports in your business? If so, read on!

• Tabular rather than graphic
• Binary comparisons with plan and/or budget
• % difference between current period/month with last period/month
• Units which are not the “common currency” of those attending
• Little or no narrative
• No historical or future context

Few people can absorb real meaning from data presented in tabular format – it is difficult to process and tends to focus the reader on the largest numbers, whether they are positive or negative, important or trivial. Appropriate graphics showing historical trends gives considerably better insight, but even better when overlaid with the forecast and plan.

The inherent bias in the review process created by reviewing tabular data sets is then compounded by making comparisons with plan or with budget. A comparison of sales numbers for, say, September 2009 with the budget created from forecasts made in September 2008, adjusted and phased politically seems to me to be rather pointless! Additionally, using percentage movement rather than absolute change can, as we all know, be very misleading.

Thinking about the S&OP; data is often presented on last month’s sales volumes in SKU’s, inventory in weeks cover, working capital, line capacity…….as a Marketing Manager, I’ve just nodded off and will find a good reason not to attend next month! The challenge is to present the same core data in the units understood by all attendees: $, €, £, or ¥ for Finance and Marketing and volume, SKU’s, weeks cover etc. for Manufacturing folks.

However, the main issue here is “context.” How much more informative would the data be if it was presented in a way which showed historical trend, future trend, historical variation, the forecast, the trend through the forecast numbers, the gap between current trend and plan/budget and the required performance to close any gap? Clear graphical presentation of exceptions is essential to focus the meeting and decision-making process.

Supporting documentation should include a narrative covering commercial intelligence and explanations for historical and forecast changes in trend or variation and, where available for any exceptions. This narrative should also include commentary concerning the tactics used to close any gaps and likelihood of success.

If any of these ideas resonate with you, please attend my session at the IBF’s Supply Chain Forecasting & Planning Conference: Europe in London 1st-2nd February 2010 – see you there!

Martin Joseph
IBF Board of Advisor
Managing Director
Rivershill Consultancy Ltd. (United Kingdom)

See MARTIN JOSEPH Speak in London at IBF’S:

$899 USD | £549 GBP for Conference Only!

1-2 February 2010
London, United Kingdom

Mark Frost

D. Reilly

The “time value of money” is at stake when you are trying to forecast demand at ATM machines and of course, customer satisfaction. Trying to get the right amount of cash for pay day and holidays requires some pretty complicated models to get this right.  The reality is that these methods and approaches of forecasting daily cash demand are just as necessary when forecasting what is perceived to be “simpler” problems.  The S&OP process has treated the importance of a baseline forecast as just a “stepping stone”.  A big reason why the S&OP process is leaned on so heavily is baseline forecasts are often generated using a simplistic model that doesn’t capture patterns into a model, but rather fits a pre-specified model to the data.  A quality baseline model and forecast can alleviate a lot of the work downstream. Another comment about adjusting forecasts is that if it is for a reoccurring reason it can be added as a causal variable to the model and utilized “in-line” or also “in-model”.

When building a forecasting model, it’s important to recognize how variables like “day of the week”, “week of the year”, “day of the month”,  and holidays can capture the swings in demand and allow you to plan for them.  It’s not just the holidays, but the days before and after the holidays that need special consideration as demand ebbs and flows around these events.

Furthermore, we often hear “I am fed up with fixing forecasts”.  This can be alleviated by taking a more rigorous approach to identifying patterns rather than have some list of 50 models to be forced onto a data-set “hoping for the best” without any care for what patterns are occurring in the data.  The “one size fits all” modeling approach by taking 50 models and forcing them on a data-set is like fitting a square peg in a round hold.  Customized suits are exactly that.  Custom suits are yes custom and they take a little work requiring the expense of a tailor, but you have a proper fitting product at the end of the process. “One size fits all” can result in a hat that just doesn’t fit your head as we have seen!

Join us at IBF’s Supply Chain Forecasting & Planning Conference in Phoenix to further discuss the above.  Plus, our discussions will also cover “motherhood and apple pie,” what you need to know to make better decisions about what makes a good baseline forecast.

Your comments and feedback are welcome here!

Mark Frost
Director of Business Strategy and Decision Science
Fiserv

David Reilly
Sr. Vice President
Automatic Forecasting Systems

See MARK FROST & DAVID REILLY Speak in Phoenix at IBF’S:

$695 (USD) for Conference Only!

February 22-23, 2010
Phoenix, Arizona USA