Sap retail forecasting and replenishment innovation overview
Assessing the Value of Demand Forecasting in Your Organization · forecasting end – although this...
Transcript of Assessing the Value of Demand Forecasting in Your Organization · forecasting end – although this...
WHITE PAPER
Assessing the Value of Demand Forecasting in Your OrganizationA Case Study
SAS White Paper
Table of Contents
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
The SVA Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Company Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Functional Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Forecasting Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Major stated issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
System Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Performance Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Road Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Return-On-Investment (ROI) Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
About SAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
The primary content provider for this white paper was Peter Dillman, an Advisory Analytical
Consultant at SAS. He can be reached at [email protected].
SAS is strongly indebted to Dr. Mark A. Moon, Dr. John T. Mentzer, Dr. Carlo D. Smith and Dr.
Kenneth B. Kahn for their contributions in defining the methodology that serves as the basis for
the forecasting assessment process.
1
Assessing the Value of Demand Forecasting in Your Organization
AbstractThe introduction of statistical forecasting rigor into corporate processes carries many challenges and complexities from business, systemic, architectural, cultural and political perspectives . Any change, no matter how beneficial in the long run, initiates a degree of uncertainty and potential resistance . Recommendations on the judicious use of statistical software need to be tempered with a thorough assessment of organizational readiness that uncovers the specific problems, gauges the applicability and benefits of enhanced forecasting, and establishes a viable road map for remediation .
The following paper provides an example of a strategic value assessment (SVA) that aligns the processes of a fictitious product manufacturing company (ManuCo) to a road map for forecast improvement . The paper’s objective is to enable you to cut through organizational complexity by providing a structured assessment approach that ultimately leads to improved statistical forecasting processes .
The SVA ProcessThe objective of this SVA is to provide ManuCo with a detailed analysis on its current supply chain analytical process, assess areas for improvement, state degrees of readiness in making these improvements and provide a road map toward a phased solution in accordance with ManuCo’s stated objectives .
The SVA process consists of the following steps, where SAS will:
1 . Conduct detailed interviews with ManuCo personnel to gauge their individual objectives within the context of team goals and to prioritize areas for remediation .
2 . Hold joint discussions with SAS and the ManuCo team during an extended meeting for purposes of understanding the ManuCo structure, aligning expectations, exploring solution paths and product/service education .
3 . Develop a four-stage “readiness assessment” that covers four primary functional areas within the ManuCo supply chain structure as they pertain to improvement of the forecasting process, namely:
I. Functional integration – Measures the degree of interdepartmental cooperation and alignment .
II. Forecasting approach – Measures the degree of statistical rigor and collaboration within the current forecasting process .
III. Systems integration – Measures the ability of the IT infrastructure to support the forecasting process .
2
SAS White Paper
IV. Performance measurement – Measures the extent to which calculations of forecast accuracy are aligned with key performance indicators on an enterprise level .
Each of these four primary functional areas is assessed for the appropriate development “stage” within the context of the following progression:
Rudimentary --> Evolving --> Improving --> Best practices
The sole purpose of this classification is to substantiate recommendations about which areas are receptive to improvement and the functions that could provide the greatest return . It is not meant to convey a judgment of any sort .
4 . Revisit high-level ManuCo objectives and assess the ability to meet those objectives within a specified time period .
5 . Restate major impediments toward achieving those objectives in the current structure .
6 . Construct a solution road map that identifies a path toward achievement of short- and long-term objectives .
7 . If possible, construct the foundation for a return-on-investment (ROI) model that could be used as the foundation for a business case .
All of the above process steps will be addressed in this document .
Participants
SAS would like to thank participants primarily from the materials management division, as well as one of their primary outsourcers/suppliers, for their time and for revealing critical and honest information about their current processes. SAS met with and interviewed representatives from the United States and Canada divisions covering a variety of tasks and functions that support these divisions – specifically in forecasting, information systems support, demand planning, metrics analysis, materials management, inventory planning, process support and fulfillment. We feel that the resulting portrait presents a true representation of the current process along with individual and team objectives for improvement.
3
Assessing the Value of Demand Forecasting in Your Organization
Company BackgroundTo establish a reference point for the relationship between the materials management division and ManuCo as a whole, it is prudent to provide a brief statement on ManuCo’s business, product line and objectives .
ManuCo is a leading provider of technologies that serve a broad global user base, including established telecommunications carriers, government entities and individual corporate customers . ManuCo delivers innovative technology solutions that encompass a variety of products and services .
ManuCo’s business units are subdivided into two divisions – United States and Canada – and include product and services components . Efforts have been under way since the early part of this decade to consolidate and streamline functions and processes for purposes of developing a unified interdepartmental alignment structure that will serve as a benchmark for improvements in efficiency, stability and improvement with customer service, and a platform for subsequent growth . In a highly competitive and volatile industry, ManuCo has sustained long-term relationships with well-established customers, kept pace with (and driven) technological advancements and positioned itself for new business .
The materials management division is organized in support of these initiatives .
Functional IntegrationThe following classification will be used when evaluating functional integration:
Stage 1
Rudimentary
Stage 2
Evolving
Stage 3
Improving
Stage 4
Best Practices
Major disconnects between departments
Formal meetings held Communication Functional integration
Multiple forecasting Forecasting isolated to one area
True consensus Integrated forecasts
No accountability for accuracy
“Dominated” consensus meetings
Forecast champion Separate forecasting function
Performance rewards for all people involved in consensus
Multidimensional performance rewards
Figure 1: Functional integration stages
4
SAS White Paper
The following diagram depicts our understanding of the ManuCo functional flow:
CustomersProduct linemanagement
Marketing & sales Suppliers
Product, materials management
Demand planning
Other management
Ful�llment Field support
Metric & inventory
management
Finance
ResponsivenessMaterial inquiriesSVC escalation
Supplier/outsourcer
Financial & material forecast
Consumption mgt.Sup. interface
Escalation mgt.Change control
Complex solutions
Order release mgt.Source validation
Prioritization
Installation & returnsCredit mgt.
Interdivision coordination
Department
Internal customer
External customer
Materials management
Figure 2: ManuCo functional flow
Background
The US and Canada flows are handled similarly with slight degrees of variation .
US products use established technology primarily for large businesses and are predominantly managed in the United States . Products are configured to specification and are subject to change due to technological advances/obsolescence as well as upgrades and retrofits . The history of these products is well-established, and long-term contractual relations with key, critical customers are the rule . Product orders are placed at “assembly levels” – which represent unique configurations of core components .
Roughly 120 assemblies exist covering the configuration of more than 2,000 components . Approximately 1,000 SKUs are actively sold . Two thousand units are sold per quarter with a typical three-month breakdown of 400/600/1,000 .
Roughly 10 percent of the products are either preconfigured or ordered from OEMs .
Approximately 35 percent of all products are ordered not as part of a configuration, but “shipped loose” individually for purposes of retrofitting or upgrading existing systems .
5
Assessing the Value of Demand Forecasting in Your Organization
Assembly and installation are, in many cases, outsourced along with the maintenance of component stock in accordance with anticipated production and sales levels . OutSourceCo is a key outsourced integrator and manages much of this function for this product line . It translates 30-, 60- and 90-day sales projections into components based on established plan of material (POM) documents .
Products produced in Canada are not as expansive, complex or volatile as the US product base, and in many cases they have streamlined configurations (“drivers”) that consist of a set combination of assemblies with common equipment . Historical maintenance of these driver configurations enables the planning process at that level .
CanadaOutCo is a primary supplier within the Canadian division .
Assessment
The process appears to have good integration among various departments on the forecasting end – although this is not a Collaborative Planning, Forecasting and Replenishment (CPFR) environment in its truest sense . Marketing and sales seem to have the predominant input into initial forecasts, although inputs from materials management and downstream suppliers are factored in as well . Additionally, there appears to be judgmental input from the demand planning team in driving forecast organization and hand-off activities . It is not known whether true customer feedback (internal/external) and technological advances are considered when driving forecast consensus . The forecast “champion” seems to serve as a point of integration for collecting information and releasing the final forecasts .
The inclusion of a metrics and inventory management leader who interfaces with all departments is an indicator of desired process alignment . From a forecasting perspective, the process seems to have multiple hand-off points with forecasting analysis isolated in one or two areas and not driven necessarily by interdepartmental consensus .
We would classify this environment as Evolving with a move toward Improving . There is good interdepartmental activity, and forecasting appears to be isolated in one specific area; however, there is room for improvement in building interdepartmental consensus by promoting:
• A forecasting feedback loop to sales to improve the collaboration process – identifying some trends and causal factors that may improve quarterly plan forecasts .
• A methodology for integrating assembly POM histories into the forecasting process, which will serve as a vehicle for alignment with engineering divisions and supplier inventory projections .
• The examination of order histories to track high-priority and escalated orders against forecast accuracy measurements .
6
SAS White Paper
• The examination of potential factors, including contract activity, returns reported from field service, technological advances or obsolescence as potential influencers that drive forecast consensus . This examination could be a monthly or quarterly review process that would help provide some degree of statistical rigor .
• Ahistoricaldatabaseofmetricscoveringforecastaccuracy.
Some of these steps may be under way, but the true goal should be interdepartmental accountability .
Forecasting ApproachThe following classification will be used when evaluating forecasting approach:
Stage 1 Rudimentary
Stage 2 Evolving
Stage 3 Improving
Stage 4 Best Practices
Plan-driven forecasts Bottom-up forecasting Top-down and bottom-up forecasting
Top-down and bottom-up reconciliation
Forecast shipments only
Forecasting invoices/orders
Forecasting POS demand
CPFR
“Black Box” Examine promotion and seasonality
Increased FC sophistication
Segmented FC
All products treated the same
Limited seasonal training
FC drives business plan
Top management support
FC and business plan reconciliation
Figure 3: Forecasting approach stages
7
Assessing the Value of Demand Forecasting in Your Organization
The following diagram depicts our understanding of the ManuCo forecasting process . Note that we will focus on the US division with an understanding that Canada may have slight deviations .
Regional order history(product line)
Prob/Sale (h/m/l)
Capital budgets
Timing/contracts
MarketShare
Data
ManuCo Process
Integrator Process
Assembly
Forecastassembly orders
(product line x)
Consolidateacross regionsproduct lines
Regional order history(product line)
Prob/Sale (h/m/l)
Capital budgets
Timing/contracts
Ship loose
Regional order history(product line)
Consolidateacross regionsproduct lines
Regional order history(product line)
Capital budgets
Capital budgets
Capital budgets
Regional order history(product line)
Figure 4: US forecasting flow
8
SAS White Paper
Background
Forecast generation and release are two of the primary responsibilities of the demand planning group, which aggregates and modifies forecasts using inputs from internal customers (product line managers, marketing, sales plans), historical demand figures, contract analysis, supplier capacities, materials management and other factors .
The basic US process is as follows:
1 . The process starts with regional account teams holding monthly revenue calls by product line . Sales enters the probability of sale (high, medium or low) when classifying commitment levels .
2 . Quarterly timing issues are key (such as the hockey stick effect at the end of each quarter) and are considered .
3 . Market share, growth potential and capital budget expenses are considered .
4 . Projections are done for major accounts and on a regional basis for an aggregate of smaller, independent companies .
5 . Reusability of materials is considered when gauging the impact of technological advances or obsolescence .
6 . Basic smoothing and moving average models are used for most of these product lines . MAPE (Mean Absolute Percent Error) is computed and 30-, 60- and 90-day forecasts are generated at the order level .
7 . All regions and product lines are consolidated into assembly-level forecasts . Some adjustments are performed prior to release to the integrator . “Ship loose” is a term used to classify components that are not associated directly with a complex order but represent items that augment or replace components of an existing installation .
8 . Integrator reviews forecasts and explodes to components using POM to gauge inventory requirements .
Canada uses a similar process initiation approach based on sales plans . A major difference is that forecasts are provided at the product driver level – roughly one step below the assembly level . Unit-level decomposition is possible prior to forecasting in Canada, but procedures would need to be in place to handle the volatility of POM mappings .
Major stated issues
• There is a strong desire for process improvement within the Sales & Operations Planning (S&OP) framework from Order Placement --> Forecasting --> Inventory Balancing --> Metrics .
• There is a desire for consistency in processes across the enterprise . Certain product lines lend themselves to some form of statistical rigor more than others, but in general there is a play for forecasting, albeit in a variety of contexts .
• Forecast accuracy is strongly tied to accuracy of component deconstruction in various forms, e .g ., account revenue, assembly, item detail or product driver, and item level . Handling POM volatility is critical .
9
Assessing the Value of Demand Forecasting in Your Organization
• There is a strong need to improve collaboration with suppliers to develop item forecasts and compare them to outsourced item forecasts .
• There is a desire to improve the accuracy of ship-loose forecasting (e .g ., upgrades and retrofits), and installment-related material (IRM) and OEM products .
• There is a strong need to link customer satisfaction and forecast accuracy . Canada is struggling more with this than the United States (which is using more predictable, well-established products) .
• The US outsourcer has stated a need for more stable forecasts . This need presumably drives to a better idea of how all the unit components can be forecast more consistently for longer periods of time (due to longer lead times and configuration changes) .
Assessment
Clearly there is significant pressure on one specific area within demand planning to serve as the final arbiter for forecast generation . There is some question, however, about whether there is significant input and data availability to promote a forecasting approach other than one heavily based on data collection, organization, manual adjustments and exporting .
The consensus seems to be that Canada would benefit more than the established US products, since historical demand at lower levels can be charted more easily . There is well-stated accuracy at the revenue level for forecasting established US product movement due to long-term contracts and established plan-driven account behavior . The volatility of assembly configurations, however, works against unit-level accuracy and has a critical downstream impact for inventory shortages or overages .
Tying ship-loose demand into assembly sales using statistical clustering and other analytical techniques would help the demand stabilization issue .
Data aggregation is a big issue because so many forecasts approach one central area and need to be organized .
The desire for process improvement should not be underestimated and puts ManuCo in the Evolving category . The company does use some form of statistical forecasting and does not rely purely on plan-driven forecasts . However, some serious data collection and mapping issues work against flexibility in driving multiple reconciliation techniques as well as unit-level forecasts, which would help downstream inventory . The skills are there to embrace a more statistically driven approach . Accordingly, the major improvements should focus on:
• Exploring the use of statistical techniques using data collected from Canada and OEM/IRM sales and product movement . This could be a sound foundation for showing immediate return and would provide insight into a dynamic forecasting process that can incorporate other market/technological factors into future projections .
• Exploring the use of data aggregation techniques to automate the forecast compilation process and thus leave more time for analysis .
10
SAS White Paper
• Exploring the use of POM decomposition techniques for translating dollar/assembly forecasts into unit levels – even for the established US products . This sets up an automated mapping process that would provide insights helpful in inventory management activities .
• Building forecasting sophistication to identify events and model unobserved components that could be used to aid sales in future plan forecasts .
System InfrastructureThe following classification will be used when evaluating the system infrastructure at ManuCo:
Stage 1 Rudimentary
Stage 2 Evolving
Stage 3 Improving
Stage 4 Best Practices
Islands of analysis Interdepartmental system linkage
Client/server Open architecture
Poor integration On-screen reports Flexible user interface EDI linkages with major customers and suppliers
Poor system understanding
Periodic system- generated reports
Common database ownership
Printed reports On-demand reporting
Figure 5: System infrastructure stages
The following diagram depicts our understanding of the US system infrastructure:
Data
ManuCo Process
Integrator Process
US
Manage logistics(4PL x)
Processnoncomplex
orders
Complex orders
Ordermanagement
Engineercon�guration
Processnoncomplex
orders
PLMforecasts
Forecast/orderhistory
Consolidateforecasts
Reconcileforecasts with
inventory
Finalforecasts
Final con�gurations
Materialsdatabase
Orderful�llment
Legacy
Legacy
ERP
outsourcer
MPE legacy
FTP to outsourcer
Noncomplex,general orders
Figure 6: System-forecast integration flow
11
Assessing the Value of Demand Forecasting in Your Organization
Background
The US process for complex orders primarily consists of an order management sequence initiated and controlled through a series of home-grown mainframe and HP legacy systems .
1 . Data from the legacy system is stored in SQL using Microsoft Access . Warehouse management data is stored in Oracle .
2 . Orders and configurations are sent through these systems with final configurations transferred via FTP to the integrator .
3 . Multiple 3PLs are managed externally .
4 . Noncomplex orders are processed through SAP R/3 and sent to the integrator . The corporate movement is toward SAP integration with SAP-BW housing information .
5 . Sales forecasts from order systems are reviewed by PLMs within sales/marketing and presented to the demand planning group in the form of Excel spreadsheets (individual by PLM, region or major account) .
6 . The demand planning group maintains histories and forwards final assembly forecasts to the integrator .
Globally there are approximately seven unique flows with associated data marts .
Assessment
The clear advantage for ManuCo is the movement toward a central ERP process and repository, which puts ManuCo in the Improving category, as it will ultimately lead to greater accessibility to data needed for collaboration and historical forecasting activities . Short-term improvements, however, should focus on the following:
• Automating the forecasting aggregation process for the demand planning group .
• Incorporating a POM-decomposition process that commences generation of unit-level historical data . This should be done immediately for the Canadian division .
• Building an analytic repository of sales history along with event calendars, market intelligence and technological touch points . The repository should support an ad hoc reporting process that can be used in a multiuser collaborative environment for forecasting and reporting .
12
SAS White Paper
Performance MeasurementThe following classification will be used when evaluating performance measurement:
Stage 1 Rudimentary
Stage 2 Evolving
Stage 3 Improving
Stage 4 Best Practices
Accuracy not measured
Accuracy measured (typically MAPE)
MAPE used with focus on supply chain impact of accuracy
Recognize external factors affecting forecast accuracy
Forecasting performance not tied to measure accuracy
Evaluation based on accuracy with no consideration for implications of accurate forecasting
Performance graphed and analyzed
Treats forecast error as indication of possible problem
Other statistics are calculated and monitored
Multidimensional metrics of forecasting performance
Figure 7: Performance measurement stages
Background
Within the materials management division is the metrics and inventory management group responsible for inventory management, audit readiness and process evaluation . A clearly stated goal of this group is process improvement: more specifically, how process improvement can be aligned with evaluation and improvement of customer satisfaction . The need to improve data sharing with suppliers for purposes of enhancing forecast accuracy and inventory projections was also mentioned as equally important .
Mean Absolute Percent Error (MAPE) is used in some capacity to measure forecast accuracy . The following figures were shared:
Forecast accuracy: 65-75 percent at component level, 65-70 percent at monthly assembly level, 80 percent at quarterly level, 90 percent at the aggregate level (disregarding assembly mix), 80-90 percent accuracy at the dollar level . Forecasts are generated at 30, 60 and 90 days . Sixty percent of the codes have +/- 25 percent accuracy based on the 60-day forecast .
Customer satisfaction is measured at 75 percent for the United States and 60 percent for Canada, which indicates a pressure point in Canada that requires immediate attention .
Component code-level forecast accuracy from the integrator is at 75 percent for roughly 60 percent of the code universe .
Assessment
There is a strong desire to drive toward a unified metrics platform and an organizational infrastructure that can support performance measurement; however, the lack of true data integration and uniformity for forecast accuracy measurements and improvement analysis places ManuCo in the Evolving category . Clearly the organization would benefit from a balanced scorecard approach that looks at the four primary dimensions of customer satisfaction, supplier satisfaction, internal process efficiency and revenue
13
Assessing the Value of Demand Forecasting in Your Organization
attainment (thus linking the process with corporate performance) . We see this as a longer-term strategy that would help ManuCo realize the goal of using superior forecasting processes to manage inventory, profitability and, ultimately, customer satisfaction .
Steps that would improve the forecasting process would include:
• Constructing a schematic of a balanced scorecard with clearly defined metrics .
• Standardizing metrics computation across all divisions .
• Measuring the impact of improved forecasting accuracy .
• Working with IT to build a metrics repository .
• Exploring the use of an inventory stock calculating tool that could be used by integrators to help model scenarios that link customer satisfaction measures (i .e ., fill rates, backorders) with required inventory levels .
• Exploring the link between forecasting accuracy and profitability (either in cost reduction, reduced lost sales and/or increases in customer satisfaction) .
Road MapThe road map is presented here as a sequential process based on stated priorities . The recommendations from the preceding sections can be used to outline a rough “statement of work” for each priority:
Modelevents
Integrate orderhistory with
market intelligence
Link forecastaccuracy topro�tability
Tie ship loose intoassembly sales
Phase 3
Phase 2
Phase 4
Integrateassembly/driver
POM
Automateforecast
aggregation
Build a metrics repository
Explore POMdecomposition
Conductpilot/proof step
Developforecast
metrics feedback
Standardizemetrics
computation
Link customersatisfaction with
inventory
Phase 1
Figure 8: ManuCo road map
Phase 1: Improve forecasting sophistication.
Phase 2: Establish data integration and reporting.
Phase 3: Improve collaboration.
Phase 4: Execute balanced scorecard.
14
SAS White Paper
Phase 1: Gain a “quick win” that uses forecast accuracy improvements to improve customer satisfaction .
Steps would include:
• Exploring the use of statistical techniques in a pilot process that uses data collected from the United States, Canada and/or OEM/IRM orders and product movement . This step could be a sound foundation for demonstrating the viability of statistical forecasting with an immediate return and would provide insight into a dynamic forecasting process that can incorporate other market and technological factors into future projections .
• Establishing a methodology for integrating assembly or driver POM histories into the forecasting process, which would serve as a vehicle for alignment with engineering divisions and supplier inventory projections .
• Buildinginforecastingsophisticationtoidentifyeventsandmodelunobservedcomponents that could be used to aid sales in future plan forecasts .
Phase 2: Improve the demand planning forecasting process and establish a data integration and reporting process, thus alleviating the burden of repetitive tasks and enhancing a statistically rigorous approach .
Steps would include:
• Using a forecasting feedback loop to sales in order to improve the collaboration process – identifying some trends/causal factors that may improve quarterly plan forecasts .
• Establishing a metrics historical database covering forecast accuracy .
• Examining order histories to track high-priority and escalated orders against forecast accuracy measurements .
• Examining a variety of potential factors, including contract activity, returns reported from field service, technological advances/obsolescence, etc ., as potential influencers driving forecast consensus . This could be a monthly/quarterly review process that could help provide some degree of statistical rigor .
• Automating the forecasting aggregation process for the demand planning group .
• Building an analytic repository of sales history along with event calendars, market intelligence and technological touch points . The repository should support an ad hoc reporting process that can be used in a multiuser, collaborative environment for forecasting and reporting .
15
Assessing the Value of Demand Forecasting in Your Organization
Phase 3: Improving the collaboration between ManuCo and the supplier integrator through data sharing, inventory analysis and forecast alignment .
Steps would include:
• Exploring the use of an inventory stock calculating tool that could be used by integrators to help model scenarios linking customer satisfaction measures (e .g ., fill rates, backorders) with required inventory levels .
• Exploring the use of POM decomposition techniques for translating dollar or assembly forecasts into unit levels – even for the established US products . This sets up an automated mapping process that would provide insights helpful in inventory management activities .
• Tyingship-loosedemandintoassemblysalesusingstatisticalclusteringandotheranalytical techniques would help with the demand stabilization issue .
Phase 4: Building and executing a complete balanced scorecard that tracks metrics across the enterprise and aligns directly with customer satisfaction .
Steps would include:
• Constructing a schematic of a balanced scorecard with clearly defined metrics .
• Standardizing metrics computation across all divisions .
• Measuring the impact of improved forecasting accuracy .
• Working with IT to build a metrics repository .
• Exploring the link between forecasting accuracy to profitability (either in cost reduction, reduced lost sales and/or increases in customer satisfaction) .
Return-On-Investment (ROI) ModelThe following model is derived using information provided by ManuCo’s finance, marketing and materials management divisions based on figures from the past fiscal year . A high-level analysis established baselines for over/under forecast product splits as well as estimates of aggregate forecast error (measured using the MAPE statistic) . The resulting calculations produce an “impact of forecasting errors” section that quantifies its effect . Based on an assessment of ManuCo’s forecasting environment, a 15 percent improvement in forecast accuracy is a plausible conservative estimate, with the potential to increase in Year Two as techniques are refined . Allowing for 5 percent growth, the net two-year benefit reflects roughly $6,000,000 in savings due to reductions in inventory costs, expedites and lost sales .
16
SAS White Paper
Figure 9: Return-on-investment model
SummaryIn conclusion, the materials management division is clearly in line with organizational objectives focusing on interdepartmental alignment, streamlining the system infrastructure and driving a metrics-driven approach to forecasting and inventory management . Despite being in a somewhat transitional, evolving stage, ManuCo, with SAS’ help, identified areas for immediate improvement that involved working within the existing infrastructure to promote a more statistically rigorous forecasting process as the initial part of a four-phase plan . It is hoped that this plan will culminate in the attainment of a primary long-range goal: an organizationally aligned scorecard that uses customer satisfaction as one of its cornerstones of performance assessment .
17
Assessing the Value of Demand Forecasting in Your Organization
ConclusionThis paper has outlined the use of a strategic value assessment (SVA) process in assessing forecast readiness . This paper should provide guidance in defining the critical issues and contributing factors that would help companies build road maps toward superior statistical forecasting processes .
About SASSAS is the leader in business analytics software and services, and the largest independent vendor in the business intelligence market . Through innovative solutions, SAS helps customers at more than 60,000 sites improve performance and deliver value by making better decisions faster . Since 1976 SAS has been giving customers around the world THE POWER TO KNOW® .
SAS customers receive a full suite of support services at no extra charge, including skilled telephone technical support and unlimited, around-the-clock online technical support . Our online customer support center provides always-on access to a wealth of technical support, reference information, educational resources and communities . Knowledge-sharing is continuously available through regular seminars, webcasts and an expansive selection of training courses .
SAS’ record of revenue growth in every year of our existence not only makes us a stable business partner, but has enabled us to reinvest more than 20 percent of revenues in R&D each year – so we can continually improve our products .
ReferencesJain, C .L . “How to Measure the Cost of a Forecast Error .” Journal of Business Forecasting . Winter 2003-4: 2, 29-30 .
Kahn, Kenneth B . “How to Measure the Impact of a Forecast Error on an Enterprise .” Journal of Business Forecasting, Spring 2003, 21-25 .
Mentzer, J .T . “Sales Forecasting Audit Protocol .” bus.utk.edu/forecasting/articles/FAU%20Protocol.PDF .
Moon, Mark A ., John T . Mentzer, and Carlo D . Smith . “Conducting a Sales Forecasting Audit,” International Journal of Forecasting, 19 (2003), 5-25 .
About SASSAS is the leader in business analytics software and services, and the largest independent vendor in the business intelligence market . Through innovative solutions, SAS helps customers at more than 60,000 sites improve performance and deliver value by making better decisions faster . Since 1976 SAS has been giving customers around the world THE POWER TO KNOW® . For more information on SAS® Business Analytics software and services, visit sas.com .
SAS Institute Inc. World Headquarters +1 919 677 8000To contact your local SAS office, please visit: sas.com/offices
SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies. Copyright © 2012, SAS Institute Inc. All rights reserved. 103375_S93664_0712