A M M I 3 ( ( P P P' : G : ER M B < L F : @ : S B G > ' < H F

=KBOBG@;>MM>K;NLBG>LL=><BLBHGL L N F F> K  +) )2

MA>>OHENMBHG
H?IKB<BG@
:?M>KFHOBG@?KHFI:I>KK>IHKMLMHLIK>:=LA>>MLMH:G:ERMB<L%PA:MÌLG>QM8

%073-27-()
 &),-2(8,)796+)Ahphk`ZgbsZmbhgl\b^g\^a^ei^]mnkgZkhng]ma^pZkbgBkZj
 463&%&-0-8=1%2%+)1)28:g^pZgZermb\_kZf^phkd_hk_ZlmZg]kh[nlmlmh\aZlmb\ZgZerlbl
 13()0-2+8,)1%6/)874%'):gZermb\ZekhZ]fZi_hk[^mm^klmkZm^`b\[nlbg^llieZggbg`
 0)77-7136)AIpbglma^>]^efZg:pZk][rmkZgl_hkfbg`bmlfZllbo^ikh]n\mihkm_hebh
B G L B =>  LMH KR
FHO>KLLA:D>KL
Some of the biggest names in
the high-tech software and con-
sulting universe – IBM, SAS, Mi-
crosoft and Fair Isaac to name just
a few – have seen the future of
business in this increasingly inter-
national, interconnected world and
the future is … analytics.
IBM, for example, recently ac-
quired ILOG, a market leader in op-
timization software, and launched
an ambitious plan to hire or retrain
as many as 4,000 analytics con-
sultants and professionals to sup-
port a recently launched network of
Analytics Solution Centers around
the world (page 7).
Earlier this year, SAS, a recog-
nized leader in business analytics
software and services with 45,000
customer sites worldwide, intro-
duced a series of promising new
products at its SAS Global Forum,
“Competing on Analytics,” in which
former Secretary of State Mad-
eleine Albright spoke on the topic
“Leading with Confidence in an Era
of Uncertainty.”
Microsoft is expanding its pres-
ence in the optimization and related
analytics markets as evidenced by
its appearance at the INFORMS
*       u      : G : ER M B < L F : @ : S B G > ' < H F
Practice Conference in Phoenix in
April, while Fair Isaac (now FICO)
bolstered its credentials in predic-
tive analytics with last year’s acqui-
sition of Dash Optimization.
Despite the worldwide economic
woes, this appears to be a good time
to be riding the analytics bandwag-
on, whether you’re on the provider
I:K:@HG
end or the benefits end.
And then there’s the story of
Hewlett-Packard, the world’s largest ?NEEI:@>:=O>KMBL>F>GM
technology company with a supply
chain to match, that both provided
and benefited from analytics in the
form of cutting-edge operations re-
G>P:KMPHKD
search and algorithms that tamed
HP’s bulging product portfolio prob-
lem and returned $500 million to the
I:@>*
bottom line at the same time. The
work earned HP the coveted 2009
Edelman Award from INFORMS for
outstanding achievement in opera-
tions research (page 32).
As always, I look forward to
your feedback. Analytics is an in-
teractive exercise. Contributed ed-
itorial ideas, articles and columns
are welcomed. ❙
ÇI>M>KAHKG>K%>=BMHK
ahkg^k 9 ebhgakmin['\hf
< H G M > G M L






 




DRIVING BETTER BUSINESS DECISIONS LNFF>K+))2

 


 
*)%896)7
8 MA>>OHENMBHGH?IKB<BG@
By Neil Biehn
The evolution of pricing, from paper reports to analytics. What’s next?

15 MA>L<B>G<>;>ABG=MA>LN<<>LL?NEËLNK@>Ì
By Doug Samuelson

8
How organization science altered military decision-making in Iraq.
?KHGMEBG>*
19 IKH;:;BEBMRF:G:@>F>GM
By Sanjay Saigal
A new concept for handling uncertainty and the “flaw of averages.”
?NEEI:@>:=O>KMBL>F>GM
24 FH=>EBG@MA>F:KD>MLI:<>
By Santiago Castro
How to represent a company’s strategic situation with MCDA.
G>P:KMPHKD
28 L>O>GLBGLH?LBFNE:MBHGIK:<MB<>
By Jerry Banks and Randall R. Gibson
Long-time consultants reveal valuable lessons learned the hard way. Risk Analysis is Easy as What-If with our Ultra-Fast,I:@>+
15 32 E>LLBLFHK>?HKA>PE>MM&I:<D:K= In true Excel 2007 style, select from galleries, drag and
By Peter Horner
HP wins the Edelman Award by taming its massive product portfolio. the
watch charts update instantly with new results.
37 BGM>E>:KGLBG?HKFLIKBS> consistent user interface makes learning and teaching
Intel Decision Technology Group saves chipmaker “billions of dollars.”
Free Trials and Support for Your Eīorts at Solver.com.
38 BG?HKFBG@MA>@EH;>
Preview of meeting where networking opportunities abound about the many other ways we can support your
teaching and learning eīorts – all at www.solver.com.
for analysts.

28
produce instant reports plus standard, Overlay, Trend
and Box-Whisker charts, without programming. Tel 775 831 0300 ‡ Fax 775 831 0314 ‡ info@solver.com

+       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

 
 



 
 

DRIVING BETTER BUSINESS DECISIONS

 


 
REGISTER FOR A FREE SUBSCRIPTION:
http://analytics.informs.org

INFORMS BOARD OF DIRECTORS

President Don M. Kleinmuntz
President-Elect Susan L. Albin
Past President Cynthia Barnhart
32 Secretary Anton J. Kleywegt
Treasurer Stephen M. Robinson
Vice President-Meetings Robin Lougee-Heimer
Vice President-Publications Terry P. Harrison
Vice President-
Sections and Societies Ariela Sofer
Vice President-
Information Technology Warren Lieberman
Vice President-Practice Activities R. John Milne
Vice President-International Activities Michel Gendreau
Vice President-Membership
and Professional Recognition
Vice President-Education
Pinar Keskinocak
Jayashankar M. Swaminathan
?KHGMEBG>+
Vice President-
Marketing and Outreach Anne G. Robinson
Vice President-Chapters/Fora John W. Fowler

INFORMS OFFICES
www.informs.org • Tel: 1-800-4INFORMS
?NEEI:@>:=O>KMBL>F>GM
39 Executive Director Mark G. Doherty
Marketing Director Gary Bennett
Communications Director Barry List
G>P:KMPHKD
()4%681)287 Corporate, Member, INFORMS (Maryland)
Publications and 7240 Parkway Drive, Suite 310
Subdivision Services Hanover, MD 21076 USA
1 BGLB=>LMHKR Tel.: 443.757.3500
E-mail: informs@informs.org
Meetings Services INFORMS (Rhode Island)
Easier, More Powerful – and Far More Visual. I:@>,
4 OB>PIHBGM 12 Breakneck Hill Road, Suite 102 From its classic Solver Parameters dialog to its modern
Lincoln, RI 02865 USA Ribbon, modeless Task Pane, and charts to visualize the
Tel.: 401.722.2595
7 G>PLF:D>KL E-mail: meetings@informs.org

39 <HKIHK:M>IKH?BE>
43 E:LMPHK=
Analytics (ISSN 1938-1697) is published quarterly by the
Institute for Operations Research and the Management
Sciences (INFORMS). For a free subscription, register at http://
analytics.informs.org. Address other correspondence to the
editor, Peter Horner, horner@lionhrtpub.com. The opinions
expressed in Analytics are those of the authors, and do not
necessarily reflect the opinions of INFORMS, its officers,
Lionheart Publishing Inc. or the editorial staff of Analytics.
Analytics copyright ©2009 by the Institute for Operations
Research and the Management Sciences. All rights reserved.
ANALYTICS EDITORIAL AND ADVERTISING and a wide array of charts and graphs. One simple,
Lionheart Publishing Inc., 506 Roswell Street, Suite 220, Marietta, GA 30060 USA consistent user interface makes learning and teaching
Tel.: 770.431.0867 • Fax: 770.432.6969
President & Advertising Sales John Llewellyn
Free Trials and Support for Your Eīorts at Solver.com.
Llewellyn@lionhrtpub.com
Tel.: 770.431.0867, ext.209
Editor Peter R. Horner
horner@lionhrtpub.com
about the many other ways we can support your
Tel.: 770.587.3172 teaching and learning eīorts – all at www.solver.com.
Art Director Alan Brubaker
albrubaker@lionhrtpub.com
Tel.: 770.431.0867, ext.218
Assistant Art Director Kat Wong
katwong@lionhrtpub.com
Tel.: 770.431.0867, ext.223 and produce instant reports, charts and graphs, without
programming. And you can do even more in Excel VBA. Tel 775 831 0300ͻFax 775 831 0314ͻŝnfo@solver.com

,       u      : G : ER M B < L F : @ : S B G > ' < H F

OB>PIHBGM
><HGHFB<K><HO>KR
MAKHN@ALNIIER
<A:BGK>:LHGBG@
:EMAHN@A:GRLNIIER With economic gloom spreading globally,
<A:BGBLBGK>:EBMR:G governments are under pressure on many fronts:
BGM>K&<HGG><M>= families want relief on mortgages in the United
G>MPHKDH?>GM>KIKBL>L States, the auto industry wants help in western
PBMAHNM:GH;OBHNL countries, and banks need more help despite
LM:KMBG@IHBGM%BMA>EIL questionable results thus far. The economy itself
MHMABGD:;HNMBMBG is a complex network that defies a clear starting
M>KFLH?=>F:G=% point. How should governments develop and ex-
plain an integrated set of initiatives as they face
IKH<>LLBG@:G=LNIIER'
a growing number of demands? Businesses too
have to make sense of the apparent pot pourri of
government initiatives.
We advocate approaching economic recov-
ery from a supply chain perspective.
%79440=',%-23*7)'8367
Viewing the economy as a supply network
of industry sectors harks back to the late 1940s
when Wassily Leontief, a Nobel Laureate in
Economics, proposed input-output modeling.
Each sector receives flows from other sectors
and transforms them into flows for other sec-
;RF:GFHA:G tors. The well-being of any sector is connected
L'LH=AB!E>?M"
to that of other sectors in this supply chain flow:
:G=<AKBLMHIA>K
L'M:G@!KB@AM" when demand for one sector falls, derived de-
mand falls like dominoes for upstream nodes
in the supply chain.
-       u      : G : ER M B < L F : @ : S B G > ' < H F
Consider the retail sector in the United
States. In January 2009, Circuit City, the sec-
ond largest consumer electronics retailer in the
U.S., announced the closure of all 567 stores.
This closure creates 18 million square feet of
vacant retail space in an already faltering com-
mercial real estate market. General Growth
Properties, one of the largest owners of 200
shopping malls in the U.S., issued a warning
in January 2009 that it may seek bankruptcy
court protection due to problematic debt pay-
ments and carried out the filing in April. The
closing of Circuit City alone has caused a re-
duction in 5 percent to 10 percent advertising
revenue for certain newspapers. As a result,
many newspapers are cutting back on their
news coverage, which causes the readership
to decline. Newsstand sales of single-copy
magazines such as Sports Illustrated and
Cosmopolitan in the U.S. had already fallen 11
percent in the second half of 2008 from a year
earlier so the drop in advertising revenue may
push various newspaper and magazine com-
panies over the edge to bankruptcy. On Feb.
23, 2009, the Wall Street Journal reported the
weekend bankruptcy filings of Philadelphia’s
two major newspapers and Journal Register
Co., publisher of the New Haven Register and
19 other dailies. Finally, there is significant loss
in account receivables for struggling vendors
who supply goods to failed stores.
Consider also that to help the auto indus-
try, the U.S. government is helping out the
entire supply chain. On Dec. 19, 2008, the
Bush administration announced a short-term
loan of $13.4 billion to General Motors to tide
it over for a few months. GM then asked the
U.S. government to bail out its finance arm,
GMAC, and on Dec. 29, 2008, the government
approved a short-term loan of $5 billion. On
Feb. 4 of this year, 400 auto suppliers, includ-
ing the giant auto suppliers American Axle and
Visteon Corp, asked for $25.5 billion of federal
aid because of cash flow problems caused by
the delayed payments from the U.S. automak-
ers and asked the government to ensure that
payment terms are 10 days rather than the
usual 55+ days. Car dealers are also consider-
ing requesting federal aid to stay afloat. In late
January 2009, two city governments in Califor-
nia approved loans to their local car dealers.
Finally, on Feb. 4, the U.S. Senate voted for a
tax deduction for sales tax and car loan inter-
est for purchasing a new car under $49,500
between Nov. 12, 2008 and Dec. 31, 2009.
Although any supply chain is in reality an
inter-connected network of enterprises with-
out an obvious starting point, it helps to think
about it in terms of demand, processing and
supply. We can do likewise with the economy
or any part of it. To revive the economy, any
government should: 1) stimulate demand from
consumers, exports and the public sector; 2)
improve the economy’s processing including
financial transactions; and 3) help improve
supply by lowering the cost of inputs.
1. Stimulate demand. To increase con-
sumer demand, companies offer lower prices.
Lowering the value-added-tax (VAT) from 17.5
: G : ER M B < L      u      L N F F > K  + ) ) 2
OB>PIHBGM
:MMA>O>KRE>:LM% percent to 15 percent in the United Kingdom
MA>@HO>KGF>GM amounted to a 2.1 percent price drop for most
LAHNE=BFIHL> products and services. However, some re-
IKH=N<MBOBMR tailers have simply maintained the post-VAT
BFIKHO>F>GM price, thus pocketing the reduction rather than
<HG=BMBHGLBG passing it on to consumers. Lowering withheld
;:BEBG@HNMMA> taxes, as effective in the U.S. starting April 1 in
;:GDBG@:G=MA> accordance with the stimulus package passed
in February 2009, also lowers prices as a per-
:NMHBG=NLMKR'
centage of disposable income.
Companies can generate new demand for
products that are more attractive to end-cus-
tomers than existing products. For example,
U.K. supermarket Tesco’s has introduced dis-
count brands to better fit the consumer’s re-
cently revised budget. Some businesses have
instead increased prices for existing products
to compensate for fewer customers, suggest-
ing a short-term outlook.
The government can support companies to
meet future demand by supporting research
into new technologies. Both the U.S. and U.K.
governments have stated their interest in do-
ing so. These governments can also help com-
panies with retraining of employees in order to
meet the changing demand. However, helping
sectors continue making products whose un-
derlying demand has fundamentally declined
cannot improve economic flows.
As regards the government’s demand
through public spending, an example is
LN;L<KB;>MH:G:ERMB<L
It’s fast, it’s easy and it’s FREE! Just visit: http://analytics.informs.org/
.       u      : G : ER M B < L F : @ : S B G > ' < H F
investing in infrastructure. Such investment
may not “cost” much if these are projects that
the government was going to do so anyway
and is only bringing that investment forward.
Upgrading the electricity grid in the U.S. as
part of a stimulus package is an example. Al-
though public expenses go up, in the long run
this may benefit the economy if the infrastruc-
ture is truly necessary.
Finally, governments can also seek to enable
an increase in exports through export guaran-
tees, especially for medium-sized companies.
The U.K. government could extend the existing
range of support services to U.K. businesses
exporting to other countries through the Export
Credits Guarantee Department (ECGD) and the
U.K. Trade & Investment. Removing sanctions
selectively on countries like Cuba or Iran, while
controversial, can also help save jobs in the
countries that maintain these sanctions.
2. Improve processing. Based on size,
the auto sector and the banking sectors are
critical to the U.K. economy. However, from a
productivity perspective, the ratio of outputs
to inputs for these two sectors are lower than
for other sectors in the U.K. such as sport-
ing goods, toys, tobacco products and alco-
holic beverages. Providing additional input
and support to these two sectors then could
generate far less output than if the aid were
given to the other sectors. At the very least,
the government should impose productivity
improvement conditions in bailing out the
banking and the auto industry.
In any supply chain, the availability of
credit and cash is essential for transactions
to take place. However, banks in both the
U.S. and the U.K. have created roadblocks
for transactions by providing neither credit
guarantees for buyers nor adequate lending
for working capital for sellers. In the U.K.,
despite government guarantees to small and
medium enterprises to the tune of £1 bil-
lion, banks had barely lent 1 percent of that
amount as of mid-February 2009.
To get credit flowing, governments may
have to break banks up by asset types or to
create new institutions altogether. The Royal
Bank of Scotland, with the U.K. government
as its majority owner, is separating out 20 per-
cent of its assets, so-called toxic ones, this
way. More needs to be done to reduce com-
plexity and get credit flowing again regardless
of whether large banks in the U.S. and U.K.
are nationalized or not. While Prime Minister
Gordon Brown has ruled out a “rigid divide”
between retail and corporate deposits on one
side and international investment banking and
trading on the other, he has opened the pos-
sibility of the “reinvention of the traditional sav-
ings and mortgage bank in Britain.” Given the
sizes of stimulus packages worldwide, creat-
ing new institutions like a World Bank for busi-
nesses could also be considered.
3. Improve supply, reduce input costs.
There is no shortage of supply given lack-
luster demand, but input costs could cer-
tainly be lower overall. Although prices of
: G : ER M B < L      u      L N F F > K  + ) ) 2
OB>PIHBGM
LMBFNE:MBG@MA> oil and many commodities are down, steel
><HGHFB<?EHPL prices are beginning to go back up despite
:EHG@:LNIIER<A:BG low demand from the auto sector. There is
PBMAHNMBFIKHOBG@MA> also the cost of labor. This is already get-
?BG:E=>F:G=BLPHKL> ting lower, for instance, in the construction
MA:GK>:KK:G@BG@MA> sector in the U.K. Governments in the devel-
oped world can lower wages, the argument
<A:BKLHGMA>=><DH?
MA>=HHF>=MBM:GB< being that a lower-paying job is better than
a non-existent one – recall former CEO Lee
ÇMA>K>BLGH;>G>?BM
Iacocca’s challenge to the unions to keep
=>LIBM><HGLB=>K:;E>
Chrysler alive. Countries like France could
BGO>LMF>GM'
lower the cost of business by making it easier
to hire and fire workers rather than stipulate
aid to automakers on the basis of not laying
off any workers. Moreover, there is the issue
of unemployment benefits: if the difference
between wages and unemployment benefits
becomes too low, the labor pool may shrink.
Thus, lowering wages is tied to lowering un-
employment benefits.
Investment in infrastructure can help if it
lowers the cost of doing business as part of
the input prices. For instance, infrastructure
development could help lower costs in coun-
tries like China and India; despite a bumper
wheat crop this year, prices may not come
down too much in India due to poor storage
and transportation infrastructure. Indeed,
China’s $586 billion package announced in
November 2008 for housing, infrastructure
and post-earthquake reconstruction over
two years received warm support from eq-
uity markets. However, infrastructure spend-
ing may not help much in the United States
or in France to reduce costs for businesses
/       u      : G : ER M B < L F : @ : S B G > ' < H F
unless it is work already planned for the fu-
ture that will be implemented earlier.
A key input for most businesses is the cost
of financing, and governments can help reduce
that as we discussed earlier.
*3'9732*03;
Stimulating the economic flows along a
supply chain without improving the final de-
mand is worse than rearranging the chairs
on the deck of the doomed Titanic – there is
no benefit despite considerable investment.
Some aspects of stimulus packages in west-
ern countries could lead to such a situation.
For instance, governments could help auto
companies make cars that do not find buy-
ers. Banks taking money from the govern-
ment to improve their balance sheets and
to give themselves bonuses, contractual or
not, without helping to get economic trans-
actions going again are simply moving the
money from government coffers to their own.
Retailers who pocket the decrease in VAT-
rate reduction in the U.K. similarly defeat the
purpose of stimulating consumer demand.
Thus, governments need to look carefully
at their plans to see if the benefits might just
be one-offs rather than steady improvements
to flow in the supply chain all the way from im-
porters through manufacturers and retailers to
household consumption or exports. ❙
ManMohan S. Sodhi (M.Sodhi@city.ac.uk) heads the Operations
Research group at the Cass Business School, City University
London, London. Christopher S. Tang (ctang@anderson.ucla.
edu) is the Edward Carter Professor of Business Administration at
the Anderson School of Business, UCLA, Los Angeles.
: G : ER M B < L      u      L N F F > K  + ) ) 2
G>PLF:D>KL
B;FMHHI>G
@EH;:EG>MPHKDH?
:G:ERMB<L<>GM>KL
B;FPBEEK>MK:BG IBM recently announced a significant
HKABK>:LF:GR:L expansion of its capabilities around busi-
-%)))G>P:G:ERMB<L ness analytics with plans to open a net-
<HGLNEM:GML:G= work of Analytics Solution Centers around
IKH?>LLBHG:EL' the world, beginning with five in the sec-
ond quarter of 2009. These initial centers
will be located in Tokyo, London, New York
City, Beijing and Washington, D.C. As part
of this initiative, IBM will retrain or hire as
many as 4,000 new analytics consultants
and professionals.
According to the company, the centers will
enable IBM “to meet growing client demand
for advanced analytics capabilities as part
of new, smarter business systems.” Much
of the demand is driven by new stimulus in-
vestments around the world in areas such
as financial risk management, smart grids,
electronic medical records and food track-
ing. These clients are embedding new sen-
sor technology into their processes in order
to gather better performance and manage-
ment data. Organizations are then leverag-
ing new analytics capabilities to turn that
data into predictive intelligence to help run
new digital infrastructures more effectively
and smarter.
0       u      : G : ER M B < L F : @ : S B G > ' < H F
For example, IBM announced an analyt-
ics engagement with the State of Arkansas to
monitor and manage the use of stimulus grants
for improving education programs. In addition,
IBM introduced new optimization software to
help clients solve complex problems and pre-
dict outcomes.
“Advanced analytics are increasingly es-
sential to help companies and organizations
confronted with vast amounts of data and sys-
temic change and who are looking to build
smarter business systems,” says Samuel J.
Palmisano, IBM’s chairman, president and
chief executive officer. “All organizations today
need to sort through myriad choices, make
smarter decisions quickly and accurately and
act decisively.”
The first three centers will sit in the
world’s financial hubs – Tokyo, London and
New York City. Staffed initially with more
than 100 consultants and mathematicians
each, these centers will serve clients seek-
ing a deeper, system-level view of financial
risk across markets as they pursue more
predictive business outcomes in the new
economic condition.
The IBM Analytics Solution Centers ini-
tially will be staffed with domain experts from
across IBM. As demand grows, IBM will shift
training investments to hire or retrain 4,000
high-skilled workers needed for these next-
generation positions.
A fourth center will be located in Washing-
ton, D.C., at the IBM Institute for Electron-
ic Government, and will focus on analytics
requirements needed to manage economic
recovery investments. It will house IBM’s
Economic Recovery Fund Tracking dash-
board, an analytics tool designed for federal,
local, and state government officials to moni-
tor and report on stimulus spending from the
American Recovery and Reinvestment Act
(ARRA). The State of Arkansas is the first
state to deploy this new analytics system to
track up to $569.9 million in grants for im-
proving education programs.
A fifth Analytics Solution Center will open
in Beijing, and will focus on rail and transpor-
tation analytics. A major stimulus opportunity
in China, transportation infrastructure such as
new high-speed trains, a hugely expanded rail
network and faster links between the country’s
major cities are in development. This center
will bring together research experts, major uni-
versities, partners and clients to develop best
practices for rail systems that are smarter,
more efficient, greener and attractive for trav-
elers all around the globe.
In other news, IBM established its first-
ever new service line in its Global Business
Services unit. IBM Business Analytics and
Optimization Services draws on the com-
pany’s expertise in vertical industries, re-
search, software and mathematics to help
clients both improve the speed and quality
of business decisions while better under-
standing the consequences and outcomes
of those decisions. The organization is
staffed with more than 4,000 IBM consul-
tants worldwide. ❙
: G : ER M B < L      u      L N F F > K  + ) ) 2
; N L B G > L L & MH & ; N L B G > L L

MA>>OHENMBHG
H?IKB<BG@
?KHFI:I>KK>IHKMLMHLIK>:=LA>>MLMH;NLBG>LL
BGM>EEB@>G<>%LF:KM>KIKB<BG@:E@HKBMAFL:K>;>BG@
=>O>EHI>=>O>KR=:R'LHPA:MÌLG>QM8K>:EL<B>G<>'

;RG>BE;B>AG

THE BUSINESS
BU -TO-BUSINESS (B2B) evolutionary path of pricing. Each step information previously unavailable. Busi- a SAS executive, Jim Davis [1]. In that talk,

8 landscape is littered with a

landsc
vast array of manufactur-
distributors and service
ers, d
providers.
id Each
E h one, unique in its own
way, has a process to create pricing
for the organization. At one company,
a new salesperson makes a gut-feel
estimate of how much to increase the
price of a high-visibility item. In another,
a staff of scientists uses analytics, data
mining and science to price all aspects
of the business. Every company has
its own story and its own place on the
of the way provides new insights, new
ways of doing business and a clear,
measurable increase in revenue and
profit dollars.
The question that hounds most execu-
tives is exactly how the tools and new ways
of thinking drive value to the bottom line
- paper reports, data specialists, spread-
sheets, databases, business intelligence
software and pricing science. Paper reports
provide the macro-level information to un-
derstand the overall business performance.
Data gurus create new tools that generate
ness intelligence and analytical software
packages bring data to a common, easy-to-
use platform that is accessible to the entire
organization. And finally, there is “pricing
science” – the power to recognize oppor-
tunities with minimal effort, providing your
team a direct set of marching orders speci-
fying where action must be taken. In this
paper, we’ll examine the benefits and frus-
trations associated with each step along
the evolutionary path.
The inspiration for this article was a pre-
sentation delivered in the summer of 2008 by
Davis spoke about a company’s competitive
advantage as it compared to the degree of
intelligence using various analytical tools, in-
cluding standard reports and science based
optimization. There have been many books
that touch upon this subject, but one in par-
ticular stands out – “Competing on Analytics:
The New Science of Winning” by Davenport
and Harris [2]. In that book, the authors rec-
ognize (and give credit to SAS) of the vari-
ous levels of complexity analytics can take
on. Moreover, there is discussion on pric-
ing optimization and the power it can wield

1       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

I K B < B G @  > OH E N M B H G
across many industries. Finally, as an im-
plementer of pricing software, I’ve gained
a detailed view of how companies actually
evolve, including the reasons to start down
the evolutionary path and the benefits of
each step. The evolution of a pricing orga-
nization is mired in huge pieces of paper,
quirky data specialists, information over-
load and “behind the scenes” equations
which explain why companies embark on
the road to pricing excellence. The journey
begins with the use of paper reports.
4%4)66)43687
MOST PEOPLE READING this article have
some sort of analytical or scientific back-
ground and might think negatively of com-
panies who rely on paper reports. But the
simple fact remains; paper reports are an
extremely common and valuable approach
to understanding the performance of vari-
ous markets, business units and product
lines. A quirky attribute of paper reports
is their unique names as many of the re-
ports are purposely named with an ironic
sense of humor. At the top of the list is the
“Timber Report.” Its name comes from
the simple fact that company employees
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
2       u      : G : ER M B < L F : @ : S B G > ' < H F
believe an entire tree is killed just to create
it (good news for environmentalists, this
company recycles). A common favorite is
the “Kevin Halpin Report” – named after
an innovative manager who began gen-
erating his own information that was sub-
sequently adopted by the entire company
(the name has been changed to protect
the innocent). This last one is ubiquitous –
I have probably seen reports named after
individuals more than a dozen times.
The most common report, especially
in manufacturing, is the detailed profit
and loss statement (P&L). The P&L al-
lows category managers, pricers and
even sales people to clearly understand
the relationship between price and cost.
As revenues and costs change, price set-
ters use the P&L to analyze the resulting
impact to margin. More importantly, the
P&L offers a glimpse into what must be
done to meet budget and forecast goals.
Over and over again, executives ask
their managers to hit targets. And where
do these managers go to first? The an-
swer is the P&L. Using a calculator, it’s
not too difficult to determine the types of
price increases you’ll need to achieve the
corresponding goal. As most P&L reports
are spread across product lines, business
units and key markets, it can be an effec-
tive tool to understand how price impacts
bottom line.
I K B < B G @  > OH E N M B H G
Despite their efficacy as sources of
more general information, paper reports
are fundamentally limited. To start, they are
at a very high level. There are few details
to support the underlying numbers. As a re-
sult, more and more columns or alternative
reports are created, often at varying levels
of granularity, in an attempt to gain further
insight into the business. To inject more
and more information on the same report,
the size of paper used to generate these
reports can take up the entire length of the
analyst’s desk. To overcome the lack of
timeliness, the burden of generality and the
lack of paper real estate, many companies
turn to spreadsheets and database as the
next step when using price as a key lever
to drive business.
746)%(7,))87%2((%8%&%7)7
AT SOME POINT, managers want access
to the data behind paper reports. Since it
is kept in a warehouse or more typically,
disparate systems, the data must reside
somewhere. In a quest for more informa-
tion, executives begin demanding more
elaborate and complex reports. Enter the
data guru. In many cases, these experts
have an enormous amount of power – and
often for good reason. With only paper re-
ports to rely on, executives and managers
clamor for more detail, specifically around
why a product line or business unit is
*)       u      : G : ER M B < L F : @ : S B G > ' < H F
underperforming. “Widget sales are down people in the organization turns to outright Data gurus become spreadsheet manu-
this quarter and we recently had a big price terror as the workload grows exponentially. facturers. With these new spreadsheets,
increase – which customers have reduced More and more technically skilled individu- non-technical analysts can use software
their spending with us?” The P&L only als are hired as the need for tools becomes they are familiar with in order to achieve
points you to the problem, the data guru evident. A major milestone in the evolution granular reporting. Paper reports are now
can help find out why. of pricing reaches a critical juncture – com- more interactive. Some spreadsheets are
The tools of the data guru are clear – panies must empower their sales organiza- over a hundred columns wide and replace
spreadsheets and databases. These spe- tions, pricers and managers with the tools their smaller paper report cousins. Others
cialists are the only members of the business they need to answer their own questions are simplified what-if analysis tools that
team that know how to query a database around pricing. The epiphany strikes every quickly show the profitability of a price in-
and can write Visual Basic code. To some, level of management and it begins with a crease. Still more spreadsheets become
they seem to wield almost magical pow- proliferation of spreadsheets. quoting tools, allowing managers to set
ers. Typically, a request comes from senior
management to join two very different data
sets in order to find an answer to an old
pricing problem. If I match our internal cost
tables, customer databases and everyday
transactions, can I determine those cus-
tomers that cost the most to serve, element
by element? Which West Coast custom-
ers get the biggest discounts? What is the
year-over-year volume growth for my high-
est selling products? With this information,
managers and executives are armed with
the power to actually do something targeted
and extremely valuable – take immediate
action. As quickly as the need for the data
guru’s expertise emerges, the challenge of
manual report gathering soon appears.
As the backlog of requests grows, data
gurus begin to realize the enormity of their
obligations. Ad-hoc reporting will not scale.
The bliss of being one of the most important
: G : ER M B < L      u      L N F F > K  + ) ) 2
I K B < B G @  > OH E N M B H G
hard floors of profitability. An exciting new
stage in pricing has begun as the entire
organization begins to wield the power of
the data guru. There’s only one problem
– the same problem of scale we see in
the creation of paper reports applies di-
rectly to spreadsheets.
I recently worked with a very forward-
thinking manufacturer with an entire
team dedicated to pricing. They are an
extremely intelligent group that knows
their business down to the stock keeping
unit (SKU) and customer ID level. After
three days of intensive discussions, we
finally understood the phenomenally rich
language they use to describe their prod-
ucts, customers and pricing practices,
including the tools they use to manage
pricing. This manufacturer used no less
than 34 spreadsheets to manage pricing.
Needless to say, it’s not easy to maintain
34 spreadsheets. Many were rock solid.
Some had data problems. Most were
quite useful but lacked the key informa-
tion to direct a fully informed decision.
A standard workflow for this manufac-
turer had the sales agent and pricer in
and out of more than 10 spreadsheets.
Some analysts barely used any of the
tools while others relied heavily on all but
a few. The gurus could not keep up. Us-
ers began forming customized versions,
halting company-wide innovation in its
**       u      : G : ER M B < L F : @ : S B G > ' < H F
tracks and, these spreadsheets had to
be maintained, thus increasing the need
for Information Technology support.
As information technology specialists
and data gurus continue to bear the in-
creasing burden of spreadsheet and data
warehouse maintenance, their inboxes
swell, inhibiting their ability to create
useful and timely information in a scal-
able way. An outward view of capabilities
provided by the software marketplace
quickly leads technology specialists to
business intelligence tools and analytics
software – the ability to view all pertinent
sales and cost information in one, acces-
sible software program.
&97-2)77-28)00-+)2')83307
THE MARKET HAS SPOKEN about the power
of business intelligence (BI) tools. Some of
the largest software companies in the world
paid huge sums of money to acquire this
technology. Users of BI software are en-
abled to take the next step on our evolution-
ary path – institutionalizing company data
into day-to-day pricing decisions available
through a simple user interface. Moreover,
these systems allow several data sources
to be unified under one roof, giving users a
single point of inquiry. Want to know which
customers who spend under a million dol-
lars a year give you the lowest margin? No
problem. Which products have seen their
I K B < B G @  > OH E N M B H G
volumes drop by more than 6 percent over
the past six months? Easy. Negotiating a new
contract and need to know if this customer
has met their volume commitments? Done.
The power of BI tools is obvious – if you have
a specific question that needs answering, just
log on and find out. No need to flip through
pages of reports or spreadsheets that may
or may not tell you what you need to know.
You can take the data guru’s mobile phone
number off your speed dial. The information
is now at your fingertips.
In the early stages, the BI tool is a sav-
ior. The benefits are measurable. The costs
of maintenance are manageable and scal-
able. However, as the BI tool becomes a
natural part of daily life in an organization,
the low-hanging fruit gets harvested. After
a while, analysts have already looked at the
data every way they know how – the prob-
lem of timely information and granularity no
longer exists. A whole new set of concerns
set in. I hear it over and over again: “I don’t
know what I don’t know.” It’s a funny tautolo-
gy that rings loudly from pricing executives.
“We can analyze our transactional data
in almost every way possible, but I simply
don’t know the right questions to ask.”
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
*+       u      : G : ER M B < L F : @ : S B G > ' < H F
With the immense power that BI tools
provide, they cannot find new pricing levers
by themselves. Alerts have been set. Stan-
dard workflows are in place. Yet, the power to
uncover new opportunities eludes the most
diligent and data-driven managers. Further,
markets begin to change. “We can no longer
dramatically improve pricing by simply look-
ing at the standard workflows in the BI tool.
There are too many variables in play. Every
time we find an outlier in how we price, it is
explained by an unforeseen wrinkle in our
markets. Is there a tool in the marketplace
that can mine our data, take into account
market trends and eliminate the excuses of
price deviation?” You bet there is.
46-'-2+7'-)2')
WITHOUT A DOUBT, the best part of my
job is seeing company after company
use pricing science with amazing suc-
cess. There is no better place to find the
value of science than B2B distribution.
What makes science so applicable in dis-
tribution? Complexity. Distributors pro-
vide a unique value proposition to their
customers – everything you need to run
your business can be held, managed and
bought through them. But with this incred-
ible value proposition comes massive
complexity. Distributors must manage
tens/hundreds of thousands of products
as well as tens/hundreds of thousands of
customers – each combination requiring
its own unique pricing methodology, strat-
egy and reasoning. Let’s list a fraction of
concerns that distributors have:
• Customer A is highly competitive while
Customer B is very loyal.
• Product No. 54 has been
commoditized while No. 54A is offered
only through us.
• Customers in the West are willing
to pay much more than those in
the Southeast.
• High cost items are highly
scrutinized while lower cost items
are often ignored.
• We need to sell our image items
(e.g. milk in a grocery store) at a
very competitive price – if I’m not
competitive on those products, I’ll
lose the business. Furthermore,
these high visibility products vary
for each customer.
These are complex issues that quickly
become problems too onerous for an ana-
lyst and a BI tool to solve. Even more impor-
tantly – am I even looking at our markets the
right way? We could have also wondered
– “do blonde salespeople get better pricing
than other salespeople?” The power of sci-
ence now becomes evident – with 95-per-
cent confidence, I can tell you that hair color
has no impact on price. So, what variables
truly impact pricing for your company?
Statistical algorithms (CHAID, clus-
ter analysis, model selection to name a
few) automatically mine your data and
reveal those key attributes that describe
how pricing differs between customers,
products and buying environments. Opti-
mally combining these key attributes pro-
duces a pricing segmentation – distinct
groupings of customers, products and
environments. Pricing segmentation soft-
ware utilizes the collective knowledge of
your pricers, sales agents, sales manag-
ers, dealers, customers and competitors
to determine the factors that drive pric-
ing behavior. We recently worked with a
distributor that had more than 200,000
distinct pricing segments. After software
implementation, the distributor began
pricing very differently. Large-scale price
increases gave way to a much more gran-
ular process with only those customers
and products that underperformed within
their own micro-segment getting the call
to move in-line with their peers. The re-
sult was an 80 percent jump in profit from
the previous year’s price increase.
In another scenario, an aftermarket auto
parts manufacturer needs to understand
the effect price has on volume. It’s a classic
question that every pricing executive would
love to answer. If I know how price reacts
to volume, then I know precisely what will
happen once that 3 percent price increase
: G : ER M B < L      u      L N F F > K  + ) ) 2
I K B < B G @  > OH E N M B H G
hits the market next month. If I have a
company directive to move more volume, I
know exactly how much cheaper I must go
to achieve that goal and not a penny more.
That’s powerful. There’s only one prob-
lem: If you plot your transactional data with
volume on one axis and price on another
axis, you get complete noise. Understand-
ing this fact is the first step to using elas-
ticity correctly – your data is messy. It has
the collective consciousness of the entire
sales force in it. The hurricane that ripped
through Houston, Texas, last September is
in your data.
The recent banking crisis reveals itself in
every report you’ve been given over the last
six months. But there is good news. These
external market factors can be statistically
identified in your data and then surgically
removed to leave the one last major vari-
able that effects volume – the price. The
transformation is quite remarkable. What
once looked like wild changes in demand
patterns now settle into a much more con-
sistent and understandable pattern. What
remains tells you how price changes im-
pact volume.
Why was this of great import to the auto
parts manufacturer? Take a second to think
about the last year of new car sales. The
volume of auto parts is extremely correlat-
ed to the number of new cars sold. Con-
sequently, volumes are down across the
*,       u      : G : ER M B < L F : @ : S B G > ' < H F
board (except for maintenance parts that
older cars require to keep running). To un-
derstand a price change that was made
this year, we must remove the effect that
new car sales have on volumes. When this
happens, uncertainty moves to clarity as
the true effect of price on volume becomes
transparent. Here, science enters your data
with a handful of external market effects in
its possession. It exits your data with clear
marching orders of where to correct poor
pricing decisions and where to continue
with the positive ones.
Science is not a panacea. Just like all
analytical analyses and software it relies
heavily on data quality and assumptions.
Even still, I can’t help but reiterate its im-
portance on the evolutionary path of pric-
ing excellence – the ability, on its own, to
provide a laundry list of action items that
will generate profits for your company.
'32'097-327
PAPER REPORTS, spreadsheets, databas-
es and business tools all have their place
on the journey toward pricing excellence.
Companies evolve from one step to an-
other for clear reasons – the pain of the
previous rung and the promise of the next.
These first four steps are all improvements
on the same basic concept – give someone
the numbers and they can make great deci-
sions. BI tools are the pinnacle of this slice
of the evolutionary chain. However, BI tools
require you to know the right questions to
ask. Without an intelligent user, BI tools be-
come useless.
Science is different. It’s statistical and
mathematical algorithms mine your data
to give you direct marching orders –while
taking into account several variables and
mining your entire data set: here is a list
of products and corresponding custom-
ers that require a pricing action. Nobody
had to be smart enough to ask the right
questions; it’s already been figured out.
Now that’s powerful. But it certainly isn’t
the end of the evolutionary scale. Smart-
er pricing algorithms are being developed
every day, and the most innovative com-
panies are asking tougher questions as
science becomes an ever-increasing part
of how companies look at pricing. ❙
Dr. Neil Biehn (nbiehn@prospricing.com) is the
Director of Science and Research at PROS, a leading
provider of pricing and margin optimization software
products. At PROS, Biehn specializes in the research,
implementation and innovation of pricing optimization for
B2B Manufacturing and Distribution companies across the
globe. Biehn received his Ph.D. in operations research
from North Carolina State University.
REFERENCES
1. J. Davis, 2008, “The Resilient Enterprise: Leveraging
Information for Profitability and Longevity,” presented
at the Revenue and Profit Management Summit, Walt
Disney World, August 2008.
2. T. Davenport and J. Harris, 2007. “Competing on
Analytics: The New Science of Winning,” Boston,
Mass., Harvard Business School Press.
: G : ER M B < L      u      L N F F > K  + ) ) 2
 >
800+
G R E AT E R T H A N EVER BEFORE

ILOG is now part of IBM. IBM and ILOG have joined forces to bring together the greatest minds in
operations
research experts Operations Research (OR). As the world’s largest OR company, with more than 800 talented and

passionate OR experts and practitioners, IBM and ILOG led many of the major developments in

the field for the last 50 years, and remain dedicated to the advancement of OR through academic

contributions and widespread dissemination of theory and practice.

S T R E N G T H
For 20 years, ILOG offered breakthrough mathematical and constraint programming solvers for

solving the most complex business and research problems across a variety of applications. IBM

ILOG CPLEX is the gold standard to which all solvers are compared and will continue to define the

50
leading edge in the industry as the world’s most trusted optimization software for mission-critical

applications. Today, the historic combination of ILOG and IBM brings together a global, full

service technology company with award-winning OR products, expertise in services and

years of
dedicated support, to better serve our customers with the latest technology and real-world
leadership
solutions. Now with the largest group of experts in the world, the possibilities for the

advancement of OR are endless.

I N N U M B E R S

possibilities
L E A R N M O R E A B O U T O U R PA S T, O U R P R E S E N T A N D O U R B R I G H T F U T U R E AT T H E
2 0 T H I N T E R N AT I O N A L S YM P O S I U M O F MAT H E MAT I C A L P R O G R A M M I N G ( I S M P ) ,
C H I C AGO, I L , AU G U S T 2 3 - 2 8 .

IBM and the IBM logo are trademarks or registered trademarks of the IBM Corporation in the United States, other countries or both. All other company
or product names are registered trademarks or trademarks of their respective companies. © Copyright IBM Corporation 2009. All rights reserved.
F B E B M : K R  : G : ER M B < L
AHPHK@:GBS:MBHGL<B>G<>
:EM>K>=MA>P:KBGBK:J
FHK>MA:GCNLM:<A:G@>H?MKHHIE>O>EL%ÉMA>LNK@>ÊP:L:<A:G@>BGAHP
MA>N'L'FBEBM:KR?NG<MBHGL:L:E>:KGBG@HK@:GBS:MBHG'
;R=HN@E:L:'L:FN>ELHG
BY MOST
M ASSESSMENTS, the
& change in the U.S. approach
in Iraq,
Iraq beginning in late 2006,
has largely
la succeeded, after
previous efforts seemed doomed to fail.
More than just a change of troop levels,
more even than a change of strategy and
resources and a willingness to buy off op-
ponents rather than confronting them, how-
ever, “the surge” was a change in how the
U. S. military functions as a learning orga-
nization. Among the differences are how
*.       u      : G : ER M B < L F : @ : S B G > ' < H F
leaders are selected and evaluated, how
military doctrine connects to civilian activity,
how the military interacts with the civilian
leadership, and how decisions are made.
What the key military leaders did reflects
recent results from organization science
and OR/MS, and the lessons they learned
will, in turn, further advance understanding
of how organizations function. In cognitive
science terms, they didn’t just develop a
different set of rules, they developed a dif-
ferent way of generating rules.
Washington Post Pentagon reporter
Thomas E. Ricks was among the most
prominent and influential journalistic crit-
ics of the U. S. effort through 2006. In “The
Gamble,” a book released early this year,
he related how a number of senior U. S.
commanders in Iraq, especially then-Lt.
Gen. Raymond Odierno, managed to reas-
sess what needed to be done, rewrite their
orders to adapt to the situation and obtain
higher-level support, up to and including
the president. Gen. Odierno is of particular
interest not only because of his central role
in developing the new approach, but also
because he had been one of the most con-
ventional thinkers among the senior com-
manders. The Fourth Infantry Division,
under his command, was widely viewed as
less concerned than many other U. S. units
about collateral damage. They also interned
large numbers of military-aged males with
no plans to release most of them, thereby,
as senior U. S. leaders now see it, helping
to motivate new insurgents.
: G : ER M B < L      u      L N F F > K  + ) ) 2
F B E B M : K R  : G : ER M B < L
2);-()%7%2(79++)78-327
AFTER THE 2006 ELECTIONS, Ricks reports,
President Bush was more receptive than he
had been to new ideas about Iraq. A senior
military leader in Iraq said of Bush, “I think
America’s view of the president, whether
they like him or dislike him, is what they
see of him reading a document at the po-
dium, which isn’t impressive, in my opinion.
In these meetings, he is masterful – good
political insights, good handle on the sub-
ject.” New ideas and suggestions, some
of whose proposers had, in fact, been try-
ing to get through to senior decision-mak-
ers for some time, began to find channels
through which they reached the right audi-
ence. Ricks adds that military leaders he
spoke with were somewhat annoyed at the
late 2008 book by his Post colleague, Bob
Woodward, attributing the new approach
to Washington, D.C., think tank analysts,
in particular the American Enterprise Insti-
tute, and policy-makers. According to his
sources, the analysts in the U.S. followed
the military’s turnaround rather than doing
much to initiate it.
In a TV interview the day “The Gam-
ble” was released, Ricks responded to a
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
*/       u      : G : ER M B < L F : @ : S B G > ' < H F
Photos courtesy of the U.S. Army.
question about the military’s reaction to
his previous book: “Yes, a lot of people
in the Pentagon hated it. But many of the
senior commanders in Iraq not only gave
me long, candid interviews, but also, at the
end, pulled a copy of “Fiasco” [Ricks’ ear-
lier book on the Iraq War] off their bookshelf
and asked me to sign it.”
While the military leaders assert that the
change of approach was essentially their
idea, certain analysts did have substantial
influence on their thinking. As long-time
readers of OR/MS Today (www.lionhrtpub.
com/ORMS.shtml) may recall, Eliot Cohen
influenced the pre-Iraq reorganization of the
U. S. military as he strongly advocated that While the surge was a success, the ultimate outcome of the Iraq War remains in doubt.
civilian leaders more deeply and strongly
challenge the military leaders’ strategic as- cases, outright misstatements. This mes- Islamic culture and countries, wrote critical
sumptions. Harry Summers had challenged sage found an audience, and Ricks stated comments about counterinsurgency that
the Army’s doctrinal emphasis on large that by 2006, McMaster had become “the got him invited to participate substantially in
conventional battles, moving a generation most influential colonel in the U. S. Army.” rewriting the U.S. Counterinsurgency Field
of new commanders toward a focus on John Nagl, who served in Gulf War I, left Manual, shifting the emphasis away from
small-unit tactics and counterinsurgency. the Army as a major to pursue a doctorate combat and large-scale internments and
Hence even at the outset of Operation Iraqi in history at Oxford, producing an unusu- toward stability and support activities and
Freedom, a different way of thinking had at ally influential book on how a military can more selective actions against hard-core
least entered the discussions. succeed or fail in a counterinsurgency set- hostiles.
Another influence, along different lines, ting depending on how well it learns, as
was H. R. McMaster, who wrote a book an organization. A number of reporters, 6)8,-2/-2+&=1-0-8%6=0)%()67
sharply critical of Robert McNamara’s as- of whom perhaps Bob Drogin is the best THUS THE STAGE was set for the rethink-
sessment of the Vietnam War. McMas- known, produced telling accounts of how ing by military leaders to connect with civil-
ter contended that military leaders should the U. S. had been misled by faulty intel- ian analytical support. According to “The
have done far more to challenge the civilian ligence, inadequately reviewed. Austra- Gamble,” when President Bush asked se-
leaders’ wrong assumptions and, in some lian Lt. Col. David Kilcullen, an expert on nior military officers who should take over
: G : ER M B < L      u      L N F F > K  + ) ) 2
F B E B M : K R  : G : ER M B < L
command in Iraq, the recommendation of
David Petraeus was virtually unanimous,
and the reason given was, “He’s learned
the fastest and adapted the most in the
commands he’s held so far.” This change
in the culture within the military is a crucial
part of the story. Also important, and not
widely reported, is the role e-mail played
in undermining high-level resistance: by
Ricks’ account, Washington analysts who
wanted the opinion of a senior military of-
ficer in Iraq generally had little trouble ob-
taining it quickly, if he trusted them.
All this did take some time. If Presi-
dent Bush was responsive to dissenting
points of view in late 2006, he had not
been as recently as June of that year, ac-
cording to several sources Ricks quoted.
As one source close to Ricks commented
to this reporter, “Bush did figure out that
he needed to be Commander-in-Chief
rather than Cheerleader-in-Chief – unfor-
tunately, four years too late.”
The consequences in Iraq are encour-
aging but not certain. Anthony Cordesmann
of the Center for Strategic and International
Studies, whose assessment (on which the
OR/MS Today article relied heavily) of the
Iraq war in 2003 proved impressively ac-
curate, notes in new reports that the insur-
gents are still committing numerous acts
of violence. He cautions that the dramatic
decline in violence in Iraq could represent
*0       u      : G : ER M B < L F : @ : S B G > ' < H F
either a genuine, lasting realignment or a Security (CNAS). John Nagl is the presi- -6%5(-**)67*631%*+,%2-78%2
decision by the insurgents to wait until the dent of CNAS. On June 11, CNAS held IN HIS JUNE 11 PRESENTATION, Gen. Petra-
U. S. pulls out, exposing the country to re- a day-long symposium in Washington, eus pointed out that the situation in Iraq is
newed risks of political and economic fail- D.C., titled “Striking a Balance: A New different from Afghanistan. The latter has far
ure. He summarized, “The present range American Security.” David Petraeus was less infrastructure, fewer places troops can
of attacks is not a reason to assume Iraq the keynote speaker. CNAS has studies establish bases to be close to the people,
is failing or reverting to civil conflict. It s underway to continue re-evaluating U. S. fewer potential recruits for their own security
a clear warning that it can do so unless command decision processes and struc- forces, more difficult terrain for ground mobile
Iraqis do move towards political accom- ture, as well as reexamining best prac- forces. His view, however, is that the lessons
modation, unless there is more rapid eco- tices for integrated counterinsurgency, learned in terms of doctrine and command
nomic progress and Iraqi forces develop in stability and security, and political and structure are the important ones, and they
ways that encourage political unity. They economic development efforts. do carry over. What we did in Fallujah didn’t
are also a warning that the U. S. still needs
to focus as much on Iraqi stability and se-
curity as on leaving Iraq.” Other analysts
agree: how disengagement is handled will
most likely have far more important and
lasting effects than almost any aspect of
the engagement. Some, like retired Gen.
Jack Keane, the former Army Vice Chief
of Staff who was an early advocate and ar-
chitect of the “surge,” are more optimistic;
Tom Ricks is less so. As Ricks concluded
in “The Gamble,” “The events for which
the Iraq War will be remembered probably
have not yet happened.”
What is perhaps even more interesting
is the consequences in the United States.
Ricks retired from the Post around the
time he completed the book and is now
a Senior Fellow at a relatively new but al-
ready highly prominent Washington, D.C.,
think tank, the Center for a New American
: G : ER M B < L      u      L N F F > K  + ) ) 2
F B E B M : K R  : G : ER M B < L
Photos courtesy of the U.S. Army.

commanders continue to phenomenon Ricks documented. Key Douglas A. Samuelson (samuelsondoug@yahoo.

learn and adapt.
Ricks wrote of how
Gen. Odierno, as one of
his earliest actions when
he returned to Iraq in 2006
as a corps commander, in-
vited Emma Sky, a female
British peace activist, to
participate in his command
players like Gens. Petraeus and Odier-
no and Ambassador Ryan Crocker have
stated that they are unlikely to write about
their experiences, as doing so could cre-
ate discomfort for them and for people
who were close to them. “If they stick to
that,” Ricks wrote, “one of the oddities of
the Iraq war will be that the officials who
failed – L. Paul Bremer, Lt. Gen. Ricardo
com) is president and chief scientist of InfoLogix, Inc.,
a consulting firm in Annandale, Va. He has worked
extensively on national security decision-making, military
manpower planning and distribution, and modeling
and analyzing mass egress, evacuations, emergency
preparedness and response, mostly for Homeland
Security Institute and Serco-NA.
REFERENCES
1. Center for a New American Security, www.cnas.org
2. Cohen, Eliot, 2002, “Supreme Command: Soldiers,
Statesmen, and Leadership in Wartime,” Free Press.

council meetings as his po- Sanchez, even Lt. Col. Nate Sassaman 3. Cordesmann, Anthony H., 2009, “Iraq: Trends in
Violence and Civilian Casualties: 2005-2009,” Center
litical advisor. “Odierno, by – will leave behind memoirs while those for Strategic and International Studies, May 2009, www.
csis.org
bringing me in, has prob- who were more successful remain offi-
4. Cordesmann, Anthony H., 2008, “The Iraq War: Key
ably brought in the most cially silent.” Trends and Developments,” Center for Strategic and
International Studies, July 2008, www.csis.org
opposite person he could Even without a memoir, though, Gen. 5. Drogin, Bob, Curveball: Spies, Lies, and the Con Man
find,” she told Ricks, add- Petraeus has shared some valuable in- Who Caused a War, Random House, 2007
6. McMaster, H.R., 1997, “Dereliction of Duty: Johnson,
Gen. D
G David
id P
Petraeus:
t “l
“learned
d th
the ffastest,
t t adapted
d t d th
the most.”
t” ing that she had surprised sights. He emphasized the change of cul- McNamara, the Joint Chiefs of Staff, and the Lies That
herself by taking the job. ture, to promote initiative and learning, in Led to Vietnam,” HarperCollins.
7. McNamara, Robert, 1995, “In Retrospect: The
work in Sadr City, he explained, and what we Whether or not Gen. Petraeus and his asso- his presentation. He noted that a flat com- Tragedy and Lessons of Vietnam,” Times Books.
did in Sadr City won’t work in rural Afghani- ciates had read the recent organizational sci- mand structure is essential, in his view, 8. Nagl, John, 2002, “Learning to Eat Soup with a
Knife: Counterinsurgency Lessons from Malaya and
stan. He does have confidence, however, ence literature or not, they have consistently not only for operations but also in dealing Vietnam,” Praeger.
that the “full-spectrum operations” concept, relied on one of its most important findings: with the press: “we have to get the truth out 9. Page, Scott, 2007, “The Difference: How the Power
of Diversity Creates Better Groups, Firms, Schools and
integrating offense, defense, stability and in group decision-making, diversity trumps there first,” which means people have to be Societies,” Princeton University Press.

security operations, respecting and exploit- ability. Too many experts with the same point authorized to respond to inquiries quickly 10. Ricks, Thomas, 2006, “Fiasco: The American
Military Adventure in Iraq, 2003-2005,” Penguin.
ing the difference between generally disaf- of view have the same blind spots. Cultivat- and candidly. He told of seeing a sign a U. 11. Ricks, Thomas, 2009, “The Gamble: General David
fected people and irreconcilable hostiles, ing different points of view and creating a S. commander in Baghdad had posted at Petraeus and the American Military Adventure in Iraq,
2006-2008, Penguin.
and cultivating and coaching civilian political culture of cooperation and initiative seem, in his headquarters. He liked it so much that, 12. Ricks, Thomas, interview on “The Daily Show with
and economic leadership, will succeed, as this context, to have proven to be keys to in a somewhat unusual exercise of the ar- Jon Stewart,” Feb. 10, 2009.
13. Samuelson, Douglas A., 2003, “The Netwar in Iraq,”
success. The extent to which this is really bitrary authority of rank, he appropriated it, OR/MS Today, June 2003.

LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
so will be a ripe field for further study; mean-
while, it looks like the advice to follow.
The study of what worked will be
complicated somewhat by yet another
and it now sits in his office at CENTCOM.
It reads: “In the absence of orders of guid-
ance, figure out what they should have
been and execute vigorously.” ❙
14. Summers, Harry, 1984, “On Strategy: A Critical
Analysis of the Vietnam War,” Dell.
15. Wooodward, Bob, 2008. “The War Within: A Secret
White House History 2006-2008,” Simon & Schuster.

*1       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

? E : P  H ?  :O > K : @ > L

[3]. To address the shortcomings of av-
erage cases analysis, the authors called
for “a shift in information management,
from single numbers to probability distri-
butions.” In addition, the article listed the
Given the rapid pace of developments,
an update seems warranted. In the re-
mainder of this article, we mention some
successful implementation, list available
and announced tools (software and hard-

IKH;:;BEBMR technological, informational and manage-

rial foundations of the field, such as inter-
active simulation, stochastic libraries in
ware) enablers, and lay out the elements
of an implementation methodology.

8,)4633* 3*8,)49((-2+

F:G:@>F>GM
the form of SIPs and SLURPs, and intro-
duced the idea of distributions certified by
a “Chief Probability Officer.”
The second article expanded on in-
PROBABILITY MANAGEMENT has helped
Shell Oil’s Exploration and Production
(E&P) division move the energy giant

BG:<MBHG teractive simulation and on the replace-

ment of classical probability distributions
by stochastic libraries. At the simplest
level, stochastic libraries are pre-gen-
erated random trials that “approximate”
;RL:GC:RL:B@:E stochastic inputs. But as likely, they could
be outputs of simulation and optimization
HANDLIN
ANDLING UNCERTAINTY repre- and end-users, it is difficult to reanalyze models. The magic in their simplicity is
from “a highly decentralized business
with regional allegiances and reward
systems to a single centralized orga-
nization managing a large portfolio of
exploration opportunities” [1]. Critical to
this reorientation, from sequential (best
first) project selection to global portfolio
management, is an “exploration cock-

, sents the most significant

need in converting quanti-
tative insight into profitable
action.
ti Yet,
Y t explicitly
li incorporating un-
certainty into analytics continues to pres-
ent theoretical and practical challenges.
Because computational results are tied
so closely to model inputs, stochastic ap-
proaches require the modeler to make
strong, sometime unwarranted, assump-
tions on key inputs such as probability
distributions. When such assumptions
are challenged by subject-matter experts
and regroup because stochastic methods
are typically computationally intensive.
Two OR/MS Today articles [1] [2] by
Savage, Scholtes and Zweidler intro-
duced Probability Management, a new
analytic framework for fast and robust
stochastic analysis, in 2006. The first
article described the key motivation for
explicitly integrating uncertainty into deci-
sion-making: the outcome at the expect-
ed (or average) input is unlikely to be the
expected outcome, which Savage clev-
erly rebranded as “the Flaw of Averages”
that unlike distribution formulas, uncer-
tainties expressed as stochastic libraries
are additive!
Probability Management has made a
great deal of progress in the three years
since the two articles were published. In
addition to numerous projects that have
profitably leveraged its concepts, soft-
ware vendors have begun to provide na-
tive support for stochastic libraries, and
quite impressively, the user community
has begun to frame a best practices-
based implementation methodology.
pit” that incorporates simulation and op-
timization, as well as newly developed
Probability Management tools such as
“distribution strings” (DISTs), a modu-
larized data format. (More on DISTs
later.) Since it was developed in 2003,
the model has undergone multiple up-
grades while continuously supporting
Shell’s worldwide exploration portfolio.
Probability Management posits the
standardization of common uncertainties
within an organization through stochastic
libraries [1]. The Olin Corporation, a major

*2       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

I K H ;: ; B E B M R  F : G :@ > F > G M
manufacturer of copper alloys, chemicals
and other products, used this idea to co-
ordinate otherwise unsynchronized func-
tions of the company [8]. In one case, an
Olin plant was frequently unable to meet
customer demand, even though separate
models for production and logistics had
showed that capacity was more than ad-
equate. A coupled stochastic model re-
vealed how the interaction worked to limit
the capability of the system as a whole,
well below expectations.
Using interactive simulation that en-
coded how line personnel actually ran
their respective silos, the Probability
Management-savvy staff helped formu-
late organic policies that everyone could
buy into. As a result, stochastic analysis
is replacing average-based “steady state”
calculations as the default at Olin.
The biotechnology giant Genentech
has used Probability Management to
address strategic supply chain planning
questions. Genentech’s first project
concerned the management of manu-
facturing assets [4], where interactive
simulation was used to help decision-
makers evaluate and derive policies that
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
+ )       u      : G : ER M B < L F : @ : S B G > ' < H F
simultaneously lower cost and delivery
risk. More recent work has leveraged
the company’s ongoing improvements
to its product pipeline management sys-
tem, and has involved using DISTs to
make long-range forecasting faster and
more accurate.
The effectiveness of Shell’s E&P port-
folio optimizer prompted the pharma-
ceutical giant Merck to adopt Probability
Management in its product portfolio plan-
ning, going so far as to bring on board
the executive responsible for Shell’s suc-
cess. Today, the use of interactive simu-
lation allows Merck’s decision-makers to
ask and answer critical questions in real-
time [5]:
• How does my opportunity contribute
to the portfolio?
• Is the portfolio balanced with regard
to the company’s overall priorities?
• Are we making the best possible
trade-offs?
• Is the overall risk of missing our
aspirations tolerable?
The projects and companies men-
tioned above are not a complete listing
of recent successes in applying Prob-
ability Management, merely ones that
have been publicly documented. How-
ever they constitute a body of evidence
that Probability Management is more
than just a theoretical framework. Next,
let’s examine the technology artifacts
that support its use.
8)',2303+=7944368
STOCHASTIC LIBRARIES CONTAIN vectors
of simulation trial data. That introduces
platform-dependence: a data vector in
Crystal Ball is not the same as a vec-
tor in Excel. To facilitate modularity,
Sam Savage, consulting professor of
management science and engineering
at Stanford University and a co- author
of the original Probability Management
articles in OR/MS Today, proposed a
new data type – the DIST. As mentioned
earlier, a DIST is an encapsulation of
a data vector (equivalently, a stochas-
tic information packet, or SIP) in a more
compact and transportable form. Any
DIST conforming to the 1.0 open stan-
dard can be stored in a single spread-
sheet cell, irrespective of the number of
trials it encodes. Further, computations
that use DISTs as inputs are faster, es-
pecially within spreadsheets. Savage
has aptly described the DIST as moving
simulation from the era of the six-shoot-
er to the Gatling gun.
DIST-related functions are now avail-
able in Excel for users of Risk Solver
from Frontline Systems. As of this writing,
Frontline Systems distributes the only
fully Probability Management-compliant
software for spreadsheets, supporting
both interactive simulation and the DIST
1.0 standard. DIST support will also be
available in AnalyCorp’s forthcoming up-
grade of its Excel add-in, XLSim.
The traditional granddaddy of spread-
sheet-based simulation, Crystal Ball, is
now an Oracle product. Crystal Ball has
been a pioneer in supporting libraries of
probability distributions, and its leader-
ship has actively influenced the develop-
ment of the DIST 1.0 standard.
Because DISTs will soon become eas-
ily translatable with an add-in from Analy-
Corp., most simulation packages should
be able to support the format relatively
easily. In fact an early application at Mer-
ck read DISTs into and out of @RISK, the
other big player in Monte Carlo for Excel.
Specialized simulation software ven-
dors have also begun to support the
Probability Management effort, as in a
real sense they were doing it before the
movement began. These include Ana-
lytica from Lumina Systems, which in
effect has used the stochastic library
approach since the late 1970s, and
Vanguard Software.
With its Enterprise focus, Vanguard
has long championed collaborative,
modular modeling, and was an early
adopter of Probability Management [6].
That focus allows Vanguard’s software
: G : ER M B < L      u      L N F F > K  + ) ) 2
I K H ;: ; B E B M R  F : G :@ > F > G M
to address more complex applications
that desktop analysis tools cannot, for
instance, estimating manufacturing
cost impacts of new product designs
at a $19 billion aerospace giant. At
the same time, the software can in-
tegrate spreadsheet models as cal-
culation engines. So, for example, a
simulation written in Excel by a finan-
cial analyst can be functionally em-
bedded in a more complex Vanguard
model. (Quite remarkably, a recent
survey found that only one in every
four simulation packages provides
any Excel support whatsoever [7].)
Vanguard also enables the creation
and free transfer of stochastic libraries
as SIPs and SLURPs, even extending
the original SIP/SLURP definitions to
include a time dimension. However,
where Vanguard really pushes the
boundary is through its grid comput-
ing capabilities. Performance speeds
up to 250 times faster than spread-
sheets allow Enterprise-level models
to be run (and even modified) interac-
tively over a Web browser.
The widening acceptance of simu-
lation has further highlighted the key
role of data scrubbing. Since this type
of simulation modeling is largely on
the desktop, Enterprise-oriented ETL
tools are not well-suited for validating
+*       u      : G : ER M B < L F : @ : S B G > ' < H F
and cleaning small-bore data. Un-
fortunately, the “garbage in, garbage
out” adage is no less true for the indi-
vidual analyst. The statistical analysis
package JMP, from SAS Institute, has
been successfully deployed for data
preparation in some of the success
stories mentioned in the previous
section. A lightweight Excel-to-JMP
DIST conversion macro, available
from AnalyCorp, makes data scrub-
bing and preparatory analysis espe-
cially simple.
Implementation Lessons
%6)%73*%440-'%8-32
IT IS NOT SURPRISING that so many
successful application of Probability
Management have been associated
with managing project portfolios and
other strategic supply chain issues.
Correlations between stochastic ele-
ments of a portfolio are notoriously
difficult to compute, especially when
the stochastics of the elements them-
selves are not well-understood. That
is why, for instance, financial port-
folio optimization uses the transpar-
ently reductive modeling assumption
of linear correlation. Unlike traditional
Monte Carlo, Probability Management
does not impose extraneous, possibly
unverifiable, relationships between
I K H ;: ; B E B M R  F : G :@ > F > G M
individual uncertainties (or classes of
uncertainties). When causations or cor-
relations are known, they can be mod-
eled explicitly. When not, relationships
encoded in historical data are still sig-
nificant inputs for the simulation engine.
(See [1] for a discussion of “coherent
modeling”.)
13()0-2+)2:-6321)28
MOST OF THE REPORTED applications of
Probability Management have been im-
plemented in Microsoft Excel. However,
almost by definition, many application in
Analytica or the Vanguard System are
examples of Probability Management. As
a modeling environment, the ubiquitous
Excel has a few benefits:
• Most analysts have at least an
intermediate level of facility with
spreadsheets. Even accounting for
the simulation add-in, ramp-up by
end-users is quick. In particular, since
data often resides in spreadsheets,
the experienced consultant can
collaborate with the business subject
matter expert (SME) to quickly
produce “reasonable” prototypes.
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
+ +       u      : G : ER M B < L F : @ : S B G > ' < H F
• As described, many simulation
environments now incorporate the
key tools of Probability Management,
specifically SIPs/SLURPs, and DISTs.
DISTs tremendously speed up Excel-
based models. Some add-ins are
so efficient that tens of thousands
of simulation trials are executed in
the time it takes for a finger to leave
the just-pressed key. There is no lag
whatsoever! For the first time ever,
true interactive simulation is possible
on the standard business desktop.
• Excel’s charting capability enables
interactivity without the need for
“programming”. This is essential for
engaging non-technical decision-
makers, not just for getting buy-in, but
also for knowledge discovery.
Of course, Excel has limitations as an
analytic workbench. These include a re-
strictive two-dimensional data format, a
non-trivial performance overhead, limited
statistical capabilities, etc. Most of those
issues have workarounds already imple-
mented by add-in software vendors such
as Frontline. One should also be open to
the possibility of using specialized soft-
ware (such as JMP) for specific needs (in
this case, for data scrubbing and prepa-
ratory validation).
Excel’s limitations as a development
environment are tougher to overcome.
While its embedded language – VBA
– can be used to build simple custom
decision-support systems (DSS), VBA’s
robustness, performance and platform-
dependence leaves much to be desired.
Further, if the DSS is intended for use in
a complex workflow, access control and
data-sharing can present a challenge.
If the prototype model is expected to
eventually lead to a DSS, it is essential
to evaluate and lay out a development
strategy, possibly using a specialized
simulation environment.
The beauty of the DIST data type
is that eventually, distributions will be
passed around between users of Excel,
Vanguard, Analytica and JMP, just the
way numbers are today. This leads to the
best of both worlds, in which individual
managers can investigate the implica-
tions of distributions generated on more
robust systems in their spreadsheets.
6)*)6)2')()7-+2)0)1)287
DUE TO THE OPEN and generous inter-
action of Probability Management prac-
titioners, the outlines of a reference
architecture are beginning to emerge.
The design is driven by two common
characteristics of problems most suited
to the framework:
1. Many classes of basic stochastic
entities (projects, products,
investments), where elements of each
class are assumed to have similarly
“shaped” uncertainties.
2. Sources of uncertainty can be
global or class-specific. Global
uncertainties, e.g., interest rates,
are typically exogenous, so they do
not lead to causative loops. (In other
words, the results of the simulation
cannot affect global factors.)
Class-specific uncertainties, e.g.,
the availability of a certain raw
material, may impact one or more
entity class. There may even be
inter-relationships, e.g., if the raw
material is internally produced, its
uncertainty may impact the price of
a finished product.
Class-specific simulations are like-
ly to be more detailed and thus, more
complex. Using Probability Manage-
ment, there is usually an opportunity to
identify and re-use models created and
knowledge collected by SMEs, say, a
plant-level planner or a brand manag-
er. Entity classes relate to each other
and to global and class-specific un-
certainties through a meta-model. The
meta-model is typically less detailed.
Cross-functional by definition, it is also
less well-understood, and thus an es-
pecially fruitful area of investigation for
this framework.
: G : ER M B < L      u      L N F F > K  + ) ) 2
I K H ;: ; B E B M R  F : G :@ > F > G M
Analytics developers tend to natu-
rally be more focused on model logic
and less interested in implementing a
presentation layer optimized for collab-
orative decision-making. Yet the end-
user “dashboard” is the most important
piece of an interactive application. The
key design criterion is that the model
should help managers make good deci-
sions, not that it should make the best
decisions. To that end, the design should
rely on visual aids to communicate the
impact of making different choices. In
particular, data tables should be, as far
as possible, hidden.
;,%8µ72)<8#
THANKS TO THE EFFORTS of a group of
committed analytics practitioners, Prob-
ability Management has notched some
high-visibility successes. The combined
efforts of the ProbabilityManagemt.
org committee members, including
software vendors putting ideas within
reach of implementers and academics
and practitioners extending the ideas
themselves, are rapidly advancing the
framework. For example:
• A DIST 2.0 standard revision to
incorporate full stochastic libraries
including a time dimension is under
discussion.
• A software-independent DIST
+ ,       u      : G : ER M B < L F : @ : S B G > ' < H F
translator is under development.
• Many more Probability Management-
inspired projects are getting
underway.
• A reference design and methodology
document is in preparation.
The coming year promises to be
one of those interesting times for
Probability Management. ❙
Sanjay Saigal (ssaigal@intechne.com) is the founder
of and CEO of Intechne, a provider of analytics-based
solutions to ill-structured problems across multiple
industries.
REFERENCES
1. S. Savage, S. Scholtes, and D. Zweidler, 2006,
“Probability Management,” OR/MS Today, February
2006, Vol. 33, No. 1.
2. S. Savage, S. Scholtes, and D. Zweidler, 2006,
“Probability Management, Part 2,” OR/MS Today,
April 2006, Vol. 33, No. 2.
3. S. Savage, 2009, “The Flaw of Averages: Why
We Underestimate Risk in the Face of Uncertainty,”
Wiley.
4. D. Johnson, 2007, “Genetech Case Study: Applied
Probability Management & Interactive Simulation,”
Proceedings of the 2007 INFORMS Conference
on O.R. Practice, Vancouver, Canada, April 2007:
268-270.
5. D. Zweidler, “R&D Strategy to Realization,” 2009,
Proceedings of the 2009 INFORMS Conference on
O.R. Practice, Phoenix, Ariz., April 2009: CD-ROM.
6. R. Suggs and B. Lewis, 2007, “Enterprise
Simulation – A Practical Application in Business
Planning,” International Journal of Simulation, Dec.
2007: 205-209.
7. E. M. O. Abu-Taieh and A. A. R. El Sheikh, 2007,
“Commercial Simulation Packages: A Comparative
Study,” Proceedings of the 2007 Winter Simulation
Conference, Vol. 8, No. 2, July 2007: 66-76.
8. D. Cawlfield, 2007, “Flaw of Averaged at Work
at Organizational Boundaries,” Proceedings of
the 2007 INFORMS Conference on O.R. Practice,
Vancouver, Canada, April 2007: 422-426.
LM K :M > @ B <  I K H ; E > F L

Advances in software have now ex- 46%'8-'%0)<%140)7

panded the basis and breadth of use
of MCDA. More complex problems can
now be modeled directly with model re-
sults being available in a few seconds.
All this has opened a new dimension for
C OGENTUS , a management consult-
ing company specialized in applying
MCDA techniques to business prob-
lems, has been engaged by different
clients to develop their business strat-

FH=>EBG@MA> the use of MCDA in the strategy arena

other than the just mentioned scenario
planning application.
egy. The focus of attention is placed
on choosing among different possible
strategic moves; in other words, the

F:KD>MLI:<>
When combined with advanced soft-
ware technologies, MCDA becomes a
very powerful methodology to represent
the strategic situation of a company in all
key market factors. Using various weight
:G>P=BF>GLBHG?HKMA>NL>H?  sets, it is now possible to capture at the
FNEMB&<KBM>KB:=><BLBHG:G:ERLBL same time the different points of views
from the Board of Directors and Executive
Team and/or simulate numerous value
sets of various customer groups. Fur-
thermore, as computers allow interacting
;RL:GMB:@H<:LMKH with complex models and as the levels
given to the different variables can be
MULT
L I-CR
ULTI CRITERIA DECISION analy- have been developed to solve strategic changed, an infinite potential for scenar-
set of managerial actions and deci-
sions that could enhance business
offerings and opening and capturing
new market spaces.
In this respect, MCDA offers a solid
methodology to model and facilitate
understanding the company’s set of
managerial actions and decision impli-
cations. In practice, a managerial deci-
sion implies an irrevocable allocation
of resources given different options.
Imagine that each option is a company
activity area or industry factor. In this
sense, a managerial decision to allo-

1 sis (MCDA)
(MC has proven to be
a sound technique to solve
complex problems where
numerous available options must be
evaluated against multiple criteria. This
includes, for example, selecting a best
project against many others or choos-
ing the most efficient portfolio of in-
vestments. Apart from those traditional
applications, more recent uses of MCDA
problems, incorporating MCDA in the
evaluation phase of scenario planning
for example. These strategic uses of
MCDA often consist of four basic steps:
identifying the decision-maker’s strategic
objectives; evaluating the performance
of each strategy-scenario on each objec-
tive; weighting the strategic objectives
and; obtaining the overall performance
of each strategy scenario.
io simulation is now available. Enhanced
scenario simulation capabilities facilitate
testing hypothesis, building common un-
derstanding, learning and adapting to
change as creativity is stimulated. In that
sense the new scenario approach devel-
oped here is more about creative thinking
rather than purely normative assessment
of options or future-oriented forecasting
of scenario planning.
cate more resources in one particular
activity or factor will have an impact
on its competitive performance in that
particular field. The implications could
be assessed relative to the position of
other competitors in the same industry
factor. In more general strategic terms,
the impact of the same allocation of re-
sources can also be measured in terms
of the company’s objectives, mission

+-       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

 F H = > E B G @  F : K D > M  L I: < >
investment and con-
sequently the offer-
ing level that buyers
receive across all
key competing fac-
tors. Different con-
sumer values are
simulated by the dif-
ferent size of the ver-
tical lines. Thus, the
same score in two
differently weight-
Figure 1: General graphical representation of competitive relative ed factors does not
positions in the market space. represent the same
opportunity for the
and vision. In other words, multiple- company. The mapping of competitors’
criteria can be used to model the mar- relative positions identifies the avail-
ket space and analyze the company’s able market space (illustrated with
most efficient actions and product port- the red circled area) where any of the
folio [1]. displayed companies could decide to
As shown in Figure 1, using MCDA move and become the leader in that
software it is possible to display in a particular offering.
two-dimension graph a particular in- More generally this type of display
dustry’s competitive performance. The constitutes an interesting tool to stimu-
horizontal dimension identifies the late analysis about the implications of
range of factors the industry competes a strategic move decision. It offers a
on and decides to invest in. The ver- comparative approach of both the di-
tical dimension represents its level of agnosis of the current state of play in
a particular market space and a pos-
LN;L<KB;>MH :G:ERMB<L sible company-wants-to-be scenario
depending on the relative performance
It’s fast, it’s easy and it’s FREE!
impact of a decision (see Figure 2).
Just visit: http://analytics.informs.org/
This approach allows companies to
+.       u      : G : ER M B < L F : @ : S B G > ' < H F
recognize where the competition is cur- On the other hand, considering a totally di-
rently investing, what customers are vergent strategic profile, as well as consider-
receiving from the existing offerings on ing possible trade-offs (i.e. reducing or even
the market and thereby where a par- eliminating the investment in some market
ticular company could be, following a factors), would liberate resources that could
possible strategic move. be used to support a better strategic move
Based on our experience, Figure 2 il- such as strategic differentiation. In sum, if
lustrates what companies tend to do first: the focus is less on fighting for competitive
increase the level of criteria performance advantage and more on strategic innova-
in order to gain a competitive advan- tion, the result is that product differentiation
tage. The final result of this type of de- can be achieved with an overall investment
cision is a general incremental change. that can stay constant (see Figure 3). This
However, the decision to only improve is a fundamental step forward, if one agrees
key industrial factors requires addition- that creating a new type of product offering
al investment and expenses without a and finding a differentiated niche where no
corresponding reduction of costs else- competition exists is the key for success.
where to compensate. Such a non-com- Instead of diffusing its efforts across
pensatory attitude tends to produce an all factors of competition, the aim is for
inefficient resource al-
location policy. At the
same time when the
benchmark is compe-
tition, the result tends
to be competitive imi-
tation resulting from a
lack of strategy inno-
vation. Consequently,
companies’ strategic
profiles tend to con-
verge, resulting in a
lack of product differ-
entiation as illustrated
in Figure 1. Figure 2: A possible strategic move: fighting for competitive advantage.
: G : ER M B < L      u      L N F F > K  + ) ) 2
 F H = > E B G @  F : K D > M  L I: < >

is done within given value weighted benefit scores (Wj*Sij) reducing or eliminating factors in the
conditions and restric- are to be added together and the total general strategy, the company could
tions. Therefore, the value adjusted benefit must be divided generate a more efficient strategic pro-
given level of available by the total cost of implementing the file and move in the direction of the left
resources will imply new business strategy. The benefit-to- (or west) oriented arrow towards the
a strategy cost limita- cost ratio is expressed in the following efficient frontier.
tion. Acknowledging additive aggregation equation: Similarly, by focusing on the par-
those limitations and ticular factors that offer more benefit
fixing a target cost can &)2)*-883'3786%8-3! to the company and deciding to raise
838%0'378 their level or create them if not yet
Figure 3: From strategic imitation to strategic differentiation.
the company to focus only on some key
factors, even though trade-offs between
multiple factors are often hard to make.
Once again MCDA techniques offer a
sound framework to implement such
fundamental trade-off between multi-
ple – and differently valued – industry
key factors. Combining MCDA with the
mappings described above can help.
The necessary trade-offs will result from
systematically assessing the business
strategy in terms of resource allocation,
measuring and always guaranteeing its
efficiency.
833083):%09%8)786%8)+-'
'327-78)2'=)**-'-)2'=
THE STRATEGIC MOVE aims to add value
to the business, but its implementation
+ /       u      : G : ER M B < L F : @ : S B G > ' < H F
help to orient the stra-
tegic analysis, strip-
ping out some costs.
Combining MCDA
and resource alloca-
tion analysis to assess how efficient
a particular strategic profile is implies
evaluating and confronting two funda-
mental criteria: the added value or ben-
efits of the strategic move against its
costs. In particular, the benefit-to-cost
ratio allows putting these elements to-
gether and enables companies to eval-
uate how to maximize the total benefit
given a target cost. Having defined
benefit criteria for each strategic move
( ← , ↑ or → ), benefit scores can then
be given. Subsequently, each score
(let Sij represent the score associated
with the consequence of a strategic
move i on criterion j) must be multi-
plied by the relative value or weight
factor (let Wj represent the weight fac-
tor associated to criterion j); finally all
∑j Wj*Sij
offered by the company, the resulting
All values in this equation can be in- strategic profile will move in the direc-
putted in MCDA software, leaving the tion of the vertical oriented arrow (or
computer to make the calculations and north) towards the efficient frontier.
produce the results. Finally the figure shows a comparison
One important output that the of two graphs where a vertical line is
computer can produce following this fixed in a different total cost level.
methodology is shown in Figure 4. This vertical line represents the tar-
The horizontal axis displays the total get cost of the company. As explained
weighted costs and the vertical axis above, given the resource restrictions
shows the total weighted benefits. As or their availability, the managers of
illustrated, an efficient frontier –repre- the company can fix a target cost.
sented by the top left
borderline of the foot-
ball shape within the
graph – displays all the
possible points that
maximize the benefit
to the company. The
red point represents
the current state of the
strategic profile. This
means that either by Figure 4: Measuring a strategic profile in relation to the most efficient strategic moves.
: G : ER M B < L      u      L N F F > K  + ) ) 2
F H = > E B G @  F : K D > M  L I: < >
Depending on where the target cost
is fixed, and whether or not it is possible
to invest additional resources, different
possible strategic moves emerge (dif-
ferent possible arrows to follow in the
graphs. It is important to say here that
with a simple click the software can dis-
play a precise measure of the strategic
profile in terms of its total benefit score,
its total cost (how much economy is
achieved and/or investment is needed)
to implement an overall strategic move-
ment. Based on those results, the deci-
sion to give priority to some strategic
moves instead of others seems very
straightforward.
To conclude, combining MCDA
with appropriate software tools can
offer considerable help to understand,
model and choose between possible
strategic choices in a market space.
However, it is only a decision aid; the
final strategic decision will depend on
various factors. For example, depend-
ing on resource constrains, as illus-
trated with the target cost (vertical line)
in Figure 4, the available choices will
be limited. Also, once the competitors’
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
+ 0       u      : G : ER M B < L F : @ : S B G > ' < H F
strategic profiles are added into the
graph, resulting relative positions will
give additional insights about some
better and some worse orientations.
In this respect creativity has an impor-
tant role in taking the rest of the deci-
sion, because creativity is needed to
invent a new offer. A fundamental part
of the strategic planning process con-
cerns in particular that aspect of gen-
eration and formulation of alternative
strategic options. Exploring options
before deciding is a great value.
Before deciding a “where do we want
to be” strategy, it is essential to imag-
ine, test and compare many “where we
could be” alternatives. Before launch-
ing the expedition into the unknown, it
is critical to have a rational vision. Once
again, as strategic situations are rep-
resented and tested, combining MCDA
with advanced software technologies
offer powerful means to stimulate cre-
ative thoughts. Confidence in the final
decision is strengthened by every ad-
ditional option considered. ❙
Santiago Castro is a senior analyst with Cogentus
Consulting Limited (www.cogentus.co.uk/) in the U.K.
REFERENCES
1. The work presented here is the result of
combining ideas taken from the textbook “Blue
Ocean Strategy” (Kim and Mauborgne, 2005) with
MCDA techniques and applying such theoretical
approach to practical cases.
? H K  I K : < M B H G > K L  H G ER
L>O>GLBGL
H?LBFNE:MBHG
IK:<MB<>
IK:<MBMBHG>KL%;>:=OBL>=3:OHB=MA>L>FBLM:D>L
HKKBLD=:F:@>MHRHNK<:K>>K'
;RC>KKR;:GDL:G=K:G=:EEK'@B;LHG
SEVEN IS A NUMBER that ap-
7 pears many places in na-
ture, music and religion.
There are seven oceans,
seven continents, seven notes on the
musical scale, seven days of the week
and many other “sevens.” In this article,
we will discuss major mistakes; we call
them the “seven sins of simulation” to
which many simulation practitioners fall
prey. Certainly, there could be more,
and others may find different ones than
those we cite here. We consider the
+1       u      : G : ER M B < L F : @ : S B G > ' < H F
seven major mistakes in this article to
be like the major causes of problems
when conducting a Pareto analysis, i.e.,
80 percent of the problems are attrib-
uted to 20 percent of the causes.
Mistake 1: Focusing more on the
model than on the problem. As it is
said, “it’s possible not to see the forest
for the trees.” Some simulation practi-
tioners get so involved in model build-
ing that they miss the real objective,
which is to solve the client’s problem.
They build an elegant model that has
too many features and too much detail.
This can occur in several ways. One of
those ways is modeling aspects of the
system that contribute little, if anything,
to the solution of the problem. Another
way is adding interface features that
aren’t needed, aren’t requested, but are
nice cosmetically or operationally. An
example would be a Visual Basic inter-
face that insulates the model user from
the model and makes it easy to alter
and run various scenarios. If the client
requests such an interface, then build
it. Otherwise, it can detract from the fo-
cus of the project.
Mistake 2: Providing point estimates.
Simulation models generally have ran-
dom arrivals and random service times.
There are additional sources of random-
ness that are possible such as break-
downs and path choices. Thus, the
output is a random variable. Running
simulation models is the same as con-
ducting a statistical experiment where
there is randomness. If you were trying
to learn the effects of water and fertiliz-
er on the growing of corn, you wouldn’t
simply conduct one trial and report your
result. The same is true of simulation.
Don’t conduct one trial and report the
result because it is highly unlikely that
the real system is going to react exactly
like the single simulation run. We as-
sume that the value of a simulation run
is one of many that come from output
that is normally distributed. You should
conduct many trials and perform a sta-
tistical analysis that includes a confi-
dence interval or a precision interval.
Mistake 3: Not knowing when to
stop. “A simulation model should be
small, but not too small. A simulation
model should be big, but not too big.”
Somewhere between those extremes is
a happy medium. In this “mistake” how-
ever, we are looking at the second part
of the statement. The simulation model
should be big enough (sufficient scope
and detail included) to respond to the
questions asked in an appropriate man-
ner. But, there is no need to continue to
refine the model after that point. Some
simulation practitioners have great dif-
ficulty in stopping the model refinement.
The model becomes larger and larger
and the verification and validation of
the model becomes more and more
difficult.
Think of the modeling effort as follows:
Say that 80 percent of the accuracy can
be obtained with 50 percent of the effort.
Another 10 percent can be gained with
25 percent more of the effort. Another 5
: G : ER M B < L      u      L N F F > K  + ) ) 2
L B F N E :M B H G  L B G L
percent can be gained with the remain-
ing 25 percent of the effort. To go from
95 percent to 99 percent accuracy may
require the same amount of funds that
have already been expended. And, to go
from 99 percent to 99.9 percent requires
an equivalent input of funds. To go from
99.9 percent to 100 percent accuracy re-
quires a vast amount of effort, perhaps,
the replication of reality!
On the other hand, many simulation
practitioners stop the verification and val-
idation effort far too short. Since it’s not
possible to declare a simulation model to
be verified and/or validated, many tests
should be applied to achieve the assur-
ance that the model can be used as a
representation of reality.
Somewhere, it’s time to stop the mod-
eling effort and the verification and vali-
dation process. If the future of the nation
depended on the model, a lot of effort
might be expended. But, that is not usually
required. Ultimate truth is not affordable.
Mistake 4: Reporting what the cli-
ent wants to hear rather than what the
model results say. This can occur in
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
+ 2       u      : G : ER M B < L F : @ : S B G > ' < H F
numerous ways, all of them inappropri- Mistake 5: Lack of understanding statistical methods, but they can’t recall
ate. The simulation practitioner should of statistics. This mistake has lots of the answer now. If they do answer, it is
act as a scientist, gathering informa- roots and lots of consequences. Virtu- likely that they will get it wrong.
tion, analyzing it and reporting it factu- ally all simulation software has output Then, there are more subtle points of
ally. But, in order to please the client, analysis capability. But, many simula- statistics, such as in multiple compari-
to help the client prove a point, or, for tion practitioners don’t take advantage sons of confidence intervals, how do
whatever reason, there are those that if it. The reason is that they don’t under- you know when there is no significant
will distort the model results. stand it! And, they might get a question difference between two of them? Here
It can occur somewhat innocently, from their manager like the following, is a concrete example: There are three
such as the statement made by the “What does a Type I error of 0.05 really speeds for a machine to accomplish a
decision-maker, “Show that we can mean?” They might have been able to task. We call them Speed A, Speed B,
achieve the desired throughput,” with answer that question when they studied and Speed C. The speeds are normally
the simulation practitioner doing what-
ever is needed to the model to make
that dream come true. Or, it can occur
upon using clearly false input such as
the reliability of a piece of equipment
that is reported to have a 99.99 percent
uptime when the simulation practitio-
ner has observed the same piece of
equipment in a down state numerous
times, and knows that the high rate of
uptime is just some advertising ploy.
Or, it can be a political issue where
two managers are arguing about a
process and the simulation practitio-
ner works for one of the managers. In
his or her zeal to curry favor with the
manager, the simulation practitioner
reports whatever outputs bring favor
on the position of his or her manager.
Whatever the cause, deliberate or in-
nocent, it’s a mistake!
: G : ER M B < L      u      L N F F > K  + ) ) 2
L B F N E :M B H G  L B G L

versus using the normal
test? What’s the differ-
ence between a c 2-test
and an F-test? When do
I use a two-tailed test
and when do I use a one-
tailed test? If it’s a one-
tailed test, is it an upper
or lower? Because these
questions exist and the
answer is unclear, many
simulation practitioners
may choose to avoid sta-
tistical methods.
In their clever book, “Freakonomics,”
by Levitt and Dubner (2005), many of
these cause-and-effect relationships
were discussed. For example, it was
shown that the presence of books in a
household is a good indicator of how
children will perform in school. But,
don’t rush out and buy a bookcase and
lots of books if your child is performing
poorly; it won’t help.
Similarly, with simulation we have
confusion with cause and effect. A client
will say something like, “At around 3
p.m. every afternoon, the entire pro-
duction system (we are discussing a
production system here, but we could
be discussing a service system) comes
to a standstill. Can you simulate that?
The answer is, “No, that’s an effect,
we simulate causes.” But, that’s not
entirely true, we simulate breakdowns
and those are effects. It would be way
too difficult to simulate the cause of
a breakdown. For example, we would

di t ib t d with
distributed ith means and d standard
t d d
deviations. Which of these machines
will result in the lowest expected wait-
ing time if the arrivals are exponentially
distributed with a mean of 50 seconds?
You asked for the confidence intervals
for differences in means between Ma-
chines A versus B and Machines C ver-
sus B. How do you know if no statistically
significant differences are present? The
answer is that there is no statistical sig-
nificance if zero is in the interval. Think
about that. Ask your local statistician if
it doesn’t make sense.
There are many questions like this
in statistics. When do we use a t-test
Mistake 6: Confusing
cause and effect. A study
reported
t d on ini the Wall Street Journal
(Sept. 13, 2003) indicated that teenage
girls who use tanning booths are more
likely to smoke and drink. But, do the tan-
ning booths cause the behavior, or do
those teenage girls who are risk takers
(they smoke and drink) also accept the
risk of the tanning booth?
Another study concluded that being
overweight isn’t unhealthy. This was fi-
nally straightened out in Newsweek (Dec.
11, 2006) among other news media. In
the Newsweek article, the question was
whether being thin caused poor health,
or whether poor health caused thinness.
The latter view prevailed.

, )       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

L B F N E :M B H G  L B G L
be simulating the erosion of a shaft
in a machine, erosion of the gears
and so on. So, we take the easy way
out. We get a distribution of time be-
tween failure and time to repair, and
we simulate the effect. The trick is to
know when it is acceptable to simu-
late effects and when we must simu-
late causes.
Mistake 7: Failure to replicate real-
ity. We will explain this mistake using
breakdowns (again, a similar discus-
sion could be carried on for a service
system) as an example. There are
many ways to simulate breakdowns.
And, these are very crucial to real-
ity in a simulation model. Let’s say
that a machine is up for an average
of 48 minutes, and down for an aver-
age of 12 minutes. Now, we provide
some complications. What happens
to the load being processed when the
breakdown occurs? When the break-
down occurs, it could be destroyed, it
could be sent to another machine for
processing, processing might need
to be restarted when the machine
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
,*       u      : G : ER M B < L F : @ : S B G > ' < H F
comes back up, or the load could be
processed with a different amount of
time required. The load could be sent
to another machine with higher prior-
ity, lower priority, or its same priority.
The objective is to replicate reality, so
choose wisely.
'32'097-32
J UST BY AVOIDING these seven mis-
takes of simulation practice does not,
of course, guarantee the practitioner
a path to valid and useful models.
However, we submit that if the prac-
titioner continues to make mistakes
such as those discussed in this ar-
ticle, he or she risks meeting simula-
tion project goals, and that could be
costly, dangerous and a threat to his
or her career. ❙
Jerry Banks (atljerry@earthlink.net) holds the title
Academic Leader at Tecnológico de Monterrey in
Monterrey, México. Previously, he retired from the
faculty of the School of Industrial and Systems
Engineering at Georgia Tech in Atlanta and then he
was Senior Simulation Technology Advisor at Brooks
Automation (now Applied Materials).
Randall Gibson (randall.gibson@me.com) is an
independent consultant based in Solana Beach, Calif.
Previously, he was Principal and Senior Vice President
in the Management and Supply Chain Consulting
group at TranSystems Corporation, where he managed
the Simulation and Analytical Modeling Practice
Area. He was founder and president of Automation
Associates, which was acquired by TranSystems in
2005. He has over 20 years experience in simulation
modeling and consulting.
> = > E F : G  : P: K =

E>LLBLFHK>
?HKAI
A>PE>MM&I:<D:K=>:KGL<HO>M>=:G:ERMB<L
IK:<MB<>IKBS>;RMK:GL?HKFBG@F:G:@>F>GM
H?F:LLBO>IKH=N<MIHKM?HEBH' Hewlett-Packard VP Kathy Chou (front, center) and other members of the Edelman Award-winning team
have plenty to smile about after capturing the “Super Bowl of O.R.”

;RI>M>KAHKG>K

THEY SAY
S THAT variety is the successfully transform its product portfo- judges, was delivered in dramatic fashion reality, beginning several years ago with

8 spice of life, but don’t tell that

to Hewlett-Packard,
He
abundance
abund
whose
of product vari-
lio program and return $500 million over
a three-year period to the bottom line, it
also earned HP the coveted 2009 Edel-
by INFORMS President Don Kleinmuntz
at the 2009 INFORMS Practice Confer-
ence in Phoenix. Robert Bixby served as
mathematics and imagination and what it
might do for HP.”
To put HP’s product portfolio prob-
ety
t bumped
b d up revenues but also gave man Award from the Institute for Opera- master of ceremonies for the Oscars-like lem into perspective, consider these
the world’s largest technology company a tions Research and the Management Edelman Awards gala. numbers: HP generates more than $135
serious case of indigestion that ultimately Sciences (INFORMS) for outstanding “This is not the success of just one per- billion annually from customers in 170
and negatively impacted its bottom line. achievement in operations research. The son or one team,” said Kathy Chou, vice countries by offering tens of thousands
The cure: operations research. announcement of the award winner, cap- president of worldwide commercial sales of products supported by the largest
HP’s ground-breaking use of ana- ping a day-long competition in which six at HP, in accepting the award on behalf supply chain in the industry. You want
lytics, including operations research, finalists from around the world made a of the winning team. “It’s the success of variety? How about 2,000 laser printers
not only enabled the high-tech giant to series of presentations before a panel of many people across HP who made this a and more than 20,000 enterprise servers

,+       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

> = > E F : G  : P: K =
and storage products. Want more? HP
offers more than eight million configure-
to-order combinations in its notebook
and desktop product line alone.
The something-for-everyone approach
drives sales, but at what cost? At what
point does the price of designing, manu-
facturing and introducing yet another new
product, feature or option exceed the ad-
ditional revenue it is likely to generate?
Just as important, what are the costs
associated with too much or too little in-
ventory for such a product, not to men-
tion additional supply chain complexity,
and how does all of that impact customer
satisfaction?
According to Chou, HP didn’t have
good answers to any of those questions
before the Edelman award-winning
work.
“While revenue grew year over year,
our profits were eroded due to un-
planned operational costs,” Chou said
in HP’s formal Edelman presentation.
“As product variety grew, our forecast-
ing accuracy suffered, and we ended
up with excesses of some products
and shortages of others. Our suppliers
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
,,       u      : G : ER M B < L F : @ : S B G > ' < H F
suffered due to our inventory issues and
product design changes. I can person-
ally testify to the pain our customers ex-
perienced because of these availability
challenges.”
Chou would know. In her role as VP
of worldwide commercial sales, she’s
“responsible and on the hook” for driv-
ing sales, margins and operational
efficiency.
Constantly growing product variety to
meet increasing customer needs was the
HP way – after all, the company is noth-
ing if not innovative – but the rising costs
and inefficiency associated with manag-
ing millions of products and configura-
tions “took their toll,” Chou said, “and we Hewlett-Packard VP Kathy Chou receives Edelman Award from INFORMS President Don Kleinmuntz.
had no idea how to solve it.”
Compounding the problem, Chou cycle times. With no fact-based, data- (HP Business Groups, HP Labs and HP
added, was HP’s “organizational di- driven tools, decision-making between Strategic Planning and Modeling) and
vide.” Marketing and sales always want- different parts of the organization was out (individuals from a handful of con-
ed more – more SKUs, more features, time-consuming and complex due to sultancies and universities) and armed
more configurations – and for good rea- these differing goals and objectives.” with analytical methodology – went to
son. Providing every possible product By 2004, HP’s average order cycle work on the problem. Over the next few
choice was considered an obvious way times in North America were nearly years, the team: 1) produced an analyti-
to satisfy more customers and generate twice that of its competition, making it cally driven process for evaluating new
more sales. tough for the company to be competi- products for introduction, 2) created a
Supply chain, on the other hand, al- tive despite its large variety of products. tool for prioritizing existing products
ways wanted less. Less to forecast, less Extensive variety, once considered a in a portfolio, and 3) developed an al-
inventory, less complexity to manage. plus, had become a liability. gorithm that solves the problem many
“The drivers (on the supply chain side) It was then that the Edelman prize- times faster than previous technologies,
were cost control,” Chou said. “Supply winning team – drawn from various thereby advancing the theory and prac-
chain wanted fast and predictable order quarters both within the organization tice of network optimization.
: G : ER M B < L      u      L N F F > K  + ) ) 2
> = > E F : G  : P: K =
The team tackled the product variety
problem from two angles: pre-launch and
post-launch.
“Before we bring a new product, fea-
ture or option to market, we want to
evaluate return on investment in order to
drive the right investment decisions and
maximize profits,” Chou said. To do that,
HP’s Strategic Planning and Modeling
Team (SPaM) developed “complexity
return on investment screening calcula-
tors” that took into account downstream
impacts across the HP product line and
EDELMAN: THE FINAL SIX
Led by Kathy Chou, vice president
of worldwide commercial sales at
Hewlett-Packard, the 2009 Edelman
Award-winning team included Ann
Brecht, Brian Cargille, Russ Chadinha,
Gavin DeNyse, Shailendra Jain, Holger
Mishal, Thomas Olavson, Cookie Pado-
vani, Kurt Sunderbruch, Robert Tarjan,
Julie Ward, Joseph Woods, Bin Zhang
of HP; Jason Amaral of Emeraldwise
LLC; Dirk Beyer of M-Factor; Chris Fry
of Strategic Management Solutions; Qi
Feng of the University of Texas at Aus-
tin; Sesh Raj of DSApps, Inc.; Krishna
Venkatraman of Intuit; and Jing Zhou
of the University of North Carolina at
Charlotte. Their presentation was en-
titled, “HP Transforms Product Port-
folio Management with Operations
Research.”
,-       u      : G : ER M B < L F : @ : S B G > ' < H F
supply chain that were never properly
accounted for before.
Once a product is launched, variety
product management shifts from screening
to managing a product portfolio as sales
data becomes available. To do that, the
Edelman-award winning team developed a
tool called revenue coverage optimization
(RCO) to analyze more systematically the
importance of each new feature or option in
the context of the overall portfolio.
The RCO algorithm and the complex-
ity ROI calculators helped HP improve its
Other finalists included:
CSX Transportation for “CSX Railway
Cashes in on Optimized Equipment
Distribution.” CSX used math model-
ing to create a system for assigning
and repositioning empty cars. The
company claims approximately $2 bil-
lion total savings from car mileage re-
ductions, car management workforce
reduction and capital avoidance, and
notes other qualitative benefits of the
system to the public.
IBM for “Operations Research Im-
proves Sales Productivity at IBM.”
IBM used operations research to help
the company identify new sales op-
portunities and to better allocate sales
resources to the best future revenue-
generating accounts.
operational focus on key products, while
simultaneously reducing the complex-
ity of its product offerings for customers.
For example, HP implemented the RCO
algorithm to rank its Personal Systems
Group offerings based on the interrela-
tionship between products and orders. It
then identified the “core offering,” which
is composed of the most critical products
in each region. This core offering repre-
sented about 30 percent of the ranked
product portfolio. All other products were
classified as HP’s “extended offering.”
Marriott International for the “Group
Pricing Optimizer.” The company’s
operations research-aided system
empowers the sales team with the in-
formation they need to profitably ne-
gotiate the price of proposed group
bookings. The system automates a
complex manual process to maximize
revenue, hotel profitability and the
quality of time spent taking care of
customers. Since its implementation
GPO has been used to contract over $1
billion in group business.
Norske Skog for “Norske Skog Im-
proves Global Profitability using O.R.”
The publication paper industry has
faced declining markets and margins
for several years. The Norwegian-based
company, with plants in 12 countries
Based on these findings, HP ad-
justed its service level for each class
of products. Core offering products are
now stocked in higher inventory levels
and are made available with shorter
lead times, and extended offering prod-
ucts are offered with longer lead times
and are either stocked at lower levels
or not at all. The net result: lower costs,
higher margins and improved customer
service. (Detailed technical accounts
of the HP Edelman-winning work, as
well as the work of all of the other 2009
on four continents, used operations
research to downsize and reduce man-
ufacturing and supply chain costs, po-
tentially $120 million per annum (~3%
of turnover). Thanks to robust analy-
sis, tough decisions were made and
implemented with minimal disruption.
Zara for “Zara Uses Operations Re-
search to Reengineer Its Global Dis-
tribution Process.” The Spanish
clothing manufacturer and retailer,
which achieves Fast Fashion by making
millions of shipments a week to stores
from its central warehouses, used op-
erations research to optimize its distri-
bution process and increase in-season
sales by an estimated 3 percent to 4
percent – in excess of $230 million in
2007 and $350 million in 2008.
: G : ER M B < L      u      L N F F > K  + ) ) 2
> = > E F : G  : P: K =
Edelman finalists, are scheduled to
appear in an upcoming issue of In-
terfaces: www.informs.org/site/Inter-
faces/ .)
The RCO software algorithm was
developed as part of HP Labs’ “analyt-
ics” theme, which applies mathematics
and scientific methodologies to help
decision-making and create better-run
businesses. Analytics is one of eight
major research themes of HP Labs,
which last year refocused its efforts
to address the most complex chal-
lenges facing technology customers
THE ‘SUPER BOWL OF O.R.’
Named in honor of a pioneer of
operations research practice at
the RCA Corporation, the Franz
Edelman Award for Achievement
in Operations Research is con-
sidered the “Super Bowl of O.R”
because it honors the best ap-
plications of operations research
in the world. The nearly eight-
month competition begins with a
call for nominees in the fall. The
nominees are asked to provide a
two-page summary of a practical
application of O.R. that has had a
significant, positive impact on the
company’s operations and bottom
line. A team of verifiers is then
sent forth to make on-site visits to
learn more about the nominated
,.       u      : G : ER M B < L F : @ : S B G > ' < H F
in the next decade. “Smart application
of analytics is becoming increasingly
important to businesses, especially
in the areas of operational efficiency,
risk management and resource plan-
ning,” says Jaap Suermondt, director,
Business Optimization Lab, HP Labs.
“The RCO algorithm is a fantastic
example of an innovation that helps
drive efficiency with our businesses
and our customers.”
In accepting the Edelman Award,
Chou emphasized not only the com-
pany-wide effort in developing elegant
work and verify claims made in the
application process.
The field of nominees is gradually
narrowed down until six finalists are
invited to present their cases before a
panel of judges at the INFORMS Prac-
tice Conference in the spring.
2009 Edelman Committee Chair Srini-
vas Bollapragada of General Electric
served as one of the judges, along with
Peter Bell of the University of Western
Ontario, Terry Harrison of Penn State
University, Russ Labe of Merrill Lynch,
Patricia Neri of Southwest Airlines,
Leon Schwartz of Yeshiva Universi-
ty, Donald (Bob) Smith of Monmouth
University, ManMohan Sodhi of City
University London and Mike Trick of
Carnegie Mellon University.
> = > E F : G  : P: K =
technical solutions to incredibly complex
problems, but also the buy-in and co-
operation of managers and C-level ex-
ecutives and the wisdom and insight of
the award-winning team to engage and
share their vision with those managers
and executives.
“For some of you who have not been a
part of a very large organization like HP,
this might sound strange, but it required
tenacity and skill to bring about major
changes in the processes of a company
LN;L<KB;>MH :G:ERMB<L
It’s fast, it’s easy and it’s FREE!
Just visit: http://analytics.informs.org/
, /       u      : G : ER M B < L F : @ : S B G > ' < H F
of HP’s size,” Chou said.
“In many of our business
[units], project managers
took the tools and turned
them into new processes
and programs that funda-
mentally changed the way
HP manages its product
portfolios and bridged the
organizational divide.”
By most accounts,
the 2009 Edelman event
was perhaps the most
competitive in its 38-year
history, with CSX Trans-
portation, IBM, Marriott
International, Norway’s Norske Skog
and Spain’s Zara (see box) all provid-
ing first-class presentations. When
asked what put HP over the top with the
judges, Randy Robinson, a former ex-
ecutive director of INFORMS and Edel-
man Committee member and frequent
Edelman coach who served as one of
the coaches of this year’s HP team, of-
fered this:
“Certainly the impact of the work on
HP was substantial and pervasive. Sec-
ondly, they had great teamwork. They
were able to handle the trouble that you
so often see in a big company with dif-
ferent departments and competing inter-
ests. They had the magic to overcome
those obstacles and get people to work comments of Shane Robison, executive
together. They received enthusiastic VP and chief strategy and technology of-
support from all levels of the company, ficer at HP. “Innovation is the lifeblood of
including senior management. Third and our company,” he said via video as part
perhaps most important, the technical of HP’s Edelman presentation. “We be-
work was outstanding. It was practical, lieve continuous innovation is just as vital
but part of it involved cutting-edge ad- to our business processes as it is to our
vances and methods. I think all of that products and services. Our work in op-
combined to set them apart from some erations research is relevant to both ar-
very tough competition.” eas and critical to retaining a competitive
HP’s Chou, who congratulated the edge in the marketplace.” ❙
other finalists in her acceptance remarks, Peter Horner (horner@lionhrtpub.com) is the
later said “that just being a part of the Edel- editor of Analytics and OR/MS Today. Barry List
(barr.list@informs.org), the director of communications
man Awards has really raised the visibility for INFORMS, contributed to this article. A version of this
of operations research at HP. We’ve seen article appeared in OR/MS Today. More information about
the Edelman competition can be found online at www.
tremendous improvements thus far, but scienceofbetter.org/Edelman .
after this, I will personally make sure that
O.R. becomes the foun-
dation of any major pro-
cess improvement going
forward. This project has
taken us to the next lev-
el. What has really come
through is senior man-
agement’s involvement
and the understanding of
how important O.R. is …
to how we do business in
the future in a much more
sophisticated and higher
impact way.”
That message is al-
ready clear based on the
: G : ER M B < L      u      L N F F > K  + ) ) 2
> = > E F : G  : P: K =
of HP’s size,” Chou said.
“In many of our business
[units], project managers
took the tools and turned
them into new processes
and programs that funda-
mentally changed the way
HP manages its product
portfolios and bridged the
organizational divide.”
By most accounts,
the 2009 Edelman event
was perhaps the most
To see a video of the announcement and presentation of the 2009 Edelman competitive in its 38-year
Award, click on arrow. history, with CSX Trans-
portation, IBM, Marriott
technical solutions to incredibly complex International, Norway’s Norske Skog
problems, but also the buy-in and co- and Spain’s Zara (see box) all provid-
operation of managers and C-level ex- ing first-class presentations. When
ecutives and the wisdom and insight of asked what put HP over the top with the
the award-winning team to engage and judges, Randy Robinson, a former ex-
share their vision with those managers ecutive director of INFORMS and Edel-
and executives. man Committee member and frequent
“For some of you who have not been a Edelman coach who served as one of
part of a very large organization like HP, the coaches of this year’s HP team, of-
this might sound strange, but it required fered this:
tenacity and skill to bring about major “Certainly the impact of the work on
changes in the processes of a company HP was substantial and pervasive. Sec-
ondly, they had great teamwork. They
LN;L<KB;>MH :G:ERMB<L were able to handle the trouble that you
so often see in a big company with dif-
It’s fast, it’s easy and it’s FREE!
ferent departments and competing inter-
Just visit: http://analytics.informs.org/
ests. They had the magic to overcome
, /       u      : G : ER M B < L F : @ : S B G > ' < H F
those obstacles and get people to work comments of Shane Robison, executive
together. They received enthusiastic VP and chief strategy and technology of-
support from all levels of the company, ficer at HP. “Innovation is the lifeblood of
including senior management. Third and our company,” he said via video as part
perhaps most important, the technical of HP’s Edelman presentation. “We be-
work was outstanding. It was practical, lieve continuous innovation is just as vital
but part of it involved cutting-edge ad- to our business processes as it is to our
vances and methods. I think all of that products and services. Our work in op-
combined to set them apart from some erations research is relevant to both ar-
very tough competition.” eas and critical to retaining a competitive
HP’s Chou, who congratulated the edge in the marketplace.” ❙
other finalists in her acceptance remarks, Peter Horner (horner@lionhrtpub.com) is the
later said “that just being a part of the Edel- editor of Analytics and OR/MS Today. Barry List
(barr.list@informs.org), the director of communications
man Awards has really raised the visibility for INFORMS, contributed to this article. A version of this
of operations research at HP. We’ve seen article appeared in OR/MS Today. More information about
the Edelman competition can be found online at www.
tremendous improvements thus far, but scienceofbetter.org/Edelman .
after this, I will personally make sure that
O.R. becomes the foun-
dation of any major pro-
cess improvement going
forward. This project has
taken us to the next lev-
el. What has really come
through is senior man-
agement’s involvement
and the understanding of
how important O.R. is …
to how we do business in
the future in a much more
sophisticated and higher
impact way.”
That message is al- To see a video summary of Hewlett-Packard’s Edelman Award-winning work,
ready clear based on the click on arrow.
: G : ER M B < L      u      L N F F > K  + ) ) 2
E d e l m a n A wa r d
of HP’s size,” Chou said.
“In many of our business
[units], project managers
took the tools and turned
them into new processes
and programs that funda-
mentally changed the way
HP manages its product
portfolios and bridged the
organizational divide.”
By most accounts,
the 2009 Edelman event
was perhaps the most
To see a video of the announcement and presentation of the 2009 Edelman competitive in its 38-year
Award, click on arrow. history, with CSX Trans-
portation, IBM, Marriott
technical solutions to incredibly complex International, Norway’s Norske Skog
problems, but also the buy-in and co- and Spain’s Zara (see box) all provid-
operation of managers and C-level ex- ing first-class presentations. When
ecutives and the wisdom and insight of asked what put HP over the top with the
the award-winning team to engage and judges, Randy Robinson, a former ex-
share their vision with those managers ecutive director of INFORMS and Edel-
and executives. man Committee member and frequent
“For some of you who have not been a Edelman coach who served as one of
part of a very large organization like HP, the coaches of this year’s HP team, of-
this might sound strange, but it required fered this:
tenacity and skill to bring about major “Certainly the impact of the work on
changes in the processes of a company HP was substantial and pervasive. Sec-
ondly, they had great teamwork. They
Subscribe to Analytics were able to handle the trouble that you
so often see in a big company with dif-
It’s fast, it’s easy and it’s FREE!
ferent departments and competing inter-
Just visit: http://analytics.informs.org/
ests. They had the magic to overcome
36 | a n a ly t i c s m a g a z i n e . c o m
those obstacles and get people to work comments of Shane Robison, executive
together. They received enthusiastic VP and chief strategy and technology of-
support from all levels of the company, ficer at HP. “Innovation is the lifeblood of
including senior management. Third and our company,” he said via video as part
perhaps most important, the technical of HP’s Edelman presentation. “We be-
work was outstanding. It was practical, lieve continuous innovation is just as vital
but part of it involved cutting-edge ad- to our business processes as it is to our
vances and methods. I think all of that products and services. Our work in op-
combined to set them apart from some erations research is relevant to both ar-
very tough competition.” eas and critical to retaining a competitive
HP’s Chou, who congratulated the edge in the marketplace.” ❙
other finalists in her acceptance remarks, Peter Horner (horner@lionhrtpub.com) is the
later said “that just being a part of the Edel- editor of Analytics and OR/MS Today. Barry List
(barr.list@informs.org), the director of communications
man Awards has really raised the visibility for INFORMS, contributed to this article. A version of this
of operations research at HP. We’ve seen article appeared in OR/MS Today. More information about
the Edelman competition can be found online at www.
tremendous improvements thus far, but scienceofbetter.org/Edelman .
after this, I will personally make sure that
O.R. becomes the foun-
dation of any major pro-
cess improvement going
forward. This project has
taken us to the next lev-
el. What has really come
through is senior man-
agement’s involvement
and the understanding of
how important O.R. is …
to how we do business in
the future in a much more
sophisticated and higher
impact way.”
That message is al- To see a video summary of Hewlett-Packard’s Edelman Award-winning work,
ready clear based on the click on arrow.
a n a ly t i c s | Summer 2009
B G G H O: M B O >  H ' K '
BGM>EPBGL
BG?HKFLIKBS>
WHILE THE EDELMAN AWARD
;
T presented the 2009 INFORMS Prize to In-
(see pag
page 32) from the Insti-
tute of Operations Research
and the Management Sci-
ences (INFORMS) each year honors an
outstanding example of analytics and oper-
ations research in practice, the INFORMS
Prize salutes organizations for “sustained
integration of operations research.” The
INFORMS Prize Committee looks for a
variety of applications of O.R. in a single
organization that provides the organization
with a competitive advantage through high-
impact work. The committee is particularly
impressed with organizations that “repeat-
edly apply O.R. in pioneering, varied, novel
and lasting ways.”
Rangananth Nuggehalli of UPS,
the INFORMS Prize Committee Chair,
, 0       u      : G : ER M B < L F : @ : S B G > ' < H F
tel at the INFORMS Practice Conference
in Phoenix earlier this year. Nuggehalli
recognized Intel’s demonstrated record of
using operations research throughout the
company’s strategic, tactical and opera-
tional levels. “The Prize Committee’s task
was challenging, and the quality of all the
submissions we considered was high,”
said Nuggehalli, adding that Intel showed
how companies could drive significant
value and competitive advantage by uti-
lizing O.R. throughout an organization.
For the past two decades, Intel’s deci-
sion technology group has worked behind
the scenes to provide sound recommen-
dations for designing factories, improving
manufacturing, making accurate sales
forecasts and prioritizing the features that
should be introduced during new product
development, he said.
“They have literally saved Intel billions
of dollars,” Intel Chairman Craig Barrett
told the gala crowd. Barrett was joined by
Karl Kempf, who leads the Intel Decision
Technology Group.
The award committee found that Intel
had an impressive track record applying
operations research methods throughout
the many distinctive business areas of
the company.
The 2009 INFORMS Prize Commit-
tee cited the Intel Decision Technologies
Group for putting O.R. inside every facet of Intel Chairman Craig Barrett says the company’s O.R
Intel’s business. “By employing an exten- group has “literally saved Intel billions of dollars.”
sive array of operation research disciplines
and an innovative process to diffuse them,
the Decision Technologies Group impacted
a vast and diverse set of Intel’s functions
such as product design, demand forecast-
ing, factory development, pricing structures,
equipment and material acquisition and
production/inventory/logistics planning,”
the award citation continued. “From tacti-
cal manufacturing operations to strategic
roadmap development, the myriad of op-
erations research applications contributed
more than $2 billion in improved decision-
making. Intel demonstrated the effective-
ness of O.R. techniques by continuing to
Karl Kempf (left), head of the Intel Decision Technology
produce better products at lower prices Group, accepts INFORMS Prize from committee chair
year after year.” ❙ Rangananth Nuggehalli.
: G : ER M B < L      u      L N F F > K  + ) ) 2
M N MH K B : E L %  L > F B G : K L    G > M PH K D B G @

BG?HKFL
<HG?>K>G<>L>M
?HKL:G=B>@H
;RLHA>BE:CHKC:GB

SITUAT
ITUATED
A E ON the famed San Tuesday evening at the famous SeaWorld of MIT, Christopher Tang of UCLA and Ome- well as corporate and university spon-

7 Diego waterfront, the brand

new H
and S
Hilton San Diego hotel
San Diego Convention
Center will be the sites for what promises
to be an exciting and eventful INFORMS
Annual Meeting Oct. 11-14. The confer-
ence them is “INFORMing the Globe.”
The conference will offer plenty of op-
portunities for networking at the Welcome
Reception on Sunday, a multitude of inti-
mate receptions to choose from on Mon-
day evening and the General Reception on
Adventure Park (with a private evening
Shamu Show for INFORMS attendees).
The four days will be packed with ple-
naries, tutorials, research seminars and
networking opportunities to further com-
mon interests. While in San Diego be
sure to avail yourself of the many recre-
ational opportunities in this sunny city on
the Pacific Ocean.
Plenary and keynote speakers include:
QUALCOMM founder Irwin Jacobs, Richard
Larson, Thomas Magnanti and Yossi Sheffi
ga Rho Distinguish Lecturer Karel Lenstra
of the Eindhoven University of Technology.
More than 20 tutorials are also planned.
Many plant tours and excursions are
scheduled, including a tour of Balboa
Park and the famous San Diego Zoo. The
Hilton and Convention Center are within
walking distance of the world-renowned
USS Midway, the beautiful Seaport Vil-
lage and Coronado Island.
Conference organizers thank Leader-
ship Sponsors SmartOps and ILOG, as
sors AIMMS (Paragon Decision Tech-
nology), California State University San
Marcos and The University of Texas at
Dallas.
For the most up-to-date information,
go to http://meetings.informs.org/sandi-
ego09/. For more information about San
Diego and area attractions visit www.
sandiego.org/nav/Visitors.
See you in San Diego! ❙
Soheila Jorjani is the general chair of the 2009
INFORMS Annual Meeting in San Diego.

, 1       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

< H K I H K :M >  I K H ? B E >
:IIEB>=
F:MA>F:MB<L%BG<'
<hglnembg`Õkf]^o^ehilZg]bfie^f^gmlfZma^fZmb\Zefh]^el
_hk^o^krmabg`_khfln[fZkbg^pZk_Zk^mhZ`kb\nemnk^'
COMPANY NAME: APPLIED MATHEMATICS, INC.
HEADQUARTERS: GALES FERRY, CONN.
BUSINESS: CONSULTANCY DEVELOPS, IMPLEMENTS MATHEMATICAL
MODELS AIMED AT SYSTEMS PERFORMANCE.
WEB SITE: WWW.APPLMATH.COM
>O>KR:IIEB>= Applied Mathematics, Inc., is a privately
F:MA>F:MB<LLM:?? held technical consulting firm with offices in
F>F;>KA:L:=>@K>> Gales Ferry, Conn., near the U.S. Naval Sub-
BGF:MA>F:MB<L% marine Base New London. Also close by on
IARLB<LHK<HFINM>K the Thames River are the Electric Boat Com-
L<B>G<>%PBMAMA> pany, a builder of nuclear submarines; the U.S.
F:CHKBMRA:OBG@F'L' Coast Guard Academy and the Coast Guard
Research and Development Center; and Pfiz-
:G=IA'='=>@K>>L'
er Central Research facilities.
Applied Mathematics develops and imple-
ments mathematical models to help our clients
better understand complex systems in order to
improve system performance. Since 1980, we
have completed more than 120 research and
development projects for U.S. and foreign gov-
ernment agencies and industry. Areas of appli-
cations include submarine warfare, search and
;RPBEEB:F tracking, search and rescue, clinical informatics
C';KHPGBG@ and vineyard analytics. Algorithms developed by
Applied Mathematics are in use on U.S. and U.K.
, 2       u      : G : ER M B < L F : @ : S B G > ' < H F
submarines. In addition, we occasionally design,
build and test prototype systems. Examples in-
anti-submarine warfare during the Battle of
the Atlantic in World War II. A good reference
clude a portable launch system for analog torpe-
does, a micro-fluidic actuation system for active
flow control of an airfoil and an environmental
for this fascinating history is “Slide Rules and
Submarines,” by Montgomery Meigs, an histo-
rian and Army general.
sensor for use by a submarine. The role played by operations analysis in
Every Applied Mathematics staff member has World War II was not lost on the U.S. Navy.
a degree in mathematics, physics or computer Throughout the long Cold War between So-
science, with the majority having M.S. and Ph.D. viet and U.S. submarines, Ph.D. analysts from
degrees. Mindful of Isaac Newton’s counsel that “it Applied Mathematics and other organizations
is better to send mathematicians to sea to collect were embedded in Navy fleet ASW and sub-
data than to send data collected at sea to math- marine commands.
ematicians ashore,” our staff have participated Some of the submarine tactics areas we
as test directors and technical observers in more have worked on include submarine search
than 100 operations and tests conducted at sea
on U.S. and U.K. submarines and naval vessels
and military aircraft throughout the world. In addi-
tion, for more than 30 years, our mathematicians
and physicists have been embedded at a subma-
rine squadron working closely with submarine of-
ficers in developing and evaluating tactics.
In our technical reports for our clients we in-
clude all the data and the assumptions made in
the analysis, and we describe the models and al-
gorithms in detail so that the client can obtain a
second opinion should he or she desire. We also
regularly conduct pro bono projects; our staff mem-
bers volunteer their time and services to numer-
ous professional and non-profit organizations.
Following are brief descriptions of three ap-
plication areas in which we have worked for
many years.
Submarine Tactics. The field of operations
research began with operational analysis of Virginia Class ship control station.
: G : ER M B < L      u      L N F F > K  + ) ) 2
< H K I H K :M >  I K H ? B E >

MAK>>MRI>LH? tactics (e.g., choice of speed, track and depth), An ambush, or barrier search, is used when Coast Guard conducts more than 50,000 SAR
L>:K<A>L:K> target tracking using passive sonar, torpedo the track of a target can be predicted. Scorpi- cases annually.
<HGLB=>K>=;R evasion tactics, submarine evasion tactics, ons and snakes use an ambush search. Mathematically, a search consists of a
;BHEH@BLMLÇ optimal allocation of sonar system signal pro- Each species has evolved to search in only probability distribution in space and time of
<KNBL>%L:EM:MHKR cessing and torpedo and cruise missile tactics. one way except the nuclear submarine. A nu- the search object, the detection range of the
:G=:F;NLA' Mathematical methods we have used include clear submarine with its modern long-range sensor(s) and constraints on search effort, ei-
simulation, game theory, control theory, opti- sensors and its unlimited source of energy is ther number of searchers, time or both. The
mization and statistical estimation. unique in that it can and does conduct each of problem is to optimize some measure of effec-
One important assignment for more than 20 these distinct types of searches for its prey – tiveness, such as probability of detecting the
years was the Submarine Search Manual. This ships and submarines. person while still alive subject to constraints
manual gives search speed, track and depth on search effort. Because SAR detection rang-
guidance for a submarine when conducting a Ocean Surface Current Estimation. A core es are usually very short and the ocean is so
search for another submarine. mission of the Coast Guard is search and res- large, it is important to obtain as good an esti-
Search is a basic activity of all organisms – cue (SAR) of people in distress at sea. The mate as possible on where the search object
the search for food and the search for shelter
from predators. The science of how organisms
search for prey is called foraging by biologists.
Three types of searches are considered by bi-
ologists – cruise, saltatory and ambush.
Energy demands – e.g., lack of endurance
and the “prey handling” problem, e.g., what to
do when you catch the prey, which is often a
matter of life and death for both parties – dictate
the type of search that organisms conduct.
A cruise, or constant speed search, maxi-
mizes range at time of detection and is used
when the large sweep rates possible from a
saltatory search are either not available due to
short detection ranges or not necessary due
to a small search area. Sharks and hawks use
a cruise search strategy. A saltatory, or sprint
and drift, search is used when the search area
is large, detection ranges are long and the prey
is not uniformly distributed. The e coli bacteria
and some fish use a saltatory search.

- )       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

< H K I H K :M >  I K H ? B E >

:<<NK:M>:G=MBF>ER may be drifting in the ocean due to the effects grape production in the U.S. is less than 5 per- range of activities including scheduling labor in
?HK><:LMLH?PBG> of wind and surface current. cent of the world production. There are almost the vineyard; procuring cooperage, labor and
@K:I><KHIRB>E=L One method used in large area open ocean 5,000 bonded wineries in the U.S., located in equipment for crush; and developing market-
:K>K>JNBK>=?HKMA> search is to drop self-locating data marker all 50 states. The retail value of wine produced ing strategies. Yield estimates are needed dur-
>??B<B>GMHI>K:MBHG buoys (SLMDB) from an aircraft at the last in the U.S. was $24 billion in 2005. ing the dormant season – late fall – and during
H?:FH=>KGPBG>KR' known position of the search object. These The inability to accurately forecast grape the growing season – at bloom, mid-season
buoys then periodically report their GPS posi- crop yield costs the wine and grape industries and at harvest.
tion via satellite. hundreds of million of dollars a year. In recent The increased use of mechanical harvest-
When measurements are scarce, expensive years, 30 percent to 50 percent errors in yield ers in California, currently more than 60 per-
or difficult to obtain, it is important to maximize estimates have occurred. cent of the acreage, has also increased the
the information obtained from the measure- Accurate and timely forecasts of wine grape demand for more accurate yield estimates in
ments. The more useful the mathematical crop yields are required for the efficient opera- order to better manage harvest.
concept, the more names it has. Kriging, geo- tion of a modern winery. Yield estimates are Yield components from the current grow-
statistics, objective analysis, weighted least used throughout the winery to support a wide ing season such as cluster count, cluster
squares and optimal interpolation all refer to
the same mathematical concept that is used in
this and many other applications.
For the Coast Guard, we developed time-
dependent kriging algorithms to blend the re-
ported GPS positions of the buoys to obtain
a surface current velocity vector field and as-
sociated uncertainty. The velocity field was
then used with the Fokker-Planck equation to
propagate in time the area of uncertainty for
possible target locations. This model resulted
in a significant improvement in the probability
distribution of the search object compared to
the procedures that had been in use.

Grape Crop Estimation. More than 23,000

farms grow grapes in the United States with
close to a million grape-bearing acres and
a total crop value in excess of $3.5 billion in
2005. About 90 percent of the grape crop in the
United States is grown in California. However,

-*       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

< H K I H K :M >  I K H ? B E >

size, berry set and berry weight are used

to estimate yield using statistical methods.
A commonly used method is cluster count-
ing. Cluster counts sampled statistically
prior to bloom are multiplied by an estimate
of cluster weight at harvest to predict yield
per vine.
We have been working with growers and
research viticulturists in California to improve
the sampling and kriging algorithms used to
estimate crop yield.
Using GPS-equipped mechanical harvest-
ers, technicians have recently conducted
precision harvesting experiments on Califor-
nia vineyards. The vineyard is sampled us-
ing a portable Near Infrared Spectrometer to
prepare a quality map of the vineyard prior to
Cabernet Sauvignon grape clusters
clusters, harvest. The mechanical harvester then au-
Napa Valley, Calif. tomatically segregates the fruit into separate
gondolas. We are currently analyzing data
from these experiments.
In summary, the ease with which mas-
sive amounts of data can be automatically
recorded today – and the availability of pow-
erful and inexpensive computer processing
– offers the promise for significant advances
and widespread use of mathematical mod-
eling in areas as diverse as agriculture and
submarine warfare. ❙
William J. Browning is the president of Applied Mathematics,
Inc. (www.applmath.com).

LN;L<KB;>MH:G:ERMB<L
It’s fast, it’s easy and it’s FREE! Just visit: http://analytics.informs.org/

- +       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

E:LMPHK=

<A:G@>
F:G:@>F>GM
MA>HGERK>?>K>G<> Recently I attended a course that cov-
MHMA>G><>LLBMRH? ered the topic of “Change Management.” It
<A:G@>P:LLBFIER is probably a sign of the times that we study
É=HGÌM<A:G@>?HK how people deal with and react to change
MA>L:D>H?<A:G@>'Ê imposed by management. The course pro-
vided many useful insights into common
reactions to change – but paid little atten-
tion to the necessity of the change itself.
Of course change is inevitable, but Change
Management is focused on “management-
imposed” change – not the “Every Day is a
New Day” flavor of change. The only refer-
ence to the necessity of change was simply Consider the following algorithm: optimal solution, but for complex (or most
“don’t change for the sake of change.” Or, Step 1: Select a one-time change real-world) problems, the greedy algorithm
in other words, change only when there is in business practice. often results in suboptimal solutions. Still,
an obvious reason to change. And this state- Step 2: If the stated change can be what are other options?
ment (or perhaps not just a statement, but reasonably expected to move the An intriguing option is Threshold Accepting
an algorithm in disguise) is where my opera- business closer to its goals, do it. [1]. The steps for this would be as follows:
tions research (O.R.) radar perked up. (Read as: there is an obvious reason Step 1: Select a one-time change
To demonstrate what I mean, suppose to change.) in business practice.
there exists some objective way (e.g., a Step 3: Go to Step 1. Step 2: If the stated change seems
model or function) that measures how well reasonable, do it. However, if the
our business processes are meeting our This seems reasonable, and at the most stated change is unreasonable,
goals (market share, profit, etc.). We find basic level, it is – it certainly fits the require- still do it a certain percentage of time.
ourselves, in a very general sense, with an ment of “don’t change for the sake of change.” Step 3: Go to step 1.
optimization problem: which variables under But in terms of an optimization methodology,
our control should be changed and by how we see we have only developed a greedy al- When your typical practitioner looks at the
;R:=:F<E:KD much? These variables include many busi- gorithm. From a purely theoretical point of Threshold Accepting algorithm in the context
ness-oriented issues like “process effective- view we know there are times when greedy of, say, determining a good lumber-harvesting
ness” and “organization.” algorithms perform well, even resulting in an plan, it is easy to see the usefulness. Some

- ,       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2

E:LMPHK=

HIMBFBS:MBHGMA>HKR un-improving moves in the short term are the

LN@@>LMLP>LAHNE= most productive in the long term. However, in
;>FHK>HI>GMH the context of business policy and business
NG&BFIKHOBG@<A:G@>L organization, the attitude changes quickly to
PBMAMA>AHI>H? “don’t change for the sake of change.”
=B:@GHLBG@:G= So here we have an interesting dynamic.
BFIE>F>GMBG@HMA>K When the top-boss enters and announces
FHK>&BFIKHOBG@ a change that seems certain to move the
business away from its goals – our initial
<A:G@>L=HPGMA>KH:='
reaction is… well… one that requires the
manager to take courses in Change Man-
agement. Yet optimization theory suggests
@:FL
we should be more open to un-improving
changes with the hope of diagnosing and
implementing other more-improving chang- ?NEEI:@>:=O>KMBL>F>GM
es down the road. Certainly more thought
needs to be put into the change – as busi-
ness changes can have large impacts on
the people and the bottom line. Still, per-
G>P:KMPHKD
haps there is sound O.R. reasoning not to
scoff too much at “change for the sake of
change.” It could be a step back that allows
I:@>--
us to see the very large step forward. ❙
Adam Clark (adamclark.usa@gmail.com) has worked as an
operations research analyst for the Department of Defense since
earning his master’s degree in operations research from Oregon
State University in 2001.

REFERENCES
1. Dueck, G., Scheuer, T., Sept 1990, “Threshold accepting:
a general-purpose optimization algorithm appearing superior
to simulated annealing,” Journal of Computational Physics,
Vol. 90, No. 1.

LN;L<KB;>MH:G:ERMB<L
It’s fast, it’s easy and it’s FREE! Just visit: http://analytics.informs.org/

- -       u      : G : ER M B < L F : @ : S B G > ' < H F : G : ER M B < L      u      L N F F > K  + ) ) 2