© 2006 Oracle Corporation – Proprietary and Confidential.

© 2006 Oracle Corporation – Proprietary and Confidential

<Insert Picture Here>

Sense, Shape and Respond to Demand

John BermudezSenor Director, SCM Product Strategy


The following is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions.The development, release, and timing of any features or functionality described for Oracle’s products remains at the sole discretion of Oracle.


Demantra Demand Management

Enables real time demand sensing and shaping …

Sense demand real-time

Improve forecast accuracy

Shape demand for profitability

Evolve to real-time sales and operations planning


Improve Forecast Accuracy

• Engine supports all statistical models - Complexity is hidden for casual users but can be fine-tuned by statisticians

• Self-tuning engine

• Better than best-fit model with unlimited causal factors

• Use key accuracy metrics– MPE, MAPE, WMAPE (Weighted

MAPE)

Self-Tuning Forecasting Engine

PhD in a box: advanced statistical models

Built-in key accuracy metrics

Designed for planners, not programmers (“PhD in a box”)



• Mixed models used in same time series adjust for multiple causal factors including seasonality, market trends, and promotions

• Each model contributes different forecast characteristic to the overall model

• Automatic model selection provides improved accuracy of “best fit” approaches

– One forecast based on multiple models instead of only using best-fit model

– Self-tuning engine

• Forecast trees automatically find level with statistically relevant data

– Forecasts stored at lowest level– Proportion rules applied when necessary

• Can incorporate external information such as weather, market drivers, forward indicators, and competitive data

Causal Analysis

Outlier Detection

Promotion Events

Seasonality

Cyclical Patterns

Trend

Historical data

Bayesian Estimator

Forecast

Multiple causal factors

Combined model

Bayesian Optimizer

Leverage Advanced Forecasting and Demand Modeling option


Automatic Short/Long Term Forecast AccuracyMixed models automatically adapt in single, precise forecast

0102030405060708090

Wee

k 1

Wee

k 3

Wee

k 5

Wee

k 7

Wee

k 9

Wee

k 11

Wee

k 13

Forecast

Actual

0102030405060708090

Wee

k 1

Wee

k 3

Wee

k 5

Wee

k 7

Wee

k 9

Wee

k 11

Wee

k 13

Forecast

Actual

Model 2 Model 4 Model 2 Model 11


MAPE – Mean Absolute

Percent Error


MPE - Forecast Bias WMAPE – Weighted MAPE by

revenue, cost, and so on

Built-in accuracy metrics


The Difference is in the Details

• Demand history and causal factors maintained at lowest level

• Coefficients calculated and maintained at the lowest level for which there is demand history

• Forecasts and promotion predictions reflect local, regional, product group customer, time period, sensitivity, and so on

• Demantra approach yields most granular analysis of demand for more accurate forecasts

Product/Group

ItemItem Item

Item at a Customer

Item at a Customer/Ship-to A

Item at a Customer/Ship-to B

Demand history Promotion LiftSeasonalityCannibalizationOther Causal Factors

Demand history Promotion LiftSeasonalityCannibalizationOther Causal Factors

Per time period (week/day/month)

Granular causal factors and coefficients









Shape Demand for Profitability

• New products present forecasting challenges– Limited or no demand history for a given item

– May combine characteristics of several previous products

– Price points, changing market conditions may be different

– Product demand changes over product life cycle

Chaining

New Product C = 30% Product A + 75% Product B

Shape Modeling

•Apply shapes, scaled for volume and time

•Re-scale base on initial demand data

Attribute-Based Forecasting

Model new item based on past behavior of other items with similar attributes

Accurately forecast demand for new products based on existing data



Derive forecast for new product and adjust forecast based on actual demand

Automatically detect outliers

View demand of comparable products based on characteristics

Accurately forecast demand for new products based on existing data



• What incremental volume will result from a marketing program?

• How will it impact the sales of other products?• How does a marketing program at a brand or

product family level impact a specific item?• What were the indirect effects such as

cannibalization and consumer stockpiling?• What is the ROI on my marketing and trade

spending?• What is the predicted impact of future activity?• How does a promotion impact shipments and DC

replenishments?

0

500

1000

1500

2000

2500

3000

3500

4000

Period 1 Period 6 Period 11

Cannibalization

Pre- and post-effects

Competitive switching

Category growth

Baseline

Actual

0

500

1000

1500

2000

2500

3000

3500

4000

Period 1 Period 6 Period 11

Ca

se

s

Cannibalization

Pre- and post-effects

Competitive switching

Category growth

Baseline

Actual

Past Future

Leverage Advanced Forecasting and Modeling to understand the real impact of promotions and sales incentives



• Baseline versus incremental volume– Provides decomposition of incremental volume from advertising,

promotions, or sales incentives

• Granular lift analytics – Incremental volume lift coefficients maintained at lowest level– Localized promotion analysis– Allows shipments/replenishments to be adjusted by ship-to location

• Cross product and customer effects– Determines cross-product cannibalization impact – Adjusts forecasts for product and customer cannibalization

• Configure system for assumption based forecasting

– Structured tracking and categorization of forecast adjustment reasons, such as market and geopolitical changes

– Evaluation of impact on demand as driver for future forecasts– Examples

Semicon: forecast based on chip design wins High-Tech: forecast based on probability of winning an opportunity Life Sciences: long-term forecast based on drug approval and

patent regulations

Baseline

Incremental

Competitiveswitching

Long termgrowth

Pre and Postpromotion

effect

Cross productcannibalization

Typical Demantra AFDM

Baseline

Leverage Advanced Forecasting and Modeling to understand the real impact of promotions and sales incentives









Evolve to Real-Time S&OP

Customers

Suppliers

Develop predictable business plans

Shape demand to meet financial goals

Align supply chain to support plan

Monitor performance to plan

Shape demand to close gaps

Drive decisions into ERP

Finance

Supply Chain Sales

Marketing

Profitably balance supply, demand, and budgets



• Seeded templates for S&OP collaboration

– Seeded consensus planning worksheets– Easily tailored to your business

• Configurable and extensible– Collaborate at any level

• Manage by exception instantly– Exceptions and visual cues easily point to important

issues immediately– Document all assumptions with a complete audit trail of

decisions taken– Leverage POS data to alert planners to exceptions real

time (versus batch month by month)

• Integrated– Seeded data streams for data commonly used in the

process

Profitably balance supply, demand, and budgets



• Continuous real-time process– Simulate demand and supply strategies– Analyze multiple business scenarios– Achieve consensus on plans through internal

collaboration– Generate and analyze exceptions– Use workflow to automate process

• Adaptable – Custom demand and supply streams– Multi-dimensional– Extensible hierarchies and dimensions– User-defined reports and exceptions

Highly interactive simulation and analysis


Inputs display from entire collaboration

group – Finance, Marketing, Operations,

Key Customers and Suppliers

Integrated approval workflow process

Each S&OP participant has a configurable role-

based view

Review historical accuracy for each

input



Achieve Incremental Value

Demand Management

E-Business Suite Advanced Planning

Inventory Optimization

StrategicNetwork

Optimization

AdvancedSupply Chain

Planning

PredictiveTrade Planning

and Optimization

Real-TimeSales and

Operations Planning

Collaborative Planning

ProductionScheduling

Demand variabilityand forecast error

Demand scenarios

Promotional lift anddecomposition

Range forecast andforecast accuracy

• Shipment history• Booking history• Sales forecast• Manufacturing forecast• Items and categories• Organizations• Customers• Calendars• UOM and currency conversions• Price lists• Hierarchies (product family and

product category, time, ship from, geography, customer, demand class, sales channel)

Demand scenarios and priority

Implement additional Advanced Planning components quickly by leveraging the same foundation



Customers


Company• Leading manufacturer of medical devices• Manufactures more than 25,000 SKUs and operating from 25 plants and office around the

world$1 billion in revenues, operating worldwide

Planning problem solved• Aligned customer demand with supply chain planning

Unique aspects of implementation• Integrated with JD Edwards ERP• Supports the companies many new product introductions with improved forecast accuracy

• Increased forecast accuracy by 5-10%• Increased customer service levels• Reduced inventory by 8%• Enabled a comprehensive S&OP process

Live on Demand Management

DeRoyal Industries


Company• Leading producer of juices and jams

Planning problem solved• Promotion planning synchronized with demand planning

Unique aspects of implementation• Sales reps drive forecasting process from trade promotion planning process • What-if scenario planning enables sales reps to test promotion before selecting it

• Increased forecast accuracy at SKU level • Enables trade promotion planning to be integrated with RT S&OP• Reduced supply chain costs• Improved HQ and sales planning productivity

Live on Demand Management, Trade Management

Welch’s


Company• $1 billion in revenues, operating worldwide• Manufacturing plants in China• Leading provider of cordless phones and electronic children’s toys

Planning problem solved• Real-time S&OP process driven by one demand number

Unique aspects of implementation• Generates forecasts from customer POS data • Compares customer and generated forecasts and routes exceptions to planner, sales

representative, or customer

• Increased order fill rate from 55% to 95%• Increased inventory turns by 100%• Reduced price protection claims by 40%• Reduced logistics costs by 65%

Live on Demand Management, Real-Time S&OP

VTech



Summary


Oracle Demantra Demand Management

• Sense demand real-time– Sense demand more frequently, closer to the point of consumption– Capture demand and forecast at more granular level (store, shelf,

attributes, product characteristics)– Achieve consensus demand number more quickly by involving all

constituents at the same time, including customers– Quickly identify and react to demand changes and exceptions

• Improve forecast accuracy– Leverage advanced statistics for more accurate demand number– Use any combination of quantitative or qualitative data to establish your

base line forecast– High precision statistical forecasting, no statistical background required –

Superior Bayesian-Markov forecast analytics– Forecast based on attributes and characteristics– Leverage Advanced Forecast Modeling for promotion lift decomposition

and causal analysis

• Shape demand for profitability– Plan new product introductions– Plan promotions and sales incentives– Identify cross selling opportunities

• Evolve to real-time S&OP– Profitably balance supply, demand, and budgets

Shipments

Marketing

forecast

Order

history

CHANNEL DATA

Customer

sales

CollaborationWorkbench

Demand Hub and Seeded Worksheets

Real-time demand sensing and collaborative consensus forecasting


Oracle Demantra Demand Management

Eliminatespreadsheets

Manage rolling forecasts

Collaborate with all constituentson one number

Use basic statistics, alerts, andseeded worksheets

Tailor worksheets for individual users

Leverage POS and channel data

Forecast new product introductions

Collaborate with customers

Use advanced statistics and causal factors

Complex alerts and custom worksheets

Forecast based on attributes and product characteristics

Compute promotional lifts and analyze impact of demand drivers

Assumption based forecasting

From less complex to best in class





Leverage POS and channel data

Forecast new product introductions

Collaborate with customers

Use advanced statistics and causal factors

Complex alerts and custom worksheets





Start anywhere

Evolve at your own pace to a best-in-class solution


Proven, scalable demand management solution

Integrated demand, supply, and sales and operations planning

Designed for planners, not programmers

Integrated performance management

Why Oracle !


AQ&Q U E S T I O N SA N S W E R S


© 2006 Oracle Corporation – Proprietary and Confidential.

Documents

confidential shape demand

forecast accuracy shape

oracle corporation proprietary

confidential slide

granular analysis of

sense demand realtime

demand modeling option

details demand history