10 Oilfield Review

Portfolio Management for Strategic Growth

Tom AdamsJeff LundKerr-McGee Oil and Gas Corp.Houston, Texas, USA

Jack A. Albers Burlington Resources InternationalHouston, Texas

Michael BackJason McVeanCalgary, Alberta, Canada

John I. Howell III Portfolio Decisions, Inc.Houston, Texas

For help in preparation of this article, thanks to FionaMacmillan, London, England; Graeme Simpson, Gaffney,Cline and Associates, Guildford, Surrey, England; and Jim Thorson, Resource Solutions, Denver, Colorado, USA.Capital Planning and Peep are marks of Schlumberger.TERAS is a mark of Landmark.

The petroleum industry is shifting emphasis from cost-cutting to more diverse asset-

management practices. Portfolio optimization, a fast and effective way of analyzing

and improving overall asset value, is one tool that can help.

Ups and downs in the exploration and production(E&P) business are accepted as facts of life in anindustry known for its risks. How high the highs,how low the lows and when they occur are diffi-cult to predict, but the cyclic nature itself is takenfor granted. During good times, companies grow,invest in riskier ventures and profit if the upturnlasts long enough. During slumps, companiesdivest and implement cost-cutting measures.Given such unpredictable swings, how can thepetroleum industry plan for long-term growth?

Many E&P companies have discovered thevalue of managing their assets as a blended col-lection, or portfolio, taking into account interde-pendence of projects rather than consideringinvestments on a project-by-project basis.1

Common practice is for a company to first rankindividual projects either by net present value(NPV) at a given discount rate or by some othermeasure of worth, then initiate those projectsthat fit the current investment budget—startingwith the best.2 This method assumes that projectsare independent, or have no common factors.

1. Ball BC and Savage SL: “Holistic vs. Hole-istic E&PStrategies,” Journal of Petroleum Technology 51, no. 9(September 1999): 74, 76, 78, 80, 82, 84.Howell JI III, Anderson RN, Boulanger A and Bentz B:“Managing E&P Assets from a Portfolio Perspective,” Oil & Gas Journal 96, no. 48 (November 30, 1998): 54-57.Anderson RN, Amaefule J, Forrest M, Howell JI III,Nelson HR Jr and Rumann HA: “Quantitative Tools LinkPortfolio Management with Use of Technology,” Oil &Gas Journal 96, no. 48 (November 30, 1998): 48-50, 53-54.

2. Net present value represents the difference betweenpresent values of cash outflows over the life of the projectand present values of cash inflows, all discounted at aselected interest rate.

Winter 2000/2001 11

The portfolio-management approach capital-izes on the fact that all projects interact, whetherthey involve exploration, development, produc-tion or acquisition. Factors such as market fluctu-ations, performance targets and technical riskare among the elements that tie one project toanother. Even if there is no apparent technicallink between projects, they interact in the sensethat pursuing one project may prevent anotherfrom being pursued, or one project’s success maycause others to be possible. The portfolio per-spective helps decision-makers throughout the

organization understand how projects interact tosatisfy balanced business requirements.

Portfolio management can be thought of as abridge between a company’s vision, or businessstrategy, and the collection of projects that willbring that strategy to fruition. Corporate strategyand metrics—standards of measurement used toquantify the strategy—along with long-term tar-gets for each metric form the foundation. Forexample, results of the company’s existing basebusiness can be compared to targets for metrics

such as earnings, net cash flow, production andreserves (above).3 Discrepancies between base-business results and targets highlight potentialproblems in business performance that must becorrected to meet the targets. However, optimizinga set of assets while simultaneously satisfyingmultiple competing metrics is not a trivial task.

The next structural element of the bridge isselection of assets to pursue, acquire, divest andreconfigure so that overall, the portfolio willmeet strategic targets (next page, top). Excessesin net cash flow, for example, might be investedto bring reserves, production and earnings in linewith target levels. However, it is not likely thatselection of one isolated project will bring base-business results up to target levels. A subset ofprojects must be selected from what is generallya much larger assortment of possible projects.Potential projects may include exploration opportu-nities, current development and production assets,and full- or part-ownership of new acquisitions,

mergers and trades. As the number of projectopportunities grows, corporate or financial plan-ners are faced with the increasingly difficult taskof choosing projects that best achieve companyobjectives.

This article describes some of the techniquesavailable for analyzing and optimizing asset port-folios, including software and consulting servicesthat help rank investments, select projects andpredict the probability of portfolio success. Thesetechniques can be used on multiple levels by avariety of decision-makers: at the highest level,for developing a business strategy; at a secondlevel, for evaluating investment opportunities;and at a project level, for supporting ongoingbusiness. First we look at how a method calledefficient-frontier analysis, designed for analyzingfinancial-investment portfolios, is being tailoredto suit petroleum-industry problems. Then wepresent case studies showing how two oil com-panies are starting to apply these optimizationmethods to asset-portfolio management.

On the Frontier of Efficient PortfoliosEfficient-frontier analysis considers the balancebetween value and risk in the selection of opti-mal portfolios. Efficient-frontier theory was origi-nally developed about 50 years ago to analyzesecurities portfolios, but it differs in somerespects when applied to petroleum portfolios.4

The original idea states that a portfolio can beworth more or less than the sum of its compo-nent projects and that there is not one best port-folio, but a family of optimal portfolios thatachieve a balance between value and risk. Thesestatements remain at the heart of efficient-frontier theory as it pertains to the oil field.

A portfolio is said to be efficient if no otherportfolio has more value while having the same orless risk, and if no other portfolio has less risk

12 Oilfield Review

1000

800

600

400

0

300

250

200

150

100

50

0

3000

2500

2000

1500

1000

500

0

250

200

150

100

50

0

1999 2001 2003 2005 2007 2009 2011 2013

1999 2001 2003 2005 2007 2009 2011 2013

1999 2001 2003 2005 2007 2009 2011 2013

1999 2001 2003 2005 2007 2009 2011 2013

3500

4000

4500

200

1200

1400

1600

1800

350

400

450

Year Year

Year Year

Mill

ions

of b

arre

ls o

f oil

equi

vale

nt

$, m

illio

ns$,

mill

ions

Mill

ions

of b

arre

ls p

er y

earEarnings Production

Net cash flowReserves

> Metrics, or standards of measurement, and strategic targets for a generic exploration and production (E&P) portfolio.Metrics for this study are earnings, production, net cash flow and reserves. Targeted levels for these metrics are displayedas vertical bars. The purple shaded area represents business as usual, or the level achieved by the base business, over a14-year period. Disparity between base-business results and targets shows where company performance falls short.(Adapted from Howell et al, reference 1.)

Winter 2000/2001 13

while having the same or greater value (right).5 Forthe purpose of this example, value is defined asthe mean net present value of the portfolio, andrisk is defined as the semistandard deviation ofthe portfolio’s possible value. Semistandard devia-tion is a statistical measure of the distribution ofpossible values a portfolio could have, given thatthe portfolio value is uncertain. It is calculated in

1000

800

600

400

0

300

250

200

150

100

50

0

3000

2500

2000

1500

1000

500

0

250

200

150

100

50

0

1999 2001 2003 2005 2007 2009 2011 2013

1999 2001 2003 2005 2007 2009 2011 2013

1999 2001 2003 2005 2007 2009 2011 2013

1999 2001 2003 2005 2007 2009 2011 2013

3500

4000

4500

200

1200

1400

1600

1800

350

400

450

Earnings Production

Net cash flow Reserves

Year Year

Year Year

Mill

ions

of b

arre

ls o

f oil

equi

vale

nt

$, m

illio

ns$,

mill

ions

Mill

ions

of b

arre

ls p

er y

ear

> A portfolio perspective. The pink line shows an optimized portfolio solution that meets strategic targets, filling in gapsleft by the base business. (Adapted from Howell et al, reference 1.)

Risk

Valu

e

Eff ic ient f ront ier

Valu

e

Risk

Eff ic ient f rontier

A

BC

D

E

> Efficient-frontier theory. Oil-industry portfolios plotted by risk and valuedelineate the efficient frontier. A portfolio is efficient if no other portfoliohas more value for the same or less risk, and if no other portfolio has lessrisk for the same or more value. Portfolios B, C, D, E and all blue dots areefficient, while A and other pink dots are not. In the securities-investmentindustry, for which efficient-frontier theory was developed, the efficientfrontier is a continuous line (insert). (Adapted from McVean, reference 4.)

3. Base business is the course of business if current pro-jects are continued but no new projects are undertaken.

4. McVean JR: “The Significance of Risk Definition ofPortfolio Selection,” paper SPE 62966, presented at the2000 SPE Annual Technical Conference and Exhibition,Dallas, Texas, USA, October 1-4, 2000.

5. Markowitz H: Portfolio Selection: Efficient Diversificationof Investments, 2nd ed. Oxford, England: BlackwellPublishing Co., 1991.Bailey W, Couët B, Lamb F, Simpson G and Rose P:“Taking a Calculated Risk,” Oilfield Review 12, no. 3(Autumn 2000): 20-35.

the same way as standard deviation, but only val-ues less than the mean are used (above).

On a plot of value versus risk, the upperboundary of the group of portfolios approachesthe efficient frontier. In the financial-investmentindustry, where each investment can representinfinitesimally small shares in a project, the effi-cient frontier plots as a continuous line. In thepetroleum-asset case, projects are often eitherdone or not done, and so the value-risk space isa collection of points rather than a continuousspace. The efficient frontier itself plots as a setof portfolios rather than a line. In this example,portfolios B, C, D and E are all relatively effi-cient—they lie on or near the frontier—while

portfolio A could become more efficient by reduc-ing its risk or increasing its value, or both.

Several companies and consultants havedeveloped software packages to calculate anddisplay efficient-frontier analysis specifically forE&P portfolios. These include the Perspectivespackage by Portfolio Decisions, Inc. (PDI), CapitalPlanning software by Merak, a Schlumbergercompany, and the Portfolio module of the TERASsoftware by Landmark.6 Merak and PDI havedeveloped a business partnership combining PDI consulting and functionality from thePerspectives package with Merak CapitalPlanning software to create an enhancedportfolio-management process.

In the Merak suite of software, analyzing port-folios in value-risk space follows three steps. First,the set of projects that could potentially beincluded in a portfolio must be collected and eval-uated. Economic evaluations are carried out by theMerak Peep software, an economic engine capa-ble of performing calculations for fiscal regimes allaround the world. Monte Carlo methods are usedto model the uncertainty inherent in each of theseprojects. If required, correlation—such as price—among projects can be set up at this stage.

In the second step, the business strategy isdefined in terms of economic, strategic andphysical requirements for the portfolio.Constraints may be in terms of maximum capitalcost, minimum production, minimum reservesgrowth, or any other strategic metric, and maybe fixed over one or more years of the life of theportfolio. Other factors, such as rig availability,geographic distribution of assets in the corpo-rate strategy and contractual obligations, can beincluded as constraints.

14 Oilfield Review

µ

-2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28

150

125

100

75

50

25

0

Freq

uenc

y

NPV, thousands

σ0.10

0.09

0.08

0.07

0.06

0.05

0.04

0.03

0.02

0.01

0

Prob

ablil

ity

1/2σ

> Standard (�) and semistandard (�1/2) deviations of a typicalportfolio risk profile. For such a probability distribution, whichis skewed about the mean (�), the semistandard deviation isnotably smaller than the standard deviation.

> The effect of different definitions of risk on efficient-frontier analysis. The left plot uses the common E&Pdefinition of risk, semistandard deviation of net present value (NPV). Efficient portfolios are blue dots, and are labeled EP01 through EP56, starting at the bottom. On the right, the same portfolios are plotted, but the definition of risk is the probability of exceeding the capital-spending limit in the first year of portfolio life. Someportfolios that were attractive under the original definitions become less so with the new definitions, and viceversa. Portfolios that were efficient before are again shown as blue dots. (Adapted from McVean, reference 4.)

6. For a list of software and service providers in decision,risk and portfolio management: Thorson J: “OpportunityManagement Resources,” Exploration Business Journal4, no. 3 (Third Quarter, 2000): 14-15.

7. A genetic algorithm can be thought of as a guided randomsearch engine. For more on its use in portfolio optimiza-tion: Fichter DP: “Application of Genetic Algorithms inPortfolio Optimization for the Oil and Gas Industry,”paper SPE 62970, presented at the 2000 SPE AnnualTechnical Conference and Exhibition, Dallas, Texas, USA,October 1-4, 2000.

8. McVean, reference 4.9. Albers JA and Howell JI III: “Portfolio Balancing to

Achieve Long Term Strategic Goals,” presented at theEuroforum International Symposium for StrategicPortfolio Management for Upstream Oil and Gas, London,England, March 22-23, 1999.

Mea

n N

PV, t

hous

ands

300

280

260

EP14EP13EP12

EP05

EP01

EP15

EP19EP21

EP23

EP24

EP27

EP30

EP35EP40

EP47EP50

EP54EP55EP56

320

340

360

380

400

160 170 180 190 200 210 220 230

Semistandard deviation of NPV0

Probability of capital investment exceeding $200,000,000

P13

P12

P10

P11

0.5 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50

Winter 2000/2001 15

The third step combines groups of projects tocreate portfolios, then compares and analyzesthe results. Portfolios can be created manually orthrough a variety of automatic techniques. One ofthese methods is the random portfolio generator,which creates a selection of portfolios that sat-isfy the business strategy. Better portfolios canalso be sought using more intelligent optimizers.For instance, linear programming—featured inthe PDI Perspectives and Landmark TERAS soft-ware—can be used when the description of theproblem and its constraints are linear. Linear pro-gramming delivers optimized solutions for abroad range of business problems. However,some problems require long solution times andmay generate suboptimal results when tackledwith linear programming.

Another optimizer known as the genetic algo-rithm has the power to handle highly nonlinearproblems.7 Because of its robustness, it can bedirected to maximize value or minimize risk forcases in which value and risk can be defined invirtually unlimited ways. With these methods,thousands of projects can be scrutinized andsorted to compile candidate portfolios.

All of these portfolios, however they are gen-erated, can be compared and examined in a vari-ety of ways. The Capital Planning softwareprovides graphical, tabular and data-manage-ment tools for examining and comparing portfo-lios. Some analysts prefer efficient-frontier plotsto evaluate portfolios, while others focus on theprobabilities of meeting metric targets. All areuseful tools for exploring the strengths andweaknesses of different portfolios.

The efficiency of a selected portfolio, or itsposition on a risk-value plot, depends on the defi-nitions of risk and value.8 In the E&P industry, valueis often defined as mean NPV at a specific discountrate, and risk is taken to be the semistandard devi-ation of NPV—representing only the downside inthe variance of NPV, or only the results that are lessthan the mean. A set of portfolios can have a com-pletely different appearance if plotted using differ-ent definitions of risk and value (previous page,bottom). In this example, one plot uses the com-mon E&P definition of risk, and the other quantifiesrisk as the probability of exceeding the capital-spending limit in the first year of portfolio life.Since the cost of each project is uncertain, there isa chance of failing to satisfy this constraint byoverspending in the first year. Under this definitionof risk, some portfolios that previously were inad-equate now appear attractive, and vice versa.

The selection of an optimal portfolio is stronglydependent on which definition of risk is selected.Consequently, it is important to explore multiplerisk definitions in order to more fully understandthe quality of a portfolio, and ultimately makewise decisions about which projects to pursue.

Stepping into Portfolio ManagementIn 1999, Burlington Resources International (BRI),the international division of Burlington Resources,began using the portfolio-management approachto evaluate both its existing properties and newopportunities. After successful implementationin the international division, modern portfolio-management techniques are now being appliedacross the corporation. Historically, decision-makers in the industry would have based oppor-tunity-evaluation decisions on intuition and

experience. However, these insights are subjec-tive and could result in different decisions fromone decision-maker to the next. Under the newapproach, projects are judged based on quantita-tive information about their contribution to thecompany’s long-term strategy and how they inter-act with other projects in the portfolio.9

At BRI, the portfolio model is used in severalways to identify why and how a particular oppor-tunity may be beneficial to the portfolio. Efficientfrontiers of the portfolio with and without thenew opportunity are analyzed to understand itsimpact on total portfolio value, which may behigher than the NPV of the opportunity alone. Theportfolio output is reviewed to determine why thenew opportunity may be valuable to the strategyand to identify any downside risks. Confidencelevels of meeting strategic goals with and with-out the opportunity are studied to evaluate howthe new project affects the likelihood of meetingthose goals. The opportunity is then characterizedfor the decision-maker in terms of its effect on thetotal business performance of the portfolio.

One example of how Burlington Resources hasused portfolio-management tools comes fromevaluation of a specific decision to acquire a pro-duction project. As a starting point, an original,optimized $5.5 billion portfolio is analyzed usingmultiple criteria. Then, economic data for the newproject are added, and the analysis is repeated.

The initial analysis compares targets, base-business values and portfolio values for six met-rics: net income, net cash flow, capital, production,exploration expense and oil reserves (above).Several of these targets were applied as con-straints to the portfolio solution. For example, the

1200

1000

800

600

400

200

01999 2002 2005 2008 2011

1200

1000

800

600

400

200

01999 2002 2005 2008 2011

300

250

200

150

100

50

01999 2002 2005 2008 2011

350300

250

200

100

500

1999 2002 2005 2008 2011

4000

2000

01999 2002 2005 2008 2011

1000

3000

400

200

01999 2002 2005 2008 2011

100

300

150

Year Year

Year Year

Mill

ions

of b

arre

ls o

f oil

equi

vale

nt

$, m

illio

nsM

illio

ns o

f bar

rels

per

yea

rNet income

Production Reserves

Year

Year

Net cash flow Capital

Exploration expense

$, m

illio

ns

$, m

illio

ns

$, m

illio

ns

2013

2013

2013

2013 2013

2013

> Comparison of metrics and targets for a Burlington Resources International (BRI) $5.5 billion portfolio. The base business (purple shading)meets or exceeds only a few of the targets (vertical bars), for example, net income and net cash flow for the years 2002 through 2005. By includingother assets and activities, an optimized portfolio (pink line) can be created that meets many more of the target levels.

net income and net cash flow targets for 2000 and2001 were constraints, and so forced the solutionto equal these values exactly. Similarly, explo-ration expense for 2005 and capital targets for theyears 2001 through 2004 were matched exactly.

The original optimized portfolio is only one ofa set of optimized solutions that can be createdusing the initial set of projects. An efficient fron-tier displays the set of portfolio solutions thatmeet the same performance metrics but have dif-ferent values and risks (above left). This plot,with risk on the vertical axis and value on thehorizontal, contains the same kind of informationas in previously shown efficient frontiers, but issimply transposed (software packages differ intheir presentation style). The $5.5 billion portfolio

was selected as the original test portfoliobecause it has the maximum value for the levelof risk that could be tolerated.

With uncertainty information provided bytechnical specialists, the original portfolio wasanalyzed to determine the probability of meetingeach performance target through the 14-yearperiod (below). The portfolio has a low chance ofmeeting near-term targets for net income, netcash flow and production. It also has a poorchance of staying within the target for capitalbetween 2002 and 2010, and likelihood of mak-ing the reserves target declines after 2006.These displays help decision-makers understandthe impact of project uncertainties at a portfoliolevel, and can be compared with results thatinclude the new project under consideration.

The production-acquisition project, whenadded to the pool of available projects, could notbe selected for inclusion in the portfolio becauseinitial testing showed that it violates the con-straint on capital for the first two years.Following discussion with and approval by thedecision-maker, that constraint was relaxed andthe project was added to the pool.

All the efficient-frontier portfolios generatedfrom the new pool include the new project, so itunquestionably adds value (above right). The newefficient frontier shifts down and to the right invalue-risk space. For the same value, a portfoliofrom the new pool has less risk, and for the samerisk, more value. The increase in value at a con-stant risk is not the same for all portfolios. For low-risk, low-value portfolios, say at risk level 480, the

16

> Efficient frontier for the pool of Burlington ResourcesInternational (BRI) portfolios. The plot has a differentshape than previously shown efficient frontiers becausethe axes are interchanged. The curve represents thefamily of minimum-risk, maximum-value solutions that allmeet the same performance constraints. The $5.5 billionportfolio has the highest value that could be achievedbefore the risk component increases significantly andbecomes too extreme.

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

Prob

abili

ty

Production ReservesExploration expense

Year Year Year

Prob

abili

ty

Prob

abili

ty

> Probability of meeting or exceeding six metric targets with the optimized portfolio. The portfolio has a low probability of meeting first-year ornear-term targets for net income, net cash flow and production. It also has a poor chance of meeting the capital target after the first two years,and a decreasing probability of making the reserves target in the long term.

> Comparison of efficient frontiers with and without thenew production-acquisition project. The efficient frontier ofportfolios that include the new project (pink curve) movestoward higher value for the same risk relative to the efficientfrontier of portfolios without the new project (black curve).

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

Year

Net income Net cash flow

Capital

Year Year

Prob

abili

ty

Prob

abili

ty

Prob

abili

ty

Oilfield Review

1200

3000

1000

800

600

400

200

03500 4000 4500 5000 5500 6000

Risk

, $ m

illio

ns

Value, $ millions

Efficient Frontier1200

3000

1000

800

600

400

200

03500 4000 4500 5000 5500 6000

Efficient Frontier

Risk

, $ m

illio

ns

Value, $ millions

Winter 2000/2001 17

increase in portfolio value if the new project isincluded is about $1.5 billion. For the higher risk,higher value portfolio at $5.5 billion, the addedvalue is $0.25 billion.

This example demonstrates the differencebetween portfolio value and NPV of a project.The net present value of a project is constant,and measures properties of the project alone.The portfolio value of a project varies as a func-tion of the portfolio and quantifies the cumulativeperformance difference the new project brings.

To understand why this project is valuable tothe strategy, the newly optimized portfolio can be

analyzed without the project (above). Lacking theproduction-acquisition project, the portfolio failsto meet targets for net income in 2005, net cashflow in 2002 and production and reserves in1999. These critical contributions are directlyattributable to the new project, and show whereit adds unique value to the portfolio.

The value is further defined by comparing theprobabilities of success for the new and originalportfolios (below). Significant improvements areclear in the probabilities of meeting short-term

targets for net income, net cash flow and produc-tion. However, the improvements are counterbal-anced by a marginal reduction in the probabilityof meeting the long-term reserves target and theshort-term capital target—not surprising sincethis constraint was relaxed earlier.

With these portfolio-optimization techniques,uncertainties in technical information can betranslated into chances of success. Decision-makers can quantify the value of each project interms of its contribution to total business perfor-mance and its interaction with other projects.

1200

1000

800

600

400

200

01999 2002 2005 2008 2011

1200

1000

800

600

400

200

01999 2002 2005 2008 2011

300

250

200

150

100

50

01999 2002 2005 2008 2011 1999 2002 2005 2008 2011

4000

2000

01999 2002 2005 2008 2011

1000

3000

400

200

01999 2002 2005 2008 2011

100

300

Year

Net income

ProductionReserves

Net cash flowCapital

Exploration expense

Year Year

Year Year Year

Mill

ions

of b

arre

ls o

f oil

equi

vale

nt

$, m

illio

nsM

illio

ns o

f bar

rels

per

yea

r

$, m

illio

ns

$, m

illio

ns

$, m

illio

ns

350300

250

200

100

500

150

2013 2013

2013 2013 2013

2013

> Portfolio performance without the new project. In the absence of the production-acquisition project, the portfolio fails in multiple instances tomeet targets. For example, net income in 2005 and net cash flow in 2002 are below target levels. The constraint on capital for the early years ofthe project had been relaxed.

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

Net income

Net cash flow

Capital

Prob

abili

ty

Prob

abili

ty

Prob

abili

ty

Year Year Year

> Increased probabilities for the new portfolio to meet targets. In some years, the improvements are small, but overall, the new portfolio (pink line)has higher probabilities of meeting targets than does the original portfolio (blue line). Specifically, short-term targets for net income, net cashflow and production are more likely to be met by the new portfolio. Probabilities for meeting long-term targets for reserves and production aremarginally lower.

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

1999 2001 2003 2005 2007 2009 2011 20130

1.0

0.2

0.4

0.6

0.8

Prob

abili

ty

Production ReservesExploration expense

Year Year Year

Prob

abili

ty

Prob

abili

ty

A Strategy for GrowthThe portfolio-management approach is helpingmanagers at Kerr-McGee Oil & Gas Corporationtest and refine strategies and communicate thosestrategies within their organization. The portfolioperspective provides a critical link between strat-egy and investment options for the teams thattake the vision of success defined by top man-agement and produce results that consistentlyplace Kerr-McGee in the top quartile among peer companies.

In 1997, Kerr-McGee began an internal pro-cess to examine peer company best practices,and top management adopted a vision to attain a top-quartile ranking among independent oil and gas companies.10 To reach that status, Kerr-McGee would consider all value-addedinvestment options within their two core busi-nesses—oil and gas E&P and the production andmarketing of titanium dioxide chemicals. Theywould also exploit a deepwater core compe-tence, optimize all existing assets and initiate

pay-for-performance incentives based on exter-nal performance benchmarking. Kerr-McGee’schallenge was the same one facing all oil and gas companies: to generate controlledgrowth in an industry that is characterized bydepleting resources.

A representative E&P portfolio example illus-trates the portfolio-modeling methodologyadopted by Kerr-McGee. The building blocks of atypical company’s economic models for availableprojects include proven oil and gas properties,exploitation and exploration projects, and com-mercial opportunities. A sample generic E&Pproduction projection from four types of assets—base production, identified development, explo-ration and commercial opportunities—indicatesa slight increase in production over an eight-yearperiod (above). Comparing production targets forthat period with the chance of reaching thosetargets—90% probability, mean and 10% prob-ability—showed that production targets wereunlikely to be met with the existing mix of assets(next page, top left).

Changing the mix of exploration, commercialopportunities and other projects can help identifyan optimized strategy (next page, bottom). In thisexample, the new asset mix contains a better bal-ance of low-risk, high-certainty opportunities withhigher risk, lower certainty projects. This yields aproduction projection that increases significantlythrough 2007, meeting short- and long-term objec-tives while adding significant value to the portfo-lio. The probability of achieving the productiontarget, as well as other metrics, also increases(next page, top right). The targets correspond tothe metrics used to define success—in this case,top-quartile performance.

In addition to providing a useful tool for test-ing various goals and objectives against the fea-sibility of generating acceptable results, theseplots are an excellent way to communicate therequired changes—both upward to the chairmanand key corporate executives and downward toregional vice presidents and staff. The portfolio-modeling concept helps to quantify many ques-tions that must be asked when determining thestrategic direction and goals of an organization:Is the strategy feasible? How likely is the strat-egy to succeed? How sensitive is the strategy toprice changes or political events? Which goalsare problematic? What other strategic alterna-tives exist?

At Kerr-McGee, the strategy acts as a compass,or general direction, for the company to pursue,and provides a focus to ensure movement towardtop-quartile results. The portfolio-managementapproach continues to be a valued tool after goalsand objectives have been set, as internal andexternal variables change and new opportunitiesarise. The portfolio-modeling concept also providesan excellent mechanism to investigate investmentoptions, determine trade-offs between opportuni-ties and help managers make better business deci-sions that add value to the portfolio at anacceptable risk. While the portfolio-modeling con-cept does not provide “the” answer, it adds disci-pline to the decision-making process.

18 Oilfield Review

> Kerr-McGee E&P production projection before portfolio optimization, showing base production (pink), identified development (dark blue), exploration (light blue) and commercialopportunities (yellow). Production volumes increase slightly from 1999 to 2007.

10. Adams T: “Using Portfolio Models to Optimise andCommunicate Strategy and Achieve Goals,” presentedat the Gulf Coast Association of Geological SocietiesConvention, Houston, Texas, USA, October 26, 2000.

11. Haspeslagh PL: “Portfolio Planning: Uses and Limits,”Harvard Business Review (January-February 1982): 58-73.

Mill

ions

of b

arre

ls o

f oil

equi

vale

nt p

er d

ay

1991

Year

1993 1995 1997 1999 2001 2003 2005 2007

Base-Case Strategy Before Optimization

Base productionIdentified developmentExplorationCommercial opportunities

Winter 2000/2001 19

The Portfolio Vision Portfolio management offers a methodology fordecision-makers to evaluate asset portfolios,assess the likelihood of meeting objectives andbridge the gap between targets and the resultsattainable under the current strategy. Many indus-tries already use these methods to reach theirgoals for long-term growth. Results of a surveyconducted almost 20 years ago indicate that sev-eral process-driven industries such as chemical,

food and paper manufacturing, as well as thedownstream petroleum-refining industry, alreadyhad years of experience with the portfolio per-spective.11 Now, 20 years later, the upstreampetroleum industry finally can take advantage ofthe approach, thanks to improved computingpower and analysis tools.

The same survey found that when companiesadopted a portfolio approach, their focus shiftedfrom short term to long term. Instead of ranking

next year’s profit objectives as most important inthe planning process, the top ranking went tolong-range profit objectives.

Some companies reach the portfolio-management stage quickly, within about threeyears, while others take longer. In all cases, astrong commitment from top management is the key to fast implementation, and success isbased on dealing with administrative andorganizational issues related to the portfolio-management approach.

Success also requires that the elegant theoryof portfolio management be tailored to fit thecomplex reality of the E&P business. Several com-putational-optimization tools, such as the MerakCapital Planning software, allow decision-makersto focus on issues that help balance business per-formance and manage diverse opportunities.

Using techniques as sophisticated as thosefound in other E&P domains, such as reservoirmodeling and simulation packages, thesetools help reduce complex problems to man-ageable ones that can be analyzed consis-tently and logically. —LS

Year20001999 2001 2002 2003 2004 2005 2006 2007

Typical Base Before Optimization

10% probabilityMean90% probability

> Production targets (vertical bars) and three curves showing probabilitiesof reaching those targets with the original portfolio. The 90% probabilitycurve falls far below the targets, and the mean and 10% curves also fallshort of the targets. Probability of meeting the targets is less than 10%.

> A new mix of assets for an optimized strategy. Including a better balance of lower risk and higher risk exploration and commercial opportunities yields a production projection thatincreases significantly over the eight-year period (1999 to 2007).

Year20001999 2001 2002 2003 2004 2005 2006 2007

Optimized Strategy

10% probabilityMean90% probability

> Production targets (vertical bars) and three curves showing probabili-ties of reaching those targets with the optimized portfolio. Probabilitiesare high that the new portfolio will meet production targets, and there issome chance that the targets will be exceeded.

Mill

ions

of b

arre

ls o

f oil

equi

vale

nt p

er d

ay

Optimized Strategy

1991

Year

1993 1995 1997 1999 2001 2003 2005 2007

Base productionIdentified developmentExplorationCommercial opportunities