One of my colleagues got in touch with someone she used to work with.  They were able to provide a private API: APPS.wip_flowUtil_priv.  My colleague tried the API call explodeRequirementsToMTI and was able to get the components to backflush.  We thought all was well.

However, upon further testing, I noticed something strange.  The original Assembly was getting stuck in the interface.  And when resubmitted, I saw both the assembly and another round of components getting transacted.  What was going on?

Further testing provided me with the answer: the originally inserted assembly record was getting slightly altered by the material transaction worker before getting marked in error.  Altered to a point which would spawn component backflush transactions if marked for processing and then processed.  I tried inserting a fresh assembly record into the MTL_TRANSACTIONS_INTERFACE table with the updated fields and was able to get the components to backflush, automatically, without any errors.

We now have a functioning R12 Work Order-less completion interface which inserts assembly completion records into the MTL_TRANSACTIONS_INTERFACE table.  Records which are correctly formed and spawn component issue (backflush) transactions when processed by the material transaction manager (Process transaction interface).

These are the two fields that were changed by the transaction manager:

OPERATION_SEQ_NUM from -1 to NULL
TRANSACTION_BATCH_SEQ from NULL to 1

Below is the insert statement we are using, with Oracle 12.1.3 changes noted.

INSERT INTO MTL_TRANSACTIONS_INTERFACE(
TRANSACTION_INTERFACE_ID,
TRANSACTION_HEADER_ID,
SOURCE_CODE,
SOURCE_LINE_ID,
SOURCE_HEADER_ID,
PROCESS_FLAG,
VALIDATION_REQUIRED,
TRANSACTION_MODE,
LOCK_FLAG,
LAST_UPDATE_DATE,
LAST_UPDATED_BY,
CREATION_DATE,
CREATED_BY,
LAST_UPDATE_LOGIN,
REQUEST_ID,
PROGRAM_APPLICATION_ID,
PROGRAM_ID,
PROGRAM_UPDATE_DATE,
INVENTORY_ITEM_ID,
ITEM_SEGMENT1,
ORGANIZATION_ID,
TRANSACTION_QUANTITY,
PRIMARY_QUANTITY,
TRANSACTION_UOM,
TRANSACTION_DATE,
SUBINVENTORY_CODE,
LOCATOR_ID,
TRANSACTION_SOURCE_ID,
TRANSACTION_SOURCE_TYPE_ID,
TRANSACTION_ACTION_ID,
TRANSACTION_TYPE_ID,
TRANSACTION_REFERENCE,
TRANSFER_SUBINVENTORY,
TRANSFER_ORGANIZATION,
ERROR_EXPLANATION,
ERROR_CODE,
ATTRIBUTE10,
ATTRIBUTE_CATEGORY,
WIP_ENTITY_TYPE,
OPERATION_SEQ_NUM, –****CHANGED****
TRANSACTION_BATCH_SEQ, –****CHANGED****
BOM_REVISION_DATE,
ROUTING_REVISION_DATE,
SCHEDULE_NUMBER,
SCHEDULED_FLAG,
FLOW_SCHEDULE)
VALUES(
MTL_MATERIAL_TRANSACTIONS_S.NEXTVAL, –TRANSACTION_INTERFACE_ID
MTL_MATERIAL_TRANSACTIONS_S.NEXTVAL, –TRANSACTION_HEADER_ID
‘BACKFLUSH TEST’, –SOURCE_CODE — Reference to legacy/interface data
-99999, – SOURCE_LINE_ID — Reference to legacy/interface data
-99999, – SOURCE_HEADER_ID — Reference to legacy/interface data
1, –PROCESS_FLAG
1, –VALIDATION_REQUIRED — 1=FULL
3, –TRANSACTION_MODE — 2=Concurrent, 3=Background
2, –LOCK_FLAG
sysdate, –LAST_UPDATE_DATE
1001, –LAST_UPDATED_BY
sysdate, –CREATION_DATE
1001, –CREATED_BY
1000001, –LAST_UPDATE_LOGIN
NULL,– REQUEST_ID
NULL,– PROGRAM_APPLICATION_ID
NULL,– PROGRAM_ID
SYSDATE,– PROGRAM_UPDATE_DATE
100001, –INVENTORY_ITEM_ID
‘ITEM_NUMBER’, –ITEM_SEGMENT1
100, –ORGANIZATION_ID
10, –TRANSACTION_QUANTITY
10, –PRIMARY_QUANTITY
‘EA’, –TRANSACTION_UOM
SYSDATE,– TRANSACTION_DATE
‘FG’, –SUBINVENTORY_CODE
10001, – LOCATOR_ID
”, –TRANSACTION_SOURCE_ID – NULL as Work Order-less
5, –TRANSACTION_SOURCE_TYPE_ID
31, –TRANSACTION_ACTION_ID
44, –TRANSACTION_TYPE_ID
‘TEST TRANSACTION ON ‘|| SYSDATE, – TRANSACTION_REFERENCE
NULL,– TRANSFER_SUBINVENTORY
NULL,– TRANSFER_ORGANIZATION
NULL,– ERROR_EXPLANATION
NULL,– ERROR_CODE
‘TEST ATTRIBUTE’, –ATTRIBUTE10
‘ORG’,– ATTRIBUTE_CATEGORY
4, –WIP_ENTITY_TYPE
NULL,– OPERATION_SEQ_NUM ****(Was:-1)****
1,– TRANSACTION_BATCH_SEQ****(Was: NULL)****
SYSDATE,– BOM_REVISION_DATE
SYSDATE,– ROUTING_REVISION_DATE
WIP_JOB_NUMBER_S.nextval, – SCHEDULE_NUMBER
2, –SCHEDULED_FLAG
‘Y’ – FLOW_SCHEDULE);

The iSetup module worked pretty well in our proof of concept testing.  But we ran into a problem when we started migrating our multi organization/business unit configurations. We have multiple business units because we are sharing the instance with another division at our company. Migrating configurations with iSetup brings all the ORGs and sites along with each configuration, however we only wanted our division’s configurations migrated.  To solve the problem, we started editing the exported XML before loading it into the new instance.

Editing XML is relatively time consuming and prone to error when editing in a text editor compared to a program designed to edit XML.  We tried to use the free Microsoft XML Notepad 2007, however, loading the exported files was causing iSetup to fail on upload.  I found the problem and solved it as described below.

Conditions

• Oracle iSetup module in Oracle E-Business suite R12
• XML Documents edited by Microsoft XML Notepad 2007

Symptoms

• iSetup reports error when uploading edited XML files
• Log file as seen here:

oracle.xml.parser.v2.XMLParseException: Start of root element expected.
at oracle.xml.parser.v2.XMLError.flushErrors1(XMLError.java:320)
at oracle.xml.parser.v2.NonValidatingParser.parseRootElement(NonValidatingParser.java:341)

at oracle.xml.parser.v2.NonValidatingParser.parseDocument(NonValidatingParser.java:303)
at oracle.xml.parser.v2.XMLParser.parse(XMLParser.java:205)
at oracle.apps.az.r12.api.BC4JAPI.createAPIEntityItems(BC4JAPI.java:330)
at oracle.apps.az.r12.extractor.cpserver.UploaderContextImpl.validateExtract(UploaderContextImpl.java:170)
at oracle.apps.az.r12.extractor.cpserver.UploaderContextImpl.upload(UploaderContextImpl.java:105)
at oracle.apps.az.r12.extractor.cpserver.UploaderCp.runProgram(UploaderCp.java:74)
at oracle.apps.fnd.cp.request.Run.main(Run.java:157)

Solution

1. Download extraction ZIP file generated by Oracle iSetup
2. Extract ZIP file to your computer
3. Edit and save extracted XML file in Microsoft XML Notepad
4. Open saved file in Notepad (Text Editor)
5. Save As ASNI encoding
6. Replace existing file within original ZIP file
   1. ZIP file compression levels and algorithms must reamin intact
   2. 7-ZIP supports replacing files within an existing file using drag and drop functionality
7. Upload updated ZIP file to Oracle iSetup as a new extraction.
8. Load extraction into source instance

Conclusion

Grocery retailers frequently use promotional sales to lure customers into stores.  Retailers hope customers will purchase higher margin products while they shop for the promoted items.  Chicago grocers Jewel-Osco, subsidiary of SUPERVALU, and Dominick’s, subsidiary of Safeway, both follow use this marketing strategy, changing promotions twice a week.  The promotional sales may increase customer traffic, but they may also lead to the demise of the store.  I suspect, retailers that use promotional sales, on every day products sold year round, increase the cost of doing business and decrease supply chain efficiency.

Used by many grocers, promotional sales are the activities, materials, devices, and techniques used in the advertising and marketing of products.  I separated their cost into two different categories; direct costs and indirect costs.  I see the money spent on the processes and material to implement the promotional sale as direct costs, and the money lost due of the effects of the promotional sales are considered indirect costs.  From my point of view, both of these costs are significantly high, high enough to possibly outweigh any benefit of the promotional sales.  If I were managing a retail grocery operation, these are the items I would consider when deciding whether or not to continue utilizing promotional sales.

Direct Costs

Promotional sales have many components, each having a cost associated with it. Planning, marketing, inventory builds, and store preparations all take time to complete and consume valuable company resources.

Planning

Planning a promotional sale must be time consuming and costly.  Stores need to decide what products to put on sale, when to put them on sale, and at which price to sell them during the sale.  They must also decide how they will market the promotion, design the promotion, and distribute the promotion. It takes time for the retailer to manage, and it takes time for suppliers to manage.  That being said, it is possible that retailers who promote frequently can manage this process relatively efficiently.  However, managing a promotional sales still requires resources; resources that do nothing other than manage and execute promotions. The planning process is only the beginning.

Marketing

Promotions need to be communicated to the public, using any number of methods, all of which cost money.  From my experience, grocers seem to prefer some form of print media, like news papers or local ad circulars. Designing, printing, and delivering these advertisements must add significantly to a retailers cost. Some grocers even compliment the print media with radio and TV commercials, all of which take additional resources to produce and distribute.

Inventory Builds

Before the promotional sale starts, retailers must increase inventory of the items being promoted. Vendors, transporters, distribution centers, and storefronts all need to take special measures to guarantee extra inventory is delivered, on time.  Not only on time, but just in time; fresh produce spoils quickly and it’s vital the delivery process be planned and executed without error.  Early or late delivery can lead to costly spoiled inventory. In addition to any spoiled inventory losses; inventory builds take significant amounts of working capital.

Store Preparations

At the start of the promotional sale, costly store preparations must occur.  I have witnessed employees rearranging store shelving to accommodate shifts in inventory in anticipation of an increase in demand for some products, and decrease in demand for others.  Price tags throughout the store need to be updated, reprinted, and reapplied.  Finally, if displays are being used, they must be built and placed.

Each activity and consumable is temporary.  As promotions end and new ones begin, I expect these four processes repeat endlessly, adding cost during each cycle.

Indirect Costs

Along with any direct costs mentioned above, I predict promotional sales also have indirect costs aswell. Indirect costs caused by poor forecast accuracy, inadequate supply chain communication, and large inventory swings.

Forecast Accuracy

As part the planning and execution phases, I suspect most vendors and retailers use Enterprise Resource Planning systems, or ERP systems, to manage the supply chain.  ERP systems use a multitude of variables, including, but not limited to forecasts, past orders, current orders, and current inventory levels, to track, plan, and manage inventory.  Recommendations to create purchase orders or manufacturing requirements are created on a regular basis. Recommendations may be edited manually and released by planners, or  they may be released automatically as purchase orders to suppliers.

Because forecasts are the foundation of ERP system demand calculations, I know the data must be clean and accurate for the systems to operate efficiently.  But, from my point of view, promotional sales cause forecast data to be dirty or inaccurate for two reasons: lack of predictability, and manual data entry.

I assume it is nearly impossible to precisely predict how inventory will move during and after a promotional sale.  The incredibly high quantity of variables makes it close to impossible to accurately forecast purchasing. Even the most nimble and elaborate forecasting software will likely have difficulty analyzing uneven sales data, not knowing what impact each dip or spike had, or if the dip or spike will happen again. During a promotion, inventory could be wiped out completely or may not be sold at all.  After the promotion, purchasing may likely decrease but it may, instead, increase.  The dip or spike in demand will last an unknown period of time.  I suspect competitors may possibly have a promotion on the same or competing products.  The promotion price may or may not impact purchasing.  Finally, it is also likely a naturally occurring or unpredicted event may also affect demand.

Even if all the variables are correct and accounted for, and the company has perfect knowledge of consumers’ thoughts and competitors’ marketing, forecasts may still be entered or edited incorrectly.  The manual data entry processes is far from six-sigma levels of accuracy and, from my experience in analyzing forecast data, will most likely have errors.  Incorrectly entered or edited data points could flow automatically from the retailer’s ERP system to the vendor’s ERP system without notice.  It is likely systems would be in place to detect erroneous data, but because the data nature of highly volatile forecasts, I would expect the detection systems would miss many errors.

Supply Chain Communications

The ramifications of incorrectly generated and/or entered forecasts could be huge and expensive.  Data often trickles down from retailers to distributors and from distributors to suppliers; Most likely traveling automatically by way of automatic releases of ERP generated purchase order recommendations.  An invalid forecast could be transfered from the retailer to distributor to vendor without anyone knowing its inaccuracy.

Correct, but changed forecasts, could also cause problems.  I expect promotion plans change all the time.  They may get canceled, dates may move, or anticipated purchasing may change.   Regardless of the change, corrections or non-standard changes could get lost, miss communicated, or miss translated when being sent from retailers to vendors.  In some cases, they may not be communicated at all.

The farther away in the supply chain a company is from the retailer, the more difficult planning and communication becomes.  At the front of the chain, retailers have direct knowledge of a promotion, and have the ability to add or edit the details in their ERP system.  As promotions are developed, retailers communicate information to suppliers.  But, as mentioned before, if a promotion changes, it may not be fully communicated to all suppliers.  To add to the confusion, suppliers have suppliers, and those suppliers also have suppliers.  The farther down the chain, the less promotion visibility is seen.

The entire chain is only as good as the weakest link.  If not handled quickly and efficiently, at all levels of the supply chain, one invalid or missed forecast entry, or forecast change, has the ability to trigger massive product spoilage.

Inventory Swings

From my experience at a manufacturing company, huge swings in inventory are nightmares for suppliers.  Inventory builds require increased production, which may require overtime pay and/or auxiliary temporary warehouse space; both impacting the supplier’s profitability.  On the other hand, purchasing reductions that occur after a promotion ends may lead to vast amounts of unused warehouse space, non-moving inventory, or unutilized resources or personnel.  These swings are not only problems for suppliers, they are disruptive to transporters as well.

Trucking is the predominant type of food transportation in the United States.  Trucking is a tough business; just like empty planes, empty or unused trucks are not profitable. To be profitable, trucking companies need to haul loads to and from sources and destinations.  And inventory swings spurred by promotion sales create irregular shipments and increase the possibility of empty trailers for flatbed trucks.  Fortunately, third party logistics companies have become a popular way to mitigate the cost of empty trucks, but I still don’t believe they are still not a perfect solution, nor do they provide free services.

The least tangible costs of inventory swings are their affect on data mining and forecast generation.  As previously discussed, forecast accuracy is crucial.  From my experience as a Supply Planner, large swings in inventory make it very difficult to produce an accurate forecast. Standard deviations for products that regularly go on sale are extremely high, and systematically generated forecasts can be incredibly inaccurate.

Any losses or costs incurred at any point in the supply chain will be passed on to the retailer by way of higher wholesale prices, regardless who is at fault.  The retailer will then pass that cost increase to the consumer via higher retail prices throughout the store, not necessarily on the product being promoted.   Keep in mind, all of these additional costs are being incurred to sell an item at a reduced price, likely a price below wholesale cost.

I believe retailers that use promotional sales as part of their marketing strategy are pregnant with cost saving opportunities.  It only requires retailers to realize how many costs are likely associated with their promotional sales in order for them to analyze their promotional sales strategy.  In a retail world ruled by everyday low prices, the analysis will hopefully come soon.  If not, retailers selling everyday products will likely fail because their promotional sales have a high probability of increasing their cost of doing business and decreasing their supply chain efficiency.

Individual Change… An individual may be confronted with a new user interface, a new “best practice”, a different way to make decisions, or any number of new requirements. When this is the case, individuals will tend to experience one or more of the following five key factors: fear, anxiety, confusion, loss of control, or erosion of power or influence in the organization. The more an individual feels any of these emotions, the less likely they are to accept the new way of doing things. Lack of acceptance will diminish or delay the benefits of the ERP effort…

Individual Change was something I dealt with on a daily basis during the implementation at each of the plants I was involved with.  To overcome the fears, anxiety, confusion, and loss of control or power, one of the things I stressed during my training sessions was how much impact the employee had during the process.  I made them aware of the many things that occurred in the background when they processed a transaction; including raw material consumption and inventory movement.  I also explained how that transaction impacted the rest of the business processes such as costing and planning.

This additional, non essential, information, made the employee realize not only how powerful the system was, the but the importance of completing the transaction accurately and in timely manner.  A side benefit of knowing what happened as the transaction processed and what impact a transaction had on other business processes,  was the ability to think, on the fly, of how to correct a problem that occurred outside of the typical business process.

Hard Drive Imaging

Before I installed Linux I wanted to make an image of my hard drive so that I wouldn’t have to go through the painful Vista setup again.  That was a process in itself.  My final solution was getting GPart, a bootable partition manager, and running it on a USB Jump Drive.  What a wondrous solution!  By far the best experience I have had during my Oracle Instillation process.  I was able to copy my Windows Vista partition from my internal hard disk to my external USB hard disk rather simply.

Linux and USB Hard Drives

I was a bit confused when installing Linux, but I was able to get through it.  And it booted and worked without any problems.  It even found my wired network interface card (NIC) *GASP*!  So far so good!

This is where the pain started.  Oracle was on a USB drive!  At first I connected it and nothing happened.  I tried searching the net and doing the whole mount thing, and I pretty much gave up after looking at this guide on mounting a USB drive in Linux and having zero luck. I tried getting the data through “Windows Networking” on my iMac.  It worked at first, but then something happened, and it stopped?  I still don’t know how to get it working again. BUT I found the solution to my USB drive problem!

First, I don’t know why it didn’t mount before, but it mounted after a reboot, and I could see all the files I had stored on the external USB hard drive.  But, as I should expect at this point, I had a problem.  I could not execute any files on the drive for some reason.  I am not a Linux expert, I wouldn’t even say I am a novice at this point. So I copied all my files over to my system disk and started the instillation from there. To get the Oracle E-business installer going I had to use terminal to launch it.  I browsed to the path the Oracle installer was at and used this command to launch the installer once I was in the rapidwiz folder.:

./rapidwiz

Wouldn’t you know it, I failed a test!  The storage space test!  I continued anyway; failed instillation of coarse.  I reinstalled Linux to start from scratch.  I also found a solution to my problem of executing from a USB drive!

The external file system was only half mounted or something.  I went back to my mounting a USB drive in Linux guide and figured out my problem!  I was using /sda1, /sda2, or /sda3.  Those are the partitions 1, 2, and 3 on storage device A!  I wanted partition 2 on storage device B.  So I tried:

mkdir /mnt/usbdrive
mount /dev/sdb2 /mnt/usbdrive

My command processed  with out an error!  I felt invincible at this point!  I launched rapidwiz and started the instillation once again.

This time the failure was rather quick. I figured this had to do with the fact that I had already attempted an instillation. I began deleting registry entries, and Oracle created files. It was really a pain. I thought I had everything solved, and I started the instillation again! And surprise surprise; another failure.

Installing Oracle in Windows Server 2003 was not something meant for me. I think it actually had something to do with the Microsoft Visual Studio I had used in the first place, the lack of initial host/domain configurations, and the prior install. Regardless of the actual problem. I thought it may be ideal to use Oracle’s Linux OS since that is what Oracle was made for. So on to installing Oracle Enterprise Linux!

Cygwin

For Cygwin, follow Metalink Note # 414992.1. If you don’t have access to Metalink (actually they changed the name to “My Oracle Support”, Oracle likes to confuse people as much as they can), just search the internet, and you can find a copy of the details somewhere. Oracle sure didn’t make this process easy! Cygwin can’t simply be installed with default settings; many additional items have to be selected and installed. Some of the server mirrors don’t have necessary files, so you have to try different ones. Additionally, there are additional steps after the Cygwin install that must be followed in order to get Oracle to accept the Linux like environment.

Microsoft Visual Studio

One of the hardest problems to solve was getting Microsoft Visual Studio installed, or at least the components that I needed. I thought I had it covered by following the guide and installing the Visual Studio 2005 redistributable file.

I thought everything was complete, so I started the Oracle installer. I followed all the directions and got to the “testing” phase where the Oracle Installer makes sure everything is in order and working correctly. Unsurprisingly, many things failed.

The first was the tools test, which I narrowed down to Visual Studio! I looked through the errors and found some information on the internet. I ended up installing Microsoft Visual C++ 2008 Express Edition. I had to do this in order to get the “CL.exe” file that was needed later in the Oracle Instillation (as mentioned in the error log). I also had to add the folder path to the “Environment Variables” and I copied the CL.exe to the System32 directory. I don’t know if all of those were necessary, but Oracle seemed to be happy once I did that.

Host/DNS

The next thing that failed was the “host/domain” test. I tried for an hour to get that to work. I thought editing my domain’s DNS settings would solve the problem, and It never did. I figured I wasn’t going to use a domain name with my install anyway, so I continued the instillation an hour or so after I changed the DNS settings and with the test continually failing.

Two hours or so into the instillation, it failed. I looked up many different things on the internet, and found the problem could be caused by my lack of correct host/domain. To fix the problem I went back and verified my A Name was “oracle.edhayes.us” in my domain names DNS settings. I also verified it resolved to my real world IP address. After an hour or so more of Internet server prorogation, I tried the Oracle instillation test again, and still no success with the host/domain portion.

I did some more searching on the internet and found some more information. I went to system properties, computer name, more, and filled in the Domain prefix (I used oracle.edhayes.us). I tried doing some pinging in the command window, and it was pulling up the correct IP address, but the test was still failing. At this point I think the domain settings had propagated. I set my DMZ on my router to my laptop’s IP address. I dont know what combination of tricks did it, but all of the DMZ, DNS, and computer’s domain prefix got the Oracle installer host/domain test to successfully complete. Lets start the install again!

Gateway NV5214u

Laptop

I am using a Gateway NV5214u I purchased at Best Buy for $479 before tax.

• Warranty: 1 year limited
• Product Weight: 6.8 lbs.
• Processor: AMD Athlon™ X2 at 2.1 GHz
• Screen: 15.6″ WXGA (1366 x 768)
• Memory (RAM): 4GB DDR2 SDRAM
• Hard Drive: 320GB SATA
• Optical Drive: Double-layer DVD±RW/CD-RW
• Digital Reader
• Webcam
• Gigabit Ethernet (RJ-45 connector)
• Wireless Networking (B+G+N)
• HDMI
• USB 2.0 Ports: 4
• Battery Life: Up to 3 hours
• ENERGY STAR Qualified

When Installing the Windows Server 2003, I would continually get the Blue Screen of Death with these codes:

STOP: 0x0000007B (0xF789EA94, 0xC0000034, 0×00000000, 0×00000000).

I solved the problem by setting the SATA Mode to “IDE Mode” in the bios (F2 on startup).

Network Drivers

After the instillation, I found that nearly all of my drivers were not found.  I called Gateway to locate the network drivers, but Gateway does not have my model in their support system yet!  Shame on them!

Knowing that Acer now owns Gateway, I headed over to their site and found a laptop similar to mine.  I then located that laptop’s drivers in their support section.  They had several drivers listed, so I downloaded them all and used a USB drive to copy them from my Mac to the laptop. After testing each driver, I found one that worked on one of the two network devices!

The Gateway NV5214u “Network Adapter” in Device Manager is the “Ralink 802.11N Wireless LAN Card”.  Head over to Acer’s website and select an Acer Aspire 6530 Notebook. Get the Windows XP Pro 32 bit Ralink Wireless LAN Driver Version 1.1.5.0.

On to Oracle!

I will be using this website as my guide.

Here is the list of things that I have downloaded and stored on an external drive.  I suggest using a download manager; I am using DownThemAll!.  Oracle is 39 files, totaling around 40 GB

• Windows Server 2003 R2  Enterprise Edition with SP2 (32-BIT)
  • Trial registration
  • Image download
• Visual C++ 2005 Redistributable Package
  • Download Page
• Cygwin (Linux Like Environment)
  • Home Page
  • Download File
• Oracle r12
  • Oracle Download Site
  • Select E-Business Suite for Microsoft Windows 32-bit
  • Choose Oracle E-Business Suite Applications Release 12.1.1 Media Pack for Microsoft Windows (32-bit)
  • Download All Files

Off to buy a computer!  To Be Continued…

I was sent a message on LinkedIn from a gentleman requesting my input on their planning methods. As I was writing my response, I thought that it would be of benefit to other people, so I am posting the question and response here. Click “read more” below to read the entire question and response.

Question:

We both share the Supply Chain group on linkedin.I understand you have experience in Demand Planning & Forecasting. We at the IBF – Institute of Business Forecasting & Planning are currently researching what companies are doing to manage their inventory, reduce operating costs, improve customer retention & fulfillment in this volatile market? Especially, when history can no longer be an indication of future outcome. It would be great if you could share some thoughts on what your company is doing to remain competitive and preserve cash. Furthermore, do you feel executives are recognizing the important of demand planning & forecasting, more so now than ever before? And if so, are they only looking for technology as the quick fix, or improving processes from your viewpoint? Of course, we see pursuing technology without having proper processes in place to be dangerous.

So far, we’re seeing companies paying more attention to forecasting for items with higher value only, and doing it at shorter interval. We’re also seeing companies truly leveraging their S&OP processes, as well as their POS and Syndicated data to make better planning decisions from having a clearer picture of consumer behavior at any given time.

Response:

Thanks for the message. I was unfortunately laid off in December. However, I can tell you that we had a similar problem, and when I left we were working on reducing working capital. The problem we had was that the production planners at the plants were all trying to fill their production schedule so that the production line was utilized when there were crews there to staff it. As a result, inventory was extremely high and all of our plants.

We did a few things to try to solve this. First, I created a tool to analyze safety stocks using data from our order history and item level fixed days supply. Second, we would discuss the new safety stock recommendations, by item, with the plants before adding the new data to our Oracle Item setups. Finally, there was a push to drive production solely off of our manufacturing plan recommendations. Recommendations were driven off of daily customer orders, forecast, and safety stock demand.

Oracle adoption was pushed from very high up in the company. The one pending item that we were working on when I left was our legacy way of thinking of inventory targets. Prior to our Oracle implementation, we had inventory targets as a metric. These targets were Monday Morning levels of inventory. We assuming production seven days a week and shipping five days a week. So these levels were historically the highest point of our inventory and were the metric at which plants were gauged.

Post oracle, this economy, and staffing changes all impacted changes to actual inventories at plants, and it was apparent that a heuristic or best judgment way of creating these targets would not work. It was also apparent, in my mind, that it was a poor metric to use. Plants should not be graded on their levels of inventory as compared to one target: temporary order surges or lulls could make a plant look poor, when in fact it was producing perfectly against the manufacturing plan.

As I was leaving, I was in the process of creating a method of calculating inventory targets using the fore mentioned safety stocks, fixed days supply, and average daily shipments. Where the total volume of safety stock would be a “minimum” inventory, and a calculation based on fixed days supply and average daily shipments added to the safety stock would be the peak of inventory. Of coarse the peak may or may not land on a Monday like our prior metrics assumed; that was a change management issue that had yet to be discussed or resolved.

When it comes to forecasting, we were in the process of upgrading to Oracle Demantra when I left. Oracle Demantra would greatly improve our forecasting methods, and accuracy. And yes, It was apparent that accurate forecasts were an important part of our planning process. But, an accurate forecast is meaningless if you do not produce exactly what you are recommended to based off all of the data inputs including, but limited to, forecasts, safety stocks, and orders.

Hope that helps.