July 2005 AMR Research Report

Advanced Retail Replenishment builds upon tradi-

Advanced Retail Replenishment: tional replenishment approaches by leveraging detailed
Demand-Driven Inventory demand forecasts and analytics to enable the automated
determination of optimized store and DC inventory
Planning Automation Is a orders. Replenishment software vendors largely focus

Competitive Necessity on analytics rather than on the merchant; none yet

provides advanced workflow and UI functionality that
by Alexi Sarnevitz, Janet Suleski, and Joyce McGovern fully reflects a retailer’s business processes.
© Copyright 2005 by AMR Research, Inc.

No portion of this report may be reproduced in whole or in part without the prior written permission of AMR Research. Any written
materials are protected by United States copyright laws and international treaty provisions.
AMR Research offers no specific guarantee regarding the accuracy or completeness of the information presented, but the professional staff
of AMR Research makes every reasonable effort to present the most reliable information available to it and to meet or exceed any
applicable industry standards.
AMR Research is not a registered investment advisor, and it is not the intent of this document to recommend specific companies for
investment, acquisition, or other financial considerations.
Advanced Retail Replenishment:
Demand-Driven Inventory Planning
Automation Is a Competitive Necessity
by Alexi Sarnevitz, Janet Suleski, and Joyce McGovern

Advanced Retail Replenishment process adopters are achieving significant

competitive advantage by driving substantial inventory productivity
The
improvements alongside a superior customer experience from reduced Bottom
out-of-stocks. Line

Replenishment is a critical part of Advanced Retail

Planning
As outlined in the AMR Research Report “Advanced Retail Planning: Achieving
Effective Demand and Merchandise Synchronization,” January 2005, the central
objective of enterprise-wide planning is to profitably synchronize all resources, such
as inventory and space, to consumer demand signals. While the merchandise, assort-
ment, and allocation planning processes covered in that Report (see Figure 1) are used
to determine Distribution Center (DC) and store orders for all types of retail goods,
replenishment capabilities are generally focused on longer lifecycle products with pre-
dictable demand that are more amenable to a highly automated process.

Examples of replenishable items include grocery products, building materials, office

supplies, and basic clothing such as denim jeans, socks, or undergarments. Less pre-
dictable short-lifecycle products such as fashion apparel and seasonal items are typically
handled with more manual merchandise and allocation planning processes.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 1

 Figure 1: Advanced retail planning process overview

Corporate Advanced Retail Planning

Planning Buying/Assortment Vendors

Merchandise Allocation and

Consumers
Planning Replenishment

Strategy
Space

Common Demand Intelligence Platform

Functional Area Tactical Planning

Finance
Financial Planning
Advertising, Pricing,
and Promotion
Marketing Supply Chain
Distribution and Channels
Transportation
Store
Operations Labor Scheduling

Covered in ARP Report Subject of this Report

ARP Element Definition

Merchandise Planning Sales and inventory planning process across multiple dimensions
(e.g., product hierarchy, channel, store hierarchy) to at least a
style level

Buying/ Buying: Vendor order quantity and timing determination process

Assortment
Assortment: Process for determining the breadth of products to
be carried in each store; also includes local market assortment
optimization

Allocation and Store inventory order quantity and timing process

Replenishment
Allocation: Initial or subsequent “push” orders

Replenishment: Subsequent “pull” orders based on consumer

demand via an automated process

Space Allocation of store space by product category and assignment of

shelf capacity by item

Common Demand Common set of forecasting engines, databases, and rules for all plan-
Intelligence Platform ning applications

Source: AMR Research, 2005

2 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Our research shows that most retailers recognize the growing importance of advanced
replenishment and have prioritized it as a key area for future investment in new pro-
cesses and enabling technology. This has been quantified by industry surveys, too.
According to the recent 2005 National Retail Federation (NRF) Retail IT Budget
Study, 72% of retailers are planning to replace, add to, or improve their replenishment
capabilities in 2005.

Common replenishment approaches provide

inadequate accuracy and usability
Traditional replenishment approaches fall short of meeting retailer needs for a number
of reasons:

Inaccurate demand forecasts

Many legacy replenishment systems rely on simple forecasting models to project

demand. These include methodologies such as basic moving averages, the application
of chain-wide seasonality to all Stock-Keeping Units (SKUs) and stores, or the manual
estimation of promotional lift. These approaches are unable to provide repeatable
accuracy, as they do not adequately reflect the multitude of causal factors that affect
true sales levels, such as SKU/store-specific seasonality, promotional lift, price elasticity,
and underlying strategic or competitive business trends.

Inadequate DC and store order synchronization

Stores and DCs are often replenished via independent processes and systems where the
DCs are treated as a supplier to the stores and DC replenishment forecasts are based
on store shipment history. This creates a bullwhip effect that leads to stockouts and
excess inventories. While some applications now derive DC forecasts from Point-of-
Sale (POS) sales data, this alone is not enough to achieve service-level and productivity
goals, as DC and store order quantities are also driven by store-level inventory short-
ages and surpluses. Additionally, constraints such as pack sizes and store order mini-
mums must also be accounted for. At a more advanced level, significant opportunity
exists to optimize supply by modeling inventory policies, such as safety-stock and ser-
vice levels, simultaneously across all DC and store tiers to determine the most efficient
inventory distributions that will achieve store service-level objectives.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 3

 Excessive manual effort

Most legacy replenishment applications are organized around functionality, not the
actual sequence of business process activities regularly repeated by busy merchants every
day. When asked to demonstrate “a day in the life” of a replenishment analyst, most
software vendors presented a workflow that started with demand forecasting, followed
by order determination and the manual review of proposed orders (see Figure 2). This is
nearly the reverse order of the actual process steps followed by a replenishment analyst.

The flow of some current replenishment applications almost assumes a need for users to
manually review and adjust the details of each automated order generated, a task that is
virtually impossible in a retail environment with millions of store/SKU combinations.
This creates unnecessary manual effort in processes that should be highly automated.

Users are also challenged by cumbersome, nonintuitive user interfaces that force them
to navigate a complex array of pop-up screens and pull-down menus that still emulate
the look and feel of spreadsheets. These unnecessary process steps further divert user
attention from more value-added activities such as aligning merchandise execution to
business strategies or conducting more in-depth analysis of promotional events and
their influencing causal factors.

Figure 2: Typical repenishment application workflow

Generate
Forecast Demand Recommended Order Review
Orders

• Store and DC levels • Vendor and • Due orders presented to

• Automated process DC orders replenishment analyst for
• User sorts through • Automated process approval
pull-down menus, • User sorts through • Manual process
pop-up screens, or complex menus • Little or no prioritization
parameter lists to to adjust rules or decision support
adjust methodology information
• Workflow typically has
user manually override
orders perceived to be
incorrect

Source: AMR Research, 2005

4 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Advanced Retail Replenishment enables process
automation with sophisticated analytics
Advanced Retail Replenishment includes three key enhancements not found in more
traditional replenishment processes:

Accurate and adaptable forecasting algorithms

Improved forecast accuracy is a critical enabler of replenishment process automation.

Advanced replenishment systems include an array of sophisticated, science-driven
forecasting algorithms that predict granular-level trend and seasonality characteristics
with significantly improved accuracy. These systems further enhance accuracy with a
best-pick capability that automatically selects the most accurate forecasting methodol-
ogy from this array. Items often shift between predictable and highly variable sales
patterns. For example, denim jeans shift from stable volume levels to a highly variable
sales pattern during the peak back-to-school period, and basic toys shift to extremely
variable volume peaks during the holiday season.

Any promoted product can show extreme variability during the promotion period.
Retailers have historically handled these sales peaks using manual allocations that
pushed product to the stores. Systems on the forefront of advanced inventory planning
use sophisticated seasonality and causal forecasting capabilities to enable a more auto-
mated replenishment process during volatile seasonal and promotional sales periods.
At the most advanced level, multiple Advanced Retail Planning (ARP) modules are
supported by a single forecasting capability.

Synchronized DC and store replenishment models

At a tactical level, the more advanced replenishment applications base DC orders on

POS sales to minimize the bullwhip effect on inventory. Additionally, DC orders
are adjusted for store-level inventory surpluses and deficits during each replenishment
cycle. Replenishment applications often accomplish both of these by calculating DC
orders from a projected store replenishment plan.

Advanced safety stock calculations incorporate both demand and supply variability to fur-
ther drive order accuracy. To achieve full cross-tier inventory optimization, policies such as
service levels and safety stock are modeled simultaneously for stores and DCs to determine
the most efficient inventory deployment approach that achieves store service-level objectives.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 5

 Automated and exception-driven processes

In an Advanced Retail Replenishment process, sales forecasting and order determina-

tion activities are executed by the automated replenishment application. For example,
a replenishment analyst equipped with this technology would begin his or her day
reviewing prioritized system-generated exception alerts for material sales and inven-
tory variances to plan (see Figure 3). When warranted by significant variances, system
parameters are adjusted to improve future sales forecast and replenishment plan accu-
racy. Adjustments can include the selection of an alternative forecasting algorithm,
a different ordering methodology (e.g., substitute a fast-mover methodology for one
appropriate for slow-moving merchandise), or the correction of errors to sourcing
parameters such as lead time, pack size, or product dimensions.

As part of a regular order policy review, additional adjustments can be made to reflect
business changes such as new service level goals, revised vendor agreements, or alterna-
tive transportation modes. Manual intervention is only required to expedite orders
with unique logistical requirements such as promotions with unique pack sizes, holiday
events that require special free-standing displays, or potential supply network con-
straints such as a dock strike threat or looming import quota limit. All of this must
be enabled by application workflow functionality that aligns to the core business pro-
cesses, supported by User Interface (UI) elements such as adaptable dashboards and
activity guides or wizards that conduct users through each process step.

Figure 3: Advanced retail replenishment process overview

Key Activities Key Activities

• Review automatically • Confirm sourcing parameters,
generated exceptions including points of origin, lead
• Evaluate order policies times, and pack and product
regularly specifications
• Maintain system Determine • Ascertain service-level
parameter settings: Ordering requirements and inventory
Manage
- Revised sourcing Methods deployment rules
Order
agrements and • Analyze alternative
Accuracy
- Inventory policies Policies inventory policies
- Alert thresholds • Select forecasting
methodology
• Expedite priority orders • Determine optimal ordering
- Unique events methodology
- Supply constraints • Configure system parameters

Source: AMR Research, 2005

6 © 2005 AMR Research, Inc. AMR Research Report | July 2005

A critical enabler of Advanced Retail Replenishment is full visibility to time-phased
sales, inventory levels, and capacities at all levels of the supply network. This visibility
must extend to inbound and in-transit product and cover all sales channels that rely on
the network.

Grocers have just begun the journey toward advanced replenishment

While general merchandise retailers typically implement a centrally driven replenishment process,
grocers, who are just beginning the journey toward advanced replenishment, often execute a
hybrid approach, referred to as Computer-Assisted Ordering (CAO). This approach requires sig-
nificant store associate involvement as part of a gradual move away from traditional decentralized
store-driven replenishment processes.

The hybrid process sends the centrally generated replenishment system order recommenda-
tions to store personnel for review and approval. Since most replenishment orders are manually
reviewed by store personnel, this approach stays far less efficient than a fully automated and
exception-driven process. Additionally, the manual adjustments often reduce order accuracy.

Grocers wanting to accelerate the journey toward an advanced, centrally driven replenishment
process need to address a combination of application functionality limitations specific to grocery,
cultural hurdles, and weak (if not nonexistent) critical enablers such as an accurate perpetual inven-
tory. To surmount these barriers, grocers must do the following:

• Overcome the cultural resistance to science-based capabilities—Similar to the merchant

psychology that general retailers struggle with, store associates do not trust the automated system
to determine accurate order quantities. This resistance is best dealt with by proving the effective-
ness of the new capabilities. One grocer we spoke with has begun to track and compare the
order accuracy impact of each manual order adjustment to the original CAO quantity to foster the
needed trust.

• Seek out applications with truly advanced functionality—Some of this concern by asso-
ciates is warranted, as our evaluation of grocery-oriented replenishment applications revealed a
prevalence of more rudimentary functionality compared to that available for general merchandise
retailers. In response, grocers should evaluate the feasibility of adapting the more advanced gen-
eral merchandise replenishment applications for their businesses. For example, one grocery chain
is leveraging the progressive forecasting capabilities of NCR/Teradata to provide the needed
functionality.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 7

 • Ready the infrastructure—Often, grocers have not fully deployed the enabling capabilities
that provide needed visibility into store-level inventories. For example, one major European grocer
only achieved half of its anticipated inventory productivity gain because of inaccurate perpetual
inventories that drove excessive stockouts after a new store replenishment system was imple-
mented. Grocers must make needed visibility, inventory accuracy, and data integrity enhance-
ments as part of any Advanced Retail Replenishment initiative. Often the infrastructure alone can
yield significant benefits. One regional U.S. grocery chain reduced shrink by 5% and increased
employee productivity 6% by implementing a comprehensive perpetual inventory across all
categories, including those with fresh products. While a number of the replenishment systems
evaluated can handle daily store replenishment of fresh items, none provides the functionality
needed to drive fresh and prepared product production systems. Grocers seeking these capabili-
ties should consider specialized Fresh Item Management (FIM) vendors, such as Park City Group,
DCM Solutions, ADC, and Invatron Systems, which will be covered in future research.

Next-generation advanced replenishment is driving

significant financial and operational benefits—today
Leading retailers across industry segments realize significant benefits from advanced
replenishment capabilities in place today:
• One European consumer electronics chain reduced store-level stockouts by 50%
without increasing inventory via use of advanced DC and store inventory modeling.
• A national specialty superstore retailer drove a 10%-plus increase in DC service level
by using store-centric POS volume as a critical forecast variable in place of DC ship-
ment history.
• A European department store chain achieved a 20% decrease in store inventories
while in-stock positions remained steady at 95% after implementing an advanced
replenishment process.
• A U.S. specialty hard goods retailer doubled the average number of SKUs a buying
team can manage by deploying a new application that provides an exception-based
order management model.
• A regional grocery chain generated measurable sales increases by halving store-level
out-of-stocks from 6% to 3% using a new centralized replenishment process.
• A European grocery chain improved inventory productivity by 15% after implement-
ing an advanced store replenishment capability.

8 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Vendors concentrate on functionality rather than
usability
As was true with software companies providing ARP technology, no single software
vendor today delivers the complete advanced set of retail replenishment capabilities.
Most application development to date has prioritized forecasting and order determina-
tion analytics over process-oriented workflow and usability.

Our research shows that most vendors have generally focused on internal development
of functionality, with the notable exceptions of SAP and SofTechnics, which rely on
partnerships with SAF to provide their replenishment analytics. Multiple progressive
applications provide an array of forecasting methodologies that include causal capabili-
ties for promotions and best-fit functionality to select the most accurate algorithm. i2
Technologies and Oracle/Retek stand out for offering an advanced forecasting and
demand intelligence platform that feeds multiple ARP modules.

A number of progressive applications also base DC orders on the store replenishment plans
or POS sales history with subsequent adjustments for inventory shortages or surpluses at
store level. Only Evant supports the simultaneous optimization of DC and store service
levels, although Manugistics and i2 can do this with available add-on modules.

The replenishment applications reviewed generally feature a spreadsheet look and feel.
These UIs were often cumbersome to use, with a complex array of pull-down menus
and pop-up screens that users were forced to search through to find needed func-
tionality. What workflow capabilities existed were limited at best. Evant and NCR/
Teradata, which were partial exceptions to this, have focused on continually improving
the user experience, with progressive workflow capabilities for managing order accuracy
in their system. NCR/Teradata deserves special note as it combines this with a broad
installed base and a high degree of overall company strength.

This vendor review has identified that most replenishment vendors have much more
adaptation to do for merchant-oriented workflows so that their UIs genuinely reflect
important business processes (see Figure 4).

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 9

 Figure 4: Process-oriented workflows are required

Current Application Workflow Required Replenishment Workflow

Generate Determine
Forecast Demand Recommended Order Review Ordering
Manage
Orders Order
Methods
and
Accuracy
Policies

Source: AMR Research, 2005

10 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Advanced Retail Replenishment vendor comparison
Tables 1 to 3 below present an evaluation of the replenishment functionality provided
by a number of the vendors covered in this Report. To be included in these tables,
vendors needed to have retail store and DC replenishment applications in general
release for general merchandise, apparel, and/or grocery retailers. The functionality
evaluation only includes vendors that could demonstrate software that is in general
release, not planned for a future version. The research methodology also requires at
least two references to validate our findings. Vendors that met all of these criteria are
in Tables 1 and 2.

Table 1: Replenishment functionality ratings—general retail vendors

Island NCR/ TXT

Functionality Evant Pacific JustEnough Manugistics Teradata NSB Group e-solutions
DC
Replen-
ishment
Store
Replen-
ishment
Allocation *

Forecasting
and Demand
Intelligence
Merchandise/
Space Planning
Integration
Workflow
Functionality
and Usability
Alerting **
and
Reporting
Source: AMR Research, 2005

*Not available as part of the currently deployed application.

Advanced
** Available in separate application from vendor.
Progressive
Basic
Notes:
Not Evaluated
• Allocation ratings for the following vendors reflect rating from the AMR Research Report “Advanced
Retail Planning: Achieving Effective Demand and Merchandise Synchronization,” January 2005,
which may be part of a separate module: Island Pacific, i2, NSB, Oracle, SAP.

• Manugistics has a separate inventory policy optimization module not included in the ratings.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 11

 Table 2: Replenishment functionality ratings—grocery specific vendors

SofTechnics
Functionality Aldata Lawson Retalix (SAF)
DC
Replenishment
Store *
Replenishment
Allocation * **

Forecasting and Demand

Intelligence
Merchandise and Space
Planning Integration
Workflow Functionality
and Usability
Alerting and
Reporting
Store Level and Fresh
Item Replenishment
Production System No No No No
Integration
Handheld Hardware Yes Yes Yes Yes
Integration
Scale No No No No
Connectivity
Source: AMR Research, 2005
Advanced *Not available as part of the currently deployed application.
Progressive ** Available in separate application from vendor.
Basic
Not Evaluated Notes:
• Retalix rating displayed is the highest rating of DAX, BICEPS, or PowerBuy for each
category. See Appendix A for more detailed information. Retalix allocation func-
tionality requires a separate module, ABS.
• Lawson Store Replenishment Module has not yet been acquired and deployed by a
retailer.

Our research did uncover a surprising finding in that four major retail suite provid-
ers, i2, JDA, Oracle/Retek, and SAP, were unable to supply all of the requested
references. i2 only provided a single national retailer. In JDA’s case, the new .NET-
enabled Portfolio Replenishment Optimization (PRO) application is in the final test-
ing stages and not yet fully rolled out. Oracle/Retek was unable to supply anyone
from Sainsbury’s, although we have confirmed that the Retek XI Advanced Inventory
Planning replenishment application was implemented by Sainsbury’s at the store level,
but not for its DCs, meaning that the cross-tier capabilities may be unproven. SAP
could only provide one reference for its Forecasting & Replenishment application, a
European retailer piloting 200 SKUs across six stores. Functionality evaluations for
these vendors are included in Table 3.

12 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Table 3: Replenishment functionality rating—partially referenceable vendors
JDA:PRO
Functionality (.NET version) JDA:E3 i2 Oracle (Retek) SAP
DC
Replenishment
Store
Replenishment
Allocation *

Forecasting and
Demand Intelligence
Merchandise and Space
Planning Integration
Workflow Functionality
and Usability
Alerting and
Reporting
Store Level and Fresh n/a n/a n/a
Item Replenishment
Production System n/a n/a n/a No No
Integration
Handheld Hardware n/a n/a n/a No Yes
Integration
Scale n/a n/a n/a No No
Connectivity
Source: AMR Research, 2005
Advanced
Progressive * Arthur Allocation is required to provide this functionality when using PRO. Rating based on a separate
Basic evaluation of the module executed subsequent to the publishing of the aforementioned ARP Report.
Not Evaluated Based on the ARP ratings methodology, this application would have received a Progressive rating.

Notes:
• Allocation ratings for the following vendors reflect rating from the AMR Research Report “Advanced
Retail Planning: Achieving Effective Demand and Merchandise Synchronization,” January 2005, which
�����������������������
may be part of a separate module: Island Pacific, i2, NSB, Oracle, SAP.

• i2 has a separate policy inventory optimization module not included in the ratings. i2’s Master Data
Management Module is needed to provide pervasiveness exceptions necessary for the Advanced rating
for alerting and reporting.

• These four companies demonstrated products, but could not supply the full number of references
required by this research's methodology.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 13

 Vendors are hooked on funded development
The functionality and usability gaps seen throughout our comprehensive retail replen-
ishment research is symptomatic of an ongoing vendor reliance on customer-funded
development. Customer-funded development is a root cause behind many of the
observed usability issues in retail merchandise and inventory planning because it drives
vendors to deliver functionality at the lowest cost possible. The least expensive approach
to meeting customer-funded business requirements is simply layering incremental capa-
bilities (e.g., more pop-up screens or menus) on top of legacy interfaces and workflows.

This shortsighted thinking breeds complexity, forcing users to wade through a complex
puzzle of buttons, screens, and menus to find the functionality for each step of the
replenishment process. The result is poor process compliance and reduced benefits.

The right approach? Design or configure the UI workflow to match the business pro-
cesses that will be enabled. In many cases, the gaps extend beyond usability into key
functionality. One vendor currently migrating onto a new technical architecture has
eliminated important legacy functionality by tailoring the new release to beta client
requirements. In another case, a vendor is marketing a retail replenishment applica-
tion that was originally built for manufacturing or wholesale clients. The giveaways
were in UI elements where store sales were actually labeled as “planned goods issued”
and “consumption,” combined with alerting that would require users to wade through
thousands of store/SKU exceptions.

A proactive product management and investment strategy can work with retail applica-
tions. Manugistics’s recent 7.X releases of Demand and Fulfillment, applications that
have been fully rearchitected to a J2EE platform with enhanced retail-oriented func-
tionality, were funded via significant internal research and development investment.
This investment has contributed significantly to Manugistics’s sales, as retail vertical
software license revenue has grown to 33.0% of fiscal year 2005’s total from only 8.8%
in 2003.

The vendor community is beginning to recognize the advantages of this approach.

After its merger with Oracle, Retek announced that it would leverage Oracle’s financial
strength to shift from a historical dependence on funded development toward a prod-
uct management approach where capabilities are built in anticipation of market needs.
Expect other financially strong vendors to seek competitive advantage via this route.

14 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Recommendations: No silver bullet, so retailers must
carefully navigate this evolving software market
With no fully advanced replenishment application available today, retailers must care-
fully evaluate tradeoffs between proven functionality, usability, and relative company
strength. Successfully navigating this journey to a successful advanced replenishment
application deployment requires a rigorous approach with the following elements:

Evaluate all replenishment analytics rigorously

Accurate forecasting and ordering analytics are key to successful replenishment process
automation. Unless there is sufficient accuracy to minimize the number of exceptions,
users will lose trust in the application and quickly revert to the manual execution of all
orders and forecasts, dramatically reducing the benefits realized.

During the replenishment application selection process, insist on the generation of statis-
tically valid forecasts and suggested order quantities under multiple test scenarios using
real data to validate application analytics. Make sure that the range of scenarios includes
all replenishable product types, such as fast movers, highly seasonal, and rapidly trending
goods. Follow with matched control pilots to further confirm effectiveness.

Emphasize UI simplicity and a “merchant-enabling” look and feel

Since no vendor has advanced workflow functionality that is fully aligned to the
replenishment process, retailers should look to the next best alternative—an intuitive
UI that is structured to communicate recommendations in an easy-to-understand for-
mat. This is available today with well-designed exception and order review dashboards
combined with dedicated what-if capabilities.

Retailers looking for functionality to guide users through the end-to-end replenish-
ment process should consider the use of third-party Business Process Management
(BPM) applications (see the AMR Research Report “Business Process Management
Landscape—Tangible ROI Now, SOA Investment in the Future,” January 2005)
layered on top of core replenishment functionality. Two vendors, Evant and i2, are
pursuing this approach of leveraging separate BPM capabilities to provide enhanced
workflow functionality.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 15

 Integrate replenishment with your overall ARP application strategy

Selecting a replenishment application should not be a standalone decision. Retailers

should look for a level of usability, look, and feel that is as close as possible to other
planning applications that will be used by their buying teams. Where possible, strive
for technical architecture consistency across the advanced planning application portfo-
lio. It is no longer necessary to follow a functional approach to ARP technology that
uses separate replenishment, merchandise, assortment, and allocation planning applica-
tions. This approach often results in a complex matrix of integration points and the
need to reconcile the results of multiple forecasting platforms.

AMR Research recommends that retailers consider an ARP application architecture

that integrates one vendor for cross-functional workflows and UI capabilities and a
second vendor for advanced demand intelligence capabilities. Retailers should also
consider using BPM as an alternative to a packaged vendor offering to create a unified
planning interface. This horizontal approach (see Figure 5) will be easier to implement
and provide faster results than a vertical integration—where functional ARP applica-
tions are blended together.

Figure 5: Recommended ARP best-of-breed architecture integration model

Planning UI Assortment UI Allocation UI Replenishment UI Space UI

Merchant
User
Interface

Workflow Common Planning Workflow Application

Demand Optimization
Intelligence Forecasting Engine Insighters/Miners
Engines
Platform

Data Data
POS Intelligence
Warehouse

Source: AMR Research, 2005

16 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Prepare your replenishment planning teams

The first step is to determine the optimal replenishment team structure. Traditionally,
replenishment analysts have held end-to-end responsibility for the entire replenishment
process. Leading retailers are making use of a best practice from consumer products
and high-tech companies by assigning forecast algorithm and order policy configuration
responsibilities to separate sub-teams with higher skill levels. These activities are more
analytically complex than the day-to-day management of order accuracy. For example,
one specialty hardlines retailer that adopted this structure assigned the most highly
skilled 25% of its replenishment analyst team to this new role. Qualified candidates
typically have a Master’s degree or even PhD for the sophisticated analyses required.

Adjusting buying-team psychology to accept the new technology and automated

processes can make or break a project. With replenishment, the greatest challenge is
achieving the user confidence needed to keep replenishment analysts from overriding
the system. Overrides can be minimized with effective tracking and reporting to prove
out system effectiveness.

Leverage available milestone and proof-of-concept programs

Some vendors, such as JDA, are offering proof-of-concept and milestone pricing
options as an alternative to charging the full software licensing fee up front. Retailers
should take advantage of these programs when available and ask for them even when
they are not proactively offered. This can help ensure that the functionality will deliver
the desired business outcomes.

Successful pilot results can then be used to counter naysayers and foster rapid user
adoption of the new processes. Retailers should negotiate pricing for a full deployment
up front or they will lose negotiation leverage if they wait for the conclusion of a suc-
cessful pilot.

Only fund development for unique functionality

Unless a retailer has very unique requirements or a strong passion for leading-edge
development, it should remain focused on acquiring fully developed functionality
that can be quickly piloted and deployed but still fit its strategic requirements. AMR
Research has followed a number of ARP projects where major retailers establish part-
nerships with a single vendor to gain capabilities through both existing applications
and new functional development. The risk created in complex arrangements like this
often overwhelms the potential benefits.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 17

 Closing Comments
Advanced replenishment is essential for any retailer looking to differentiate itself in the eyes
of consumers through the best deployment of both inventory and staff resources. While no
vendor currently offers a fully advanced replenishment application, sufficient functionality
is now available to warrant adoption of these capabilities, with significant enhancements to
come as vendors proactively invest in evolving their replenishment products.

18 © 2005 AMR Research, Inc. AMR Research Report | July 2005

 Appendices
Appendix A: Retail replenishment vendors

Aldata

Aldata is headquartered in Finland, with 51% of its customer base in Europe. Much
of the functionality in the company’s G.O.L.D. product was originally developed
for European hypermarket chains, and this is reflected in its ability to manage food,
apparel, and hard goods. G.O.L.D. is designed to support a centralized DC replenish-
ment process while the store-level functionality enables a decentralized approach.

The application includes a wide array of forecasting algorithms and partially integrates
stores with DCs via the use of POS sales to calculate DC demand. The application
UI is based on a spreadsheet metaphor. Reviewing exceptions via G.O.L.D.’s preset
Adobe Acrobat reports may require additional manual effort in comparison to other
applications.

G.O.L.D. has progressive grocery functionality for the replenishment of Fast-Moving

Consumer Goods (FMCG) and fresh items at DCs and stores, and includes an ingre-
dients and recipe management module.

AquiTec

AquiTec declined to participate in the research for this Report, and as a private entity,
limited information was available for inclusion. The company achieved its current
structure with the November 2000 acquisition of Uniteq Application Systems, for-
merly Worldwide Chain Store Systems, which was founded in 1970.

AquiTec’s Strategic Replenishment application provides functionality for DC replen-

ishment, with a few retailers applying this DC functionality at the store level. This
grocery-oriented application includes solid functionality specific to DC-level fresh item
replenishment, and a short-cycle ordering feature that enables store-level replenishment
of fresh items, a function currently being deployed by a wholesale company responsible
for replenishment at its retail clients. The UI is browser based with a spreadsheet look
and feel.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 19

 Churchill Systems

Founded in 1989, privately held Churchill was born out of an artificial intelligence
project sponsored by IBM. While Churchill’s products are analytic engines rather
than full applications, and therefore not rated, the company has been included in this
Appendix because its Demand Forecast Manager (DFM) suite offers sophisticated fore-
casting capabilities for short lifecycle and promotional demand. The company sports a
roster of Tier 1 clients that includes Wal-Mart.

Retailers feed the results from Churchill’s forecasting tools into separate packaged, legacy,
or internally developed replenishment applications. These separate applications typically
provide all needed workflow and UI functionality, although reporting is available via
open database files and the export of data to Comma-Separated Value (CSV) files.

Evant

Evant today is the result of a merger in 2002 with NONSTOP Solutions. Evant
Replenishment stands out for its advanced cross-tier optimization capabilities that syn-
chronize DC and store inventory policies. DC demand forecasts are based on store-
level replenishment plan projections to provide additional cross-tier integration.

The software includes a single exponential smoothing forecasting algorithm, with

causal forecasting for promotions planned for an unspecified future release. Progressive
alert-driven order review workflows are supported by a comprehensive dashboard and
process activity “breadcrumb-trail” to facilitate navigation. Alerting functionality is
well designed for the complex store/SKU environment of most retailers.

GERS

GERS, a wholly owned subsidiary of Symphony Technology Group since July 2003,
purchased its planning applications from Mitek in 1999. The GERS Planning applica-
tions are also known as the PLANalyst suite. GERS’s replenishment capabilities were
not formally evaluated for this Report as they are not available on a standalone basis.
PLANalyst does include some replenishment capabilities that may be suited for hard-
lines retailers with basic needs.

20 © 2005 AMR Research, Inc. AMR Research Report | July 2005

i2 Technologies

i2’s replenishment applications have their roots in the company’s purchase of

INFOREM from IBM in 1999. i2’s global retail client base is among the more geo-
graphically varied of the vendors researched.

i2 Replenishment Planner (RP) provides progressive DC and store replenishment

capabilities backed up by one of the most advanced demand intelligence platforms
available—i2’s Demand Manager application. While a separate inventory optimiza-
tion module is needed to provide advanced cross-tier optimization capabilities, the
application does tactically integrate stores and DCs by determining DC orders from
store-level replenishment plan projections. i2 Profiler is available as a separate module
to provide seasonality capabilities for forecasting. Advanced allocation capabilities
are available from the Merchandise Allocator module that was rated separately in the
AMR Research Report “Advanced Retail Planning: Achieving Effective Demand and
Merchandise Synchronization,” January 2005.

The Replenishment Planner UI is more difficult to use than average, with a complicated
array of screens and menus. Advanced alerting and reporting functionality, supplied
through i2’s Performance Manager module, is well designed for the complex store/SKU
environment of most retailers. i2 Studio, an additional module available for workflow
overlay, was recently made available, and, according to i2, is currently being imple-
mented by two retailers.

Island Pacific

Island Pacific, formerly SVI Solutions, is 38% owned by The Sage Group plc, a
£687M UK-based company that sells accounting and business management software
for small and midsize businesses. Island Pacific’s Island Pacific Merchandising System
(IPMS) is only sold as a comprehensive merchandise management suite that includes
a full range of merchandise, assortment, and allocation planning modules. The IPMS
replenishment functionality is appropriate for a specialty soft goods retailers or general
merchandise retailers that have basic DC and store replenishment needs.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 21

 JDA

JDA’s replenishment functionality is based on the capability acquired from E3

Associates in 2001. The recently released .NET application, Portfolio Replenishment
Optimization (PRO), consists of three modules: Advanced Store Replenishment,
Advanced Warehouse Replenishment, and Network Optimization. DC replenishment
orders are tactically integrated with store needs by aggregating store-level POS forecasts
and adjusting for store-level inventory surpluses and shortages.

PRO does not include the allocation capabilities available in E3—the presumption is
that retailers will leverage progressive Arthur Allocation functionality. Forecasting capa-
bilities include a strong seasonal profiling capability, two sophisticated forecasting algo-
rithms, and the ability for users to upload promotional lifts that have been calculated at
the store/SKU level. PRO has a spreadsheet look and feel, and users navigate using tabs
that highlight various elements of functionality. Future releases are planned with sig-
nificant functionality and workflow improvements over the next 12 to 18 months.

JustEnough

Founded in South Africa in 1994 and now headquartered in California, JustEnough

was acquired by Tier90 Investments in 2001. Approximately 80% of its client base is
in the Europe, Middle East, and Africa (EMEA) region with the remainder in North
America. Its application, JustEnough Automated Forecasting-to-Replenishment, is tar-
geted primarily at the Grocery and Specialty Hard Goods retail segments.

Progressive DC and store replenishment functionality leverages projected store replen-

ishment plans to tactically integrate DC orders with store needs. JustEnough has
embedded Forecast Pro’s logic into the application. It uses proprietary multithreaded
capabilities that allow thousands of forecasts to run in parallel to achieve the scale
needed for retail. A separate merchandise planning module is fully integrated into the
application. The application UI uses a spreadsheet metaphor. Exceptions can be iden-
tified through a filter wizard, which may prove complex for some users. Crystal Reports
from Business Objects is used as a reporting engine.

22 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Lawson

The Store Operations and Supply Chain Replenishment replenishment capabilities are
built on assets acquired from UK-based Armature Holdings in 2002 and are designed
for the Grocery retailing market. The customer base is evenly split between Europe
and North America. Replenishment functionality spans six modules and the Lawson
reporting suite.

The applications support traditional DC and store replenishment abilities; the available
store-level replenishment functionality was not rated for this Report as it has yet to be
acquired and deployed by a retailer. These capabilities support a centralized DC pro-
cess combined with a decentralized store replenishment approach. Separate but similar
forecasting engines power DC and store forecasts.

The DC replenishment UI is more difficult to use than average, with a complicated

array of screens and menus. Lawson’s progressive alerting capabilities include the auto-
mated generation of prioritized exception alerts. Manual ordering and order calcula-
tions are conducted by in-store staff via handhelds, based on on-order information and
observable on-hand inventories.

Manugistics

Manugistics was founded in 1969 as Scientific Time Share Corporation (STSC)

and has built out a rich array of cross-industry supply chain applications over the
years. Replenishment functionality resides in three modules: NetWORKS Demand,
Fulfillment, and Collaborate. The newly released Inventory Policy Optimization applica-
tion enables cross-tier inventory policy optimization, but has not yet been deployed in
Retail. Manugistics has retail replenishment clients in the Grocery, Hard Goods, and
Soft Goods segments.

The most significant capability enhancement for Retail over the past few years has
been achievement of the scalability needed for store-level replenishment from the
replatformed 7.X release series. The progressive DC and store replenishment capabili-
ties determine DC orders using store replenishment plans to tie DC inventory require-
ments with store needs.

Manugistics’s forecasting capabilities have been enhanced with additional algorithms,

a best-pick capability, and demand classification functionality that automatically classi-
fies demand as seasonal, stable, lumpy, etc. and recommends appropriate parameter set-
tings. Users start with a checklist of key tasks from which they can navigate to various
elements of functionality. The interface is designed for the detailed review of informa-
tion for all Demand Forecasting Units (DFUs), which may prove to be cumbersome
for users dealing with large numbers of store/SKU combinations.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 23

 NSB Group

NSB Group was formerly known as STS Systems. The Connected Retailer Allocation
& Replenishment application is especially well suited for Tier 2 fashion retailers
looking for replenishment capability that can replenish to demand after the initial
allocation. While DC inventory requirements are calculated independently of store
needs, this application includes progressive store-level replenishment functionality.
Both DCs and stores are able to make use of nine forecasting algorithms combined
with best-pick functionality. All reporting and alerting are provided via other modules
in the Connected Retailer merchandising suite. This functionality is now available on
NSB’s new .NET platform.

Oracle (Retek)

Oracle completed its acquisition of Retek in April 2005. The company has replen-
ishment customers in the Grocery and Specialty Hard Goods retail segments.
Replenishment functionality resides in Advanced Inventory Planning (AIP) and Retek
Demand Forecasting (RDF) applications within the Xi release, with allocation capabili-
ties residing in Retek Allocation version 10, which was evaluated in the AMR Research
Report “Advanced Retail Planning: Achieving Effective Demand and Merchandise
Synchronization,” January 2005.

AIP offers progressive functionality for DC and store replenishment planning, with
DC planning driven by aggregated constrained store replenishment plans. RDF’s
advanced forecasting and demand intelligence functionality, which is well suited for
basic hardlines merchandise, is used to feed multiple ARP modules. Users navigate a
complex array of screens and menus to execute business processes.

The applications have a rich array of accuracy measurements available via online
reporting. Alerting is accomplished with variable filtering criteria, and all exceptions
are treated equally. Post-merger, Oracle/Retek continues to emphasize its strong alli-
ance with Accenture, which often supplies the systems integration assistance typically
required with each implementation.

Retalix

Retalix, with dual headquarters in Israel and Texas, has grown through the acquisi-
tion of vendors such as OMI International and, more recently, TCI Solutions and
Integrated Distribution Systems. As a result, the company has three distinct appli-
cations for grocery replenishment: BICEPS and PowerBuy for DC replenishment,
and DemandAnalytix (DAX ) for store replenishment. Connectors between DAX and
BICEPS have been built, providing cross-tier integration, but are not yet deployed.

24 © 2005 AMR Research, Inc. AMR Research Report | July 2005

BICEPS and PowerBuy provide traditional DC replenishment functionality using DC
shipment history as the basis of forecasted needs. DAX offers progressive store replen-
ishment functionality supported by sophisticated proprietary forecasting algorithms.
DAX also possesses a progressive, alert-driven order review workflow with tabs that
highlight various elements of functionality, and automated prioritized exception alert-
ing functionality.

Users of BICEPS and PowerBuy must navigate a complex array of menus and screen
layouts to execute business processes. Both BICEPS and PowerBuy have prioritized
exception alerting. DAX has progressive FMCG and fresh item replenishment for gro-
cers, with integration to handheld devices.

SAF

Founded in 1997, Switzerland-based SAF has created a name for itself by work-
ing directly with major European retailers in addition to partnering with SAP and
SofTechnics. SAF’s forecasting and optimization engines, Superwarehouse for DC
replenishment and Superstore for store replenishment, may be used either standalone
with a custom-built front end or integrated to a retail enterprise system with its own
UI. DC Link allows POS sales forecasts to be made available to Superwarehouse for
DC replenishment calculations.

SAF’s engine offers solid forecasting capabilities, with all forecasts happening at the
item/location level. The engine has 20 statistical forecasting algorithms with 150
unique configurations and causal capabilities. SAF includes the logic to generate auto-
mated prioritized exception alerts. SAF has also been used by grocers to enable decen-
tralized replenishment for FMCG and fresh item replenishment processes.

SAP

SAP’s Forecasting & Replenishment (F&R) is a recently released application that is cur-
rently under development or in pilot with six ramp-up grocery and general retail cus-
tomers. SAP partners with SAF, using SAF’s applications to provide forecasting and
order determination functionality. The basic replenishment functionality uses SAF’s
logic to determine store and DC inventory needs independently. A complex replenish-
ment workbench is configured to allow detailed review of orders by individual DC and
store inventory planning locations, which may prove to be cumbersome for users deal-
ing with large numbers of store/SKU combinations.

Retailers can select from over 600 pre-configured exception types for each role-specific
view. Progressive store-level FMCG and fresh item replenishment functionality for gro-
cers is deployed using the Perishables Workbench within SAP’s Merchandise Management
module, which generates orders centrally and allows store-level review and finalization.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 25

 SAS/Marketmax

Marketmax, a division of SAS Institute since October 2003, offers a full suite of retail
planning applications, including progressive forecasting capabilities in SAS Forecasting
for Merchandise Planning that are applicable to replenishment. 30 base forecasting
algorithms, along with solid causal capabilities and best-pick functionality, make this
one of the stronger forecasting applications available today. These capabilities are an
option for retailers looking to power packaged, legacy, or internally developed replen-
ishment applications with improved forecasting analytics.

SofTechnics

SofTechnics has been a wholly owned subsidiary of METTLER TOLEDO

International, a global $1.4B precision instrument manufacturer headquartered in
both Switzerland and Columbus, Ohio, since March 2002. The company sells to both
grocery and general merchandise retailers, and targets its replenishment capabilities to
grocers. SofTechnics partners with SAF, with SAF’s applications providing the fore-
casting and order determination functionality behind SofTechnics’s SoftGrocer replen-
ishment application. SoftGrocer has progressive FMCG and fresh item replenishment
functionality, with automated suggested order calculations that are manually reviewed
and adjusted at the store level combined with perpetual inventory maintenance sup-
port. The system also integrates to handheld devices.

The references made available for this research were all SAF application users; we were
not able to speak with any retailers using the SoftGrocer application.

Tomax

Tomax, well known for its POS, price management, and workforce management
applications, also offers time-phased replenishment functionality in its Retail Resource
Planning (RRP) application that is built on the functionality acquired from Retail
Pipeline. While the application was not rated, as it had not yet been acquired and
deployed at a retailer at the time the evaluation was conducted, the demonstrated
functionality shows promise, as illustrated by the early interest from retailers in the
Grocery, General Merchandise, and Specialty Hard Goods segments.

RRP’s overall approach to synchronizing the supply chain is to determine inventory

requirements at the store/SKU level and aggregate these needs at the DC level. The
application uses three basic time series forecasting algorithms and solid order determina-
tion logic to achieve this. While the UI has a spreadsheet look and feel, it is supported
by prioritized alerts and extensive help menus that provide process guidance to users.
RRP is integrated to handheld devices to support decentralized replenishment processes.

26 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Teradata

Teradata, a division of NCR, is best known for its scalable enterprise data warehous-
ing architecture. Teradata’s Demand Chain Management (DCM) replenishment suite,
developed from the Stirling Douglas Group’s (SDG) PREDICT and PROMPT
applications, which were acquired by NCR in 2002, is used by general merchandise,
apparel, and grocery retailers.

Teradata offers a unique combination of progressive functionality, strong workflow

and usability, and advanced alerting combined with a broad installed base and a high
degree of overall company strength. DCM’s progressive DC replenishment capabilities
integrate store needs into DC orders by aggregating forecasted store orders.

Forecast variability is accounted for in DCM; lead time variability can be fed in from
Teradata’s Supply Chain Intelligence module. Dynamic order cycle functionality is
available in Load Builder version 3.2, but is not yet deployed at a retailer. Constrained
product logic is available in DCM version 3.2, but is not yet deployed. Progressive
forecasting functionality includes five forecasting algorithms, a best-pick capability,
and the ability to automatically adjust sales history for lost sales and known causal
events. Intelligent profile clustering allows for unique seasonal profiles for store/SKU
clusters and is available as part of version 3.2.

DCM possesses strong, dedicated what-if capabilities and a well-integrated order

management dashboard. A menu of order policy elements facilitates navigation. The
Exception Monitor includes comprehensive alerting with user-defined thresholds. DCM
stands out for advanced exception reporting and prioritization capabilities that have
been designed to handle a complex retail environment with millions of store/SKU
combinations.

TXT e-solutions

Founded in 1989, publicly traded TXT is based in Milan, Italy, with the vast major-
ity of its client base in Europe. The company’s original clients were consumer goods
manufacturers, and the company today has more than a dozen retail clients. TXT is
strongest in the Fashion Retail segment, where it is frequently deployed by manufac-
turers that operate their own retail stores.

TXT’s replenishment application, TXT Plan, has progressive DC and store replenish-
ment functionality, where DC orders are integrated with store needs via the aggrega-
tion of store replenishment plans. The company’s progressive forecasting application,
TXT Demand, supports up to four planning dimensions and an unlimited number of
hierarchies with multiple forecasting algorithms and a best-pick capability. The spread-
sheet-like UI is more difficult to use than average, with a complicated array of screens
and menus.

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 27

 Table 4: Replenishment vendor demographics

Revenue
from soft-
Total Percentage Client Base Office ware license
Revenue, of Retail (total retail Network (% of total
Vendor 2004 Revenue apps.) Headquarters Summary revenue)
Aldata 66M euros 85% 100 or more Paris 7 European 25%
offices;
Atlanta; and
Bangkok
Churchill Privately Does not Less than 25 Troy, MI Markets in Privately
Systems held—does disclose Europe with a held—does
not disclose partner not disclose
Evant Privately 50% 25-50 San Francisco Atlanta and Privately
held—does London held—does
not disclose not disclose
GERS Privately 100% 100 or more San Diego, CA Redwood Privately
held—does Shores, CA held—does
not disclose not disclose
i2 Technologies $389M 15% 50-100 Dallas 27 offices 15%
in over 20
countries
Island Pacific $21.7M 95% 100 or more Irvine, CA Folsom, CA, 2004—76%;
Centennial, 2003—55%
CO, and Essex,
UK
JDA $217M 83% 100 or more Scottsdale, AZ 33 offices in 53%
Canada and
Latin America,
Europe,
Middle East,
South Africa,
and Asia-
Pacific
JustEnough Privately 60% Less than 25 Irvine, CA Sunninghill, 36%
held—does South Africa
not disclose
Lawson $364M 13% 100 or more St. Paul, MN Worldwide 23%
office network
Manugistics $206M (last 4 33% 100 or more Rockville, MD Worldwide 23%
quarters) office network
NCR/Teradata $5.98B (total Does not Less than 25 Dayton, OH Worldwide 26%
NCR); disclose office network (Teradata)
$1.4B
(Teradata)
NSB Group $84M 100% 100 or more Pointe-Claire, 9 offices Does not
Quebec in North disclose
America and
Europe
Source: AMR Research, 2005

28 © 2005 AMR Research, Inc. AMR Research Report | July 2005

 Table 4: Replenishment vendor demographics (continued)

Revenue
from soft-
Total Percentage Client Base Office ware license
Revenue, of Retail (total retail Network (% of total
Vendor 2004 Revenue apps.) Headquarters Summary revenue)
Retalix $124M 100% 25-50 Plano, Texas Worldwide Does not
office network disclose

Oracle (Retek) $174M 100% 100 or more Minneapolis, Worldwide 56%

MN office network
SAP $10.179M 3.8%* 100 or more Walldorf, Worldwide 30% (all SAP)
Germany office network
SAS/Marketmax $1.530M 3.5% 100 or more Cary, NC Over 100 More than
SAS offices 95%
worldwide
SofTechnics Wholly owned 100% 25-50 Columbus, OH Two 65%
subsidiary development
of METTLER offices in
TOLEDO the U.S.; SAF
$1.405M has HQ in
Tägerwilen,
Switzerland
Tomax Privately 100% 50-100 Salt Lake City, Bangalore, 35%
held—does UT India
not disclose
TXT e-solutions $60M 10% Less than 25 Milan, Italy 13 European 11%
offices

* AMR Research estimate Source: AMR Research, 2005

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 29

 Table 5: Replenishment product name and version

Vendor Retail Replenishment product(s) Version

Aldata G.O.L.D. 5.0.3
Churchill Systems Demand Forecast Manager 3.0
Evant Evant Replenishment: Demand Planning, Store 5.0
Replenishment, DC Replenishment, Forecasting
GERS GERS Planning (Planalyst) 8.9
i2 Technologies i2 Replenishment Planner (RP) 6.1
i2 Demand Manager (DM) 6.1
i2 Merchandise Allocator 6.1
Island Pacific Island Pacific Merchandising System (IPMS) 2.0
JDA Portfolio Replenishment Optimization (PRO) 1.0
Advanced Replenishment by E3 5.1
JustEnough Forecast-to-Replenishment, Planning & Allocation 6.7
Lawson Store Operations, Supply Chain Replenishment, Supply 2.12
Chain Distribution
Manugistics Net Demand, Fulfillment, and Collaborate 7.2.1.1
NCR/Teradata Demand Chain Management (DCM) (also Supply Chain 3.2
Intelligence)
NSB Group The Connected Retailer, Allocation & Replenishment, 3i
Forecasting
Retalix DemandAnalytix 4.0
PowerBuy 5.5/5.6
BICEPS 5.5/5.6
Oracle (Retek) Store Replenishment Planning 11.3
Retek Allocation 11.1
Warehouse Replenishment Planner 11.2
Retek Demand Forecasting 11.1
Retek Visual Space Planning 5.0
SAP Supply Chain Planning: Forecasting 4.1
and Replenishment Module
SAS/Marketmax Marketmax Allocation 5.7
SAS Forecasting 9.1
Marketmax Space Planning 5.7
Marketmax Space Optimizer 5.7
SofTechnics SoftGrocer 1.0.21
Superwarehouse & Suprestore (SAF) 3.0
Tomax Retail.net RRP (Retail Replenishment Planner) 5.1
TXT e-solutions TXT Plan, TXT Demand 2005.2
Source: AMR Research, 2005

30 © 2005 AMR Research, Inc. AMR Research Report | July 2005

 Table 6: Geographic distribution of retail client base

Central
North and South
Vendor America Europe Asia-Pacific America Rest of World
Aldata √ √ √ * *

Churchill Systems √
Evant √ √
GERS √
i2 Technologies √ √ √ √ √
Island Pacific √ √ * *

JDA √ √ √ √
JustEnough √ * √
Lawson √ √ √ √ √
Manugistics √ √ √ √ *

NCR/Teradata √ √ √ √
NSB Group √ **

Retalix √ √ √ * √
Oracle (Retek) √ √ √ √ √
SAP √ √ √ √ √
SAS/Marketmax √ √ √ √
SofTechnics √ ***

Tomax √ **

TXT e-solutions √
Source: AMR Research, 2005
√ Both clients and offices
* Clients but no offices
** Offices but no clients
*** SAF office

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 31

Table 7: Technology architecture

Largest
Deployed
Vendor Architecture Deployment OS Client Type Databases Instance
Aldata J2EE Hosted, BTF AIX V5.1, V5.2, Thin Oracle V9.2.0.3, 1,400 stores x
Solaris V9, UX V9.2.0.4 70,000 SKUs
V11
Churchill BTF Windows, UNIX n/a SQL Server, Numbers not
Systems Oracle, disclosed;
Teradata, DB2 Wal-Mart
is DFM v3.0
customer
Evant J2EE, SOA Hosted, BTF AIX, Solaris, Thin Oracle, 33 million
HPUX, iSeries DB2/400 active SKU/
locations
GERS Visual Basic/ BTF UNIX Rich, Oracle 300 stores,
Visual C developing 30,000 SKUs
Thin
i2 Technologies J2EE, 3 tier, Hosted, BTF Windows All i2 Oracle, DB2 15,000,000
SOA Server 2003 EE, applications SKU/locations
Solaris 9, HPUX have a Thin- replenished on
11.11, AIX 5.1 client Web-UI. a daily basis
Rich clients planned for 14
are optionally days
available
for data
manipulation
intensive
applications.
Island Pacific RPG/ILE, Java BTF OS/400 Rich DB2/400 460 stores x
125,000 active
SKUs

Source: AMR Research, 2005

32 © 2005 AMR Research, Inc. AMR Research Report | July 2005

 Table 7: Technology architecture (continued)

Largest
Deployed
Vendor Architecture Deployment OS Client Type Databases Instance
JDA: E3 i-Series: RPG/ Hosted, BTF UNIX, iSeries, n/a DB2/400, 500 million
ILE and C++ Windows Oracle store/SKU
combinations
JDA: PRO .NET/UNIX: C# BTF UNIX, iSeries, n/a Oracle n/a
Windows
JustEnough Borland Delphi BTF Windows 98, Thin MSDE (2000), 8 million
6; currently ME, 2000, Microsoft SQL SKU/store
being recoded 4.0, and XP; Server combinations
in C# Microsoft
Terminal
Server
Lawson Client/Server Hosted, BTF Sun Solaris 8, Warehouse Oracle 1.2 million
9AiX 5.1, 5.2 5.3 replenishment: SKU/DC
Thick, Store combinations
replenishment:
Thin
Manugistics J2EE Hosted, BTF Windows n/a Oracle n/a
Server 2000,
2003; IBM AIX
5.2, 5.3; Sun
Solaris 2.8, 2.9;
HP/UX 11.11

NCR/Teradata C, VB, 3 tier BTF Server : Thin Teradata, 64 million

Windows 2000 Oracle store SKU
(middle tier combinations
apps server),
Windows
2003, UNIX for
RDBMS; NCR
MP-RAS, HPUX,
AIX, Linux
(1Q06)
NSB Group .NET Hosted Windows Rich/Thin Oracle 9i, SQL 100 million
Server 2000 SKU/stores
combination
Source: AMR Research, 2005

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 33

Table 7: Technology architecture (continued)

Largest
Deployed
Vendor Architecture Deployment OS Client Type Databases Instance
Retalix: BICEPS Cobol Hosted, BTF Mainframe, Rich (Green 390,000 SKUs,
Programming, UNIX, Screen) 14 locations
.NET, J2EE Windows, or
AS400; J2EE
Retalix: Power RPG WebSphere AS400 Rich/Thin DB2/AS-400 n/a
Buy
Retalix: DAX Distributed N- Windows Windows Thin MS SQL 38 stores x
tier Application environment 35,000 SKUs
Oracle (Retek) RPAS (C++), Hosted, BTF Server : UNIX, Rich (RPAS RPAS, Oracle 4,500 stores
Allocation AIX, HPUX, Gui), Thin (for allocation) and 35,000
(J2EE), N tier Solaris; Client : (allocation) SKUs per
Windows 2000 location
SAP J2EE, SOA Hosted, BTF AIX, OS/400, Rich/Thin IBM DB2, SQL 6 milllion SKUs
.NET support UNIX, Linux, Server 2000, and 190 stores
through Windows, Informix,
NetWeaver HPUX, Solaris Oracle
SAS/ N-tier, J2EE, BTF Windows 2000, Thin IBM UDB 7.2 1,600 locations
Marketmax .NET XP EEE(AIX only), by 62,500 SKUs
Oracle, MS SQL
2000
SofTechnics Client/Server BTF Unix, Linux, Rich/Thin DB2/6000, 165 stores x
Windows UDB, Oracle, 100,000 SKUs
Server 2003 Informix, SQL
& XP
Tomax J2EE Both Linux, UNIX Rich/Thin Oracle, Oracle 130,000 SKUs
(Solaris, AIX, (Retail.net with RAC, DB2, DB2 per store for
etc.), Windows, the exception DPF, Informix, 225 stores
iSeries (native of RRP is thin) Informix XPS
and Linux).
TXT Client/Server, BTF Windows Rich/Thin Oracle, SQL 3,000 stores x
e-solutions SOA via Web 2000/2003/XP 1,100 SKUs
services, .NET

Source: AMR Research, 2005

34 © 2005 AMR Research, Inc. AMR Research Report | July 2005

Appendix B: Replenishment vendor evaluation
methodology
The vendors included in this Report met the following criteria:

Application in the following family of capabilities: DC replenishment, store replenish-

ment, and forecasting/demand intelligence.

Unless otherwise noted, functionality was evaluated for vendors that were able to meet
all of the following criteria:
• Complete a detailed request for information
• Provide a live demonstration of the applications evaluated
• Provide two or more retailer reference customers with revenue of over $500M that
had fully installed the evaluated applications
• Evaluated applications met all criteria listed for the assigned functionality ratings.

Table 8: Replenishment functionality ratings criteria

Criteria Basic Progressive Advanced

DC Replenishment Independent Replenishment Tactical cross-tier synchroni- Simultaneous cross-tier
at each tier zation: DC orders adjusted optimization of inventory
for store inventory surpluses policies
Order plans based on local and shortages
demand signal (e.g. DC ship- Dynamic adjustment of
ments) Order plans based on store safety stock for forecast and
POS sales supply variability
Safety stock set manually
Fixed order cycles Dynamic order cycles per
economics
Safety stock determined auto-
matically per order policies

Store Replenishment Independent calculation of Safety stock determined Dynamic adjustment of

store replenishment orders automatically per inventory safety stock for forecast and
policies supply variability
Safety stock set manually for
store groups Constrained product logic
applied when available
inventory does not meet
store needs
Allocation Basic top-down functionality Top-down functionality Sophisticated attribute-
based bottom -up allocations
Rounding logic Basic bottom-up based on
like items and rounding logic
-OR-
Pack optimization
Source: AMR Research, 2005

AMR Research Report | July 2005 © 2005 AMR Research, Inc. 35

Table 8: Replenishment functionality ratings criteria (continued)

Criteria Basic Progressive Advanced

Forecasting and Demand Basic time-series forecasting Multiple sophisticated Common forecasting
Intelligence algorithms (moving forecasting algorithms engine feeding multiple ARP
averages) within modules (e.g. expo- modules
nential smoothing, trend,
Single product hierarchy seasonality)

Causal forecasting capabili-

ties (e.g promotions)

Best-fit calculations (auto-

matic selection of algorithm
with lowest forecast error)

Multiple product hierarchies

Merchandise and Space Able to push out or upload Integrated workflow Suite includes fully- inte-
Planning System Integration data with upload and analysis grated merchandise and
capabilities space-planning modules
No upload & screens to with workflow necessary
analyze the data Either merchandise or space to integrate replenishment
planning integrated to processes
Replenishment
Workflow Functionality and Limited workflows Dedicated what-if function- Process activity guides
Usability capabilities ality
Comprehensive exception/
Spreadsheet metaphors Integrated exception and order review workflow
order review dashboard
No what-if capabilities Cohesive order strategy
dashboard and workflow
Alerting and Reporting Basic reports with drill-down Automated prioritized Flexible, user-definable alert
capabilities exception alerting thresholds
-OR-
Ability to set variable filter- Full suite of accuracy mea-
ing criteria surements
Store-level Grocery/ Manual ordering per sales Automated suggested order Fully automated order
FMCG and fresh item history and forecast data, calculation determination
replenishment and/or inventory levels
Orders manually reviewed Post-order exception driven
Wastage and spoilage input and adjusted at stores review process
manually
Supports perpetual Automated calculation of
maintenance wastage and spoilage
projections by item
Production System
Integration

Handheld hardware
Yes=Functionality available No=Functionality not offered in current release
integration

Scale connectivity (for

inventory control)
Source: AMR Research, 2005

36 © 2005 AMR Research, Inc. AMR Research Report | July 2005

 Company List
ADC www.adctech.com

Adobe www.adobe.com

Accenture www.accenture.com

Aldata www.aldata-solution.fi

AquiTec www.aquitecintl.com

Business Objects www.businessobjects.com

Churchill Systems www.churchillsys.com

DCM Solutions www.dcmsolutions.com

Evant www.evant.com

GERS www.gers.com

i2 Technologies www.i2.com

IBM www.ibm.com

Invatron Systems www.invatron.com

Island Pacific www.islandpacific.com

JDA www.jda.com

JustEnough www.justenough.com

Lawson www.lawson.com

Manugistics www.manugistics.com

NCR/Teradata www.ncr.com

NSB Group www.nsbgroup.com

Oracle/Retek www.retek.com

Park City Group www.parkcitygroup.com

Retalix www.retalix.com

SAF ww.saf-ag.com/english

SAP www.sap.com

SAS/Marketmax www.sas.com

SofTechnics www.softechnics.com

Tomax www.tomax.com

TXT e-solutions www.txt.it

Acronyms and Abbreviations
ARP Advanced Retail Planning

BPM Business Process Management

CAO Computer-Assisted Ordering

AMR Research provides world-class research
CSV Comma-Separated Value
and actionable advice for executives tasked
with delivering enhanced business process DC Distribution Center
performance and cost savings with the aid
of technology. 5,000 leaders in the Global DFU Demand Forecasting Unit
1000 put their trust in AMR Research’s integ-
rity, depth of industry expertise, and pas- EMEA Europe, Middle East, and Africa
sion for customer service to support their
most critical business initiatives, including
FIM Fresh Item Management
supply chain transformation, new product
FMCG Fast-Moving Consumer Goods
introduction, customer profitability, com-
pliance and governance, and IT benefits
NRF National Retail Federation
realization. More information is available at
www.amrresearch.com. POS Point of Sale
Your comments are welcome. Reprints
are available. Send any comments or
SKU Stock-Keeping Unit
questions to:
UI User Interface
AMR Research, Inc.
125 Summer Street
Boston, MA 02110
Tel: +1-617-542-6600
Fax: +1-617-542-5670
www.amrresearch.com

555 Montgomery Street

Suite 650
San Francisco, CA 94111
Tel: +1-415-217-3737

Whittaker House
Whittaker Avenue
Richmond, Surrey TW9 1EH
United Kingdom
Tel: +44 (0) 20 8822 6780
Fax: +44 (0) 20 8822 6790

AMR-R-18055