SAS® Demand Forecasting for Retail
Outperform the competition and improve business results with SAS’
next-generation demand forecasting platform
To thrive in the 21st century, retail
What does SAS® Demand organizations must strive to satisfy the
Forecasting for Retail do? unique demands of every customer. • Highly Accurate Forecasts – SAS
SAS Demand Forecasting for Retail uses Gone are the good old days of the provides the most robust library of
SAS software’s unique high-performance models in the retail software industry
mass market, when a single assort-
forecasting engine to automatically
ment, standard pricing and a single that generate forecasts based on
diagnose, model, execute and reconcile
forecasts across multiple merchandise “average location” forecast would underlying trend, seasonality promo-
levels and locations. It allows creation of satisfy consumer demand in all stores. tions, inventory effects and other
forecast projects and manages forecasts Today, with so many factors affecting known causal factors. Proprietary
with a built-in repository and versioning consumer behavior, predicting demand “best pick” functionality automates
system. SAS Demand Forecasting for
for a single item in a single store can be model selection.
Retail produces results at any level of the
product and location hierarchies, down to a difficult proposition. Forecasting de- • Drives Business Results – By seam-
the SKU/store level. mand for all stock-keeping units (SKUs) lessly integrating with existing planning
across all stores and all geographies is processes, SAS Demand Forecasting
Why is SAS® Demand Forecasting
for Retail important? a greater challenge by far. for Retail helps identify opportunities
Accurate, consistent forecasts are a key in- and improve profitability.
put to financial planning, assortment plan- This increased complexity must be
• Softlines Ready – Includes models
ning, allocation, replenishment and store addressed in an ever-more demanding
for fashion forecasting, short lifecycle
labor planning processes. With a common business environment. Dynamics such
demand forecasting platform, SAS delivers products and intermittent demand;
as globalization, increasing customer
consistent, accurate and detailed aggregate leverages SAS Merchandise
expectations, high-efficiency supply
forecasts. Improved forecast accuracy Assortment Planning for end-to-end
and consistency empower better decision chain models and diversification of
fashion forecasting workflow.
making, allocation and replenishment, and sales channels have made the retail
reduce stock-outs and manual intervention • Highly Scalable – Supports the
industry intensely competitive.
in the supply chain – the results of which entire organization down to the indi-
drive both top- and bottom-line growth. vidual store SKU level, for all types
This competition, in conjunction with
of retailers. By exploiting a parallel
For whom is SAS® Demand economic conditions, has increased
Forecasting for Retail designed? execution design with a high degree
pressure on profit margins, forcing re-
Demand forecasts are leveraged in of automation, SAS delivers reliable
tailers to focus on better understanding
preseason and in-season planning activities performance in the most demanding
the demand for merchandise through-
by merchandise buying, planning and environments.
out its retail life cycle. The result is a
allocation teams, as well as dedicated
need for better forecasting and more • Grid Ready – Operability with
forecast analysts and promotion managers.
Of course, understanding of consumer reliable methods for managing demand. SAS Grid Computing and SAS Grid
demand is critical to processes enter- Ultimately, the goal is better planning – Manager ensures that you can easily
prisewide, from financial planning through for fewer stock-outs, improved margins leverage grid technology to add more
the supply chain. SAS forecasts supply the and more solid overall business results. computing horsepower and reduce
business accuracy that enables nearly
every decision maker in your organization to
drive competitive advantage. Effective forecasting in this environ- • Reduced Workload – Automated
ment requires sophisticated capabili- exception-based forecast manage-
ties delivered via a highly automated ment reduces the need for manual
solution. inputs and forecast updates.
SAS has the most robust library of
The Bottom Line: Reducing Product Details
forecasting models for a wide range of
Forecast Errors Saves Money Agile Retail Technology retail business situations. This library
Forecast errors can result in significant Statistically sound demand forecasts includes intermittent demand models,
costs to retailers. Potential sales losses serve as the foundation for all levels of unobserved components models,
can range from 20 to 25 percent, as a your business, from merchandise plan- ARIMAX models, dynamic regression,
result of stores selling out of a product ning to replenishment. SAS Demand exponential smoothing models with
or not stocking enough of a popular Forecasting for Retail provides the optimized parameters, and user-de-
brand, color or style. But with accurate foresight your organization needs to fined custom models. For the forecast
forecasts, you can reap every sales op- survive in an increasingly consumer- analyst, SAS provides a quick and easy
portunity without wasting inventory or driven economy. Forecasts can be way to manage forecasting through the
other resources – which translates into generated for any dimension of your new SAS Forecast Studio graphical
great competitive advantage, especially business (SKU, store, region, etc.), user interface.
in challenging times. appropriate for the user and need; the
SAS forecasting solution adapts to Adaptable Forecasting Execution Workflow
The costs of poor forecasting include:
your business. When deployed with Our demand forecasting solution
Over-forecasting: SAS Integrated Merchandise Planning offers a detailed and flexible forecast-
• Excess inventory and obsolescence and SAS Merchandise Assortment ing workflow. It offers the capability of
• Higher inventory holding costs Planning, it enables planners to build defining the analyst control points and
• Reduced inventory turns merchandise, financial and assortment parameterization for this best-practice
plans that better anticipate consumer workflow, supporting all forecast
• Unexpected transshipment costs
demand. Additionally, when coupled execution scenarios and configura-
• Lower margins with SAS Merchandise Assortment tions. This is critical for analysts who
Planning, the powerful forecast man- need to adapt forecasts to individual
agement functionality lets retailers products within the model and across
• Costs for order expediting
reconcile demand in a centralized the forecast horizon. This solution
• Higher product costs platform. also leverages the workflow from SAS
• Lost sales from out-of-stocks Merchandise Assortment Planning for
Built on High-Performance preseason forecasting of new fashion
• Lost companion product sales
or short lifecycle items.
• Customer satisfaction issues
Built on the SAS Forecast Server, SAS
A truly comprehensive forecasting solu- Demand Forecasting for Retail auto- SAS Demand Forecasting for Retail
tion provides an accurate picture of key matically chooses the best model and follows an integrated approach to IT,
demand signals to all levels of the retail optimizes its parameters to produce the working with SAS Merchandise Plan-
organization. An integrated workflow forecasts. The solution can generate a ning or an external replenishment sys-
is required to provide a viable demand starting-point forecast as a means of tem to automate the forecast execution
forecast without embroiling users in seeding financial plans, as well as help process. This process, which includes
detailed statistical forecasting. In short, with in-season re-forecasting so that data preparation, diagnosis, model se-
SAS Demand Forecasting for Retail retailers can better react to sales. SAS lection, fit/forecast, automated evalua-
delivers on all of these challenges. Forecast Server reduces the overall tion and reconciliation, can be executed
workload by rapidly generating multiple for each distinct level of merchandise
What if you could develop accurate, statistically based forecasts without the and location as well as for each distinct
preseason plans based on expected need for human intervention, but also KPI. Analysts can choose to skip one or
customer demand and other statisti- provides the capability to override gen- more of the steps in this process either
cally generated demand insights, using erated forecasts in special instances. while working on a forecast or during
large-scale, automated forecasting the batch execution. Users can also
that reduces errors, increases effi- define unique performance metrics to
ciency and improves your bottom line? suit specific business needs.
With SAS Demand Forecasting for
Retail, these capabilities are possible
in a common platform.
Designed to Meet Large-Scale
Forecasting Needs Key Features
SAS® Demand Forecasting for Retail
The focus is on speed, without sacrific-
• A stand-alone solution for any retail forecasting need, such as replenishment, store labor
ing accuracy. The solution automatically
scheduling and supply chain operations.
adjusts forecasts to changing demand • Native integration with SAS Integrated Merchandise Planning solution provides financial,
and quantity patterns – easily managing assortment and allocation planners:
seasonality, including difficult-to-predict • Direct access and management of demand forecast versions.
demand for slow-moving goods. To do • Use of dynamic time set for automatic adjustment of the forecast time frame based on
this, SAS Demand Forecasting for Re- a number of periods after current time period.
tail provides a pre-defined framework • Easy use of merchandise and location filters, clusters and time events.
for demand forecasting data models • Access to forecast accuracy reports and metrics.
that ideally suit the retail industry. This • Segmentation of SKUs, based on sales patterns and other variables, to identify individual
framework stores data for points of SKUs that can be addressed one at a time where similar forecasting strategies are applicable.
• Pre-forecasting analysis to determine the seasonality of SKUs and relevant historical data.
sale, events, promotions, price, weather
• Statistical forecast generation at any level in the merchandise (SKU, subcategory, category,
and other effect parameters that can
etc.), time (day, week, month, quarter and season) and location (retail outlet, postal code, city,
have an impact on demand. Combined
state region, country) hierarchies.
with the solution’s industry-leading • Demand forecasts that incorporate the base forecast, seasonality, trends, promotions and
analytics, the framework ensures events that predictably affect demand.
super-efficient demand forecasting. As • Introduce new products, define “like item(s)” and manage demand signals with integrated
a result, SAS Demand Forecasting for merchandise planning workflow.
Retail has proven extreme scalability in • Decomposition of sales to clearly see the historical demand patterns of merchandise through
highly complex benchmark scenarios, seasonal component, trend component and irregular component analysis. Helps investigate
running multiple concurrent forecast trends and seasonal variation in sales.
jobs at retail enterprise levels. • Support for top-down, bottom-up and middle-out forecast reconciliation.
• Control of the models and workflow at a forecast job level for specific merchandise, location
levels or external criteria, as well as for a particular KPI.
Parallel processing capabilities deliver
• SAS Forecast Studio provides exception-based forecasting to simulate and monitor any un-
even greater assurance of high per-
usual parameters that are input in manual mode.
formance in a large-scale forecasting
• Ability to specify flexible effect parameters that influence forecasts. For example: event, price,
environment. SAS Demand Forecast- weather and promotions, including halo/cannibalization, pantry and delay promotional effects.
ing for Retail automatically separates • Ability to allow defined KPIs to be executed within a forecast job. For example: sales units,
workloads with millions of time series sales dollars, inventory count, etc.
forecasts into smaller jobs that are • Model revisit: saving the model in a repository for subsequent forecast runs. This helps
executed in parallel and then merged analysts review the soundness of a particular forecast model based on changes in business
back together. By using SAS/CON- conditions.
NECT® to drive these jobs, the system • In-season planning: forecast revisit and progressive weeks re-forecast based on new history,
can seamlessly process forecasts on each week.
remote and local servers. Thus you can • Archive forecasts. Ability to archive a configurable number of prior forecasts so that forecast
accuracy reports can be generated.
easily add hardware to meet forecast-
• Configurable standard forecast reports, including options for plots and statistical measures. Re-
ing needs as they grow.
ports may be saved as HTML, PDF or Excel documents and are accessible from SAS Merchandise
• Supports multiple country formats and database configurations for dates, currency and language.
• Includes effective error-handling mechanism to report any issues occurring within the forecast-
• Operability with SAS grid technology allows you to add more computing horsepower while
keeping hardware costs low.
Comprehensive Retail Intelligence SAS® Demand Forecasting for Retail About SAS
SAS Demand Forecasting for Retail is SAS is the leader in business analytics
just one part of SAS’ vision for com- SAS® Forecast Server software and services, and the largest
plete retail intelligence. Retailers can SAS® BI Server independent vendor in the business
employ a wide range of powerful SAS SAS/ACCESS® intelligence market. Through innovative
solutions, all built atop a common retail solutions delivered within an integrated
business intelligence platform, to allow framework, SAS helps customers at
for rapid integration and lower total cost SAS/SHARE® more than 45,000 sites improve per-
of ownership. These solutions include formance and deliver value by making
Technical Requirements better decisions faster. For nearly three
nearly three decades of SAS retail ana-
lytics experience and merchandising Client environment decades, SAS has been giving retail
best practices. No other vendor offers • Microsoft Windows (x86-32): organizations THE POWER TO KNOW ® .
such a comprehensive, integrated set Windows XP Professional www.sas.com
of solutions for retailers.
• Internet Explorer
SAS is an associate member of the
• SAS Customer Intelligence — Server environment
National Retail Federation.
Identify, acquire, activate, serve and SAS servers, including Base SAS and
retain profitable customers. SAS Metadata Server, can be installed
• SAS Merchandise Intelligence — on one or more hardware systems in a
Drive revenue, protect margins multitier configuration.
and earn customer loyalty with • AIX (64-bit), Release 5.1+
optimized merchandise plans,
• Solaris (64-bit), Version 8, 9 or 10
assortments, demand forecasts,
pricing, promotions, space plans
• Microsoft Windows (x86-32):
Windows Server 2003
• SAS Operational Intelligence —
Leverage organizational assets to Web tier
serve customers and trade with • SAS includes a reference
vendors efficiently and profitably. implementation of Apache Tomcat.
• SAS for Performance Management
— Analyze, forecast and maximize
profits across the entire value chain
and monitor performance toward
SAS Institute Inc. World Headquarters +1 919 677 8000
To contact your local SAS office, please visit: www.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 © 2009, SAS Institute Inc.
All rights reserved. 103199_537991.0509