Improving upstream forecasts

Companies today are imposing internal performance contracts based on the reliability of these forecasts. Major construction project contracts may include performance and penalty elements that require realistic forecasts. Moreover, the investment community and other stakeholders are holding management accountable for their predictions. These pressures are compounded by the fact that many oil and gas projects are becoming more complex, requiring the integration of more components with greater technical uncertainty, combined with persistent political risk.

With effective dialog about projects and employing the right tools, E&P financial managers and decision makers can interact with their project evaluation teams, asking the questions that lead to more accountability and better forecasting.

Framing the problem: taking off the blinders

Question: How well is your project described? Have you accounted for all critical uncertainties and value drivers?

Experience shows that businesses tend to overlook a number of uncertain things in their particular market environment, as well as the magnitude of these uncertainties and the complexity of the relationships among them.

Take Encyclopedia Britannica: For more than 200 years after it was founded in 1768, management believed it was a book publishing company. During the five years between 1989 and 1994, its sales dropped 53%. CD-ROM versions that were more interesting, less expensive, and easier to use captured its market share. Management was locked into the limiting frame of publishing only bound volumes, which had disastrous consequences for the company. There are countless examples of businesses that have been diminished or disappeared for this very reason.

Credible E&P forecasts begin with an effective description of a given project, a process called “framing the problem”.¹

Good problem framing simply means asking the right questions that will help identify the important value drivers, risks, and upside opportunities. Asking these questions allows for encompassing previously unconsidered facts and variables. Then, with analysis and insight, you can distill the frame into a simpler description of the problem to be solved (see Fig. 1).

null

Ignoring this process is called “frame blindness.”

In the oil and gas industry, for example, evaluation teams engage in frame blindness when they ignore rig availability in evaluating a portfolio of projects and falsely assume any given number of drilling rigs will be available when they are needed. Drilling schedule, production timing, and capital investment phasing can vary substantially when the proper rig market forecast is incorporated. Failure to accurately account for rig availability can lead to costly consequences, such as performance contract penalties and reduced valuation by external stakeholders.

In a real-world case of frame blindness, a project team ignored the disposal of a small amount of associated gas as a key project timing and value driver. This project was in a remote location, far removed from existing infrastructure in a country that prohibited gas flaring or venting. The cost of dealing with that small volume of remote associated gas through a pipeline or other technology was not considered in the original forecast.

When external conditions changed unexpectedly, a new gas disposal plan was rushed through at considerable unexpected cost and erosion of value. Had the right questions been asked, it is probable that a number of alternatives would have been considered early in the planning phase and management would not have been left with a single-option plan.

Questions to test for frame blindness:

1. Have threats and strategies been considered that go beyond routine thinking?
2. What are the primary risks and value drivers for this project?
3. Have all important constraints been considered?
4. What will be the concerns of those who must implement this plan?

Correcting for bias: expanding viewpoints

“Who wants to hear actors talk?” – H. M. Warner, Warner Brothers, 1927

Related to frame blindness is the problem of bias. It naturally creeps into planning and forecasting because it is natural human behavior. A common example of bias is a forecaster’s singular view of the state of affairs at a given point in time. This is often thought of as “overconfidence bias.” Accurate forecasting in all industries and businesses requires recognizing and intentionally correcting for bias.

Overconfidence bias in the oil and gas business is caused by failure to account sufficiently for all the uncertainties associated with forecast parameters, such as reserves, costs, and prices. Since the price of oil is influenced by factors beyond our control, it tends to be fixed for evaluation consistency, even though it is usually the single largest value driver for an E&P project.

In one case of bias, a company established specific, firm constraints when negotiating a production sharing contract with a government. One of the main constraints management set was that a supposed field size had to be economically viable at a fixed low oil price. The negotiator asked about the tradeoffs that could be contemplated toward achieving this objective, including giving concessions in the upside event oil prices rose. The company management’s view of the future simply did not include high oil prices, so conceding this possibility seemed reasonable. (Company management was experiencing another form of bias known as “recency bias” by reacting to a recent price crash and its major negative impact on the company.)

In today’s world of high oil prices (Fig. 2), the concessions this company made to keep the project viable at low oil prices have caused real problems. The contract triggered production sharing at a lower threshold and, despite high oil prices, the value of the project decreased substantially, leading to a deferral and missed forecasts of production and cash flow. As can be imagined, this outcome is very counter-intuitive to senior management and equity analysts.

Testing the impact of uncertainties is a way to make a forecast robust and to identify project elements that should be tested for bias. Sensitivity analysis or Monte Carlo simulation techniques can be applied, depending on the situation and company culture. Key questions could include: Is the project still profitable on the downside and are there enough rigs and other facilities to accommodate drilling and production on the upside? Using this technique, projects can be made modular or phased with pilots, etc., to take advantage of upside opportunities and mitigate downside risks.

Questions to test for bias:

1. Are we biased? Where? (TIP: Beware of subject mat- ter experts who may be the most biased.)
2. What are the possible highs and lows of forecast inputs? Once bias has been identified, have we created a dialog to broaden the expected range surrounding a biased forecast?
3. Have we overcompensated as a substitute for insuficient expertise? (TIP: Always be suspect of wide ranging guesses disguised as estimates on variables. This can lead to “analysis paralysis,” described in the next section. )

Completing the analysis, without the paralysis

“I made this letter longer than usual because I lack the time to make it short.” – Blaise Pascal

Overcompensating for frame blindness can lead to “analysis paralysis”: a condition in which a team becomes so overwhelmed in data and computational models that insight is lost. This is often described as the “refinement of unimportant details”.2

Getting caught in the analysis paralysis trap can result in a loss of focus on the important issues and possibly a lack of insight into the project being evaluated. This can occur when a team rushes into the analysis of a problem. Forecasts developed under these conditions tend to be complex and inflexible when new scenarios and sensitivities are introduced.

It is always better to avoid analysis paralysis at the outset, instead of trying to correct for it later on. Pulling back from the downward slope on the right hand side of the “frustration curve” (see Fig. 3) can be an extremely difficult maneuver.

Avoiding paralysis can be done by asking questions to develop intuition about the project before plunging into analysis. If an analyst begins a sentence saying “The model says,” you may suspect that you are heading away from intuition and toward analysis paralysis.

Questions to test for analysis paralysis:

1. What needs to be done to ensure success of the proect?
2. Are we overanalyzing by dwelling on “why a result occurs”?
3. Are we taking adequate time to develop the frame and intuition for this challenge, or are we rushing to analysis and forecasting?

Effective integration: synthesis not summation

“In complex projects, the whole is greater than the sum of its parts.”

A typical engineering approach in developing forecasts is to consider several project elements sequentially and ultimately arrive at a summation that is often quite sophisticated. Sometimes forecasters use a mathematical optimizer to develop a representation of the combined factors. For simplification, forecasters apply risk factors to costs and flow rates to compensate for uncertainty and capacity limitations.

In oil and gas terms, true project integration is the interplay among reservoir description, well performance, drilling sequence and rig availability, facility capacity, schedule timing and critical linkages, commercial terms and spending limitations. These factors cannot be fully represented by sequential summation because each one influences all the others. Figs. 4a and 4b show the comparison between common summation and integration. This inter-dependency is a key element of integration.

Consider a company that attempted to balance its production output with available markets based on reservoir and production output alone. This led to inaccurate insights, which led to incorrect initial priorities.

The company first determined that gas field production would outpace limited markets at the time. Thus, it rushed to create demand for its projected gas supply. However, the company’s forecasters developed the gas volume predictions using a non-integrated production forecast. They did this using a simple production profile roll up that ignored such factors as infrastructure constraints, rig availability, and human resource limitations.

Finally, the company constructed a business simulation model to take into account these integration matters, especially the pipeline flow constraints and the availability of drilling and workover rigs as well as the personnel to operate them. This integrated forecast revealed that the company did not have gas excesses and limited markets, but rather, gas deliverability limitations and, as a result, market over-commitments (see Fig 5). The company had committed to more markets than it could supply. To correct this, it had to shift priorities, along with the associated work effort and human resource focus, from the marketing side to the development side of the equation.

Questions to uncover lack of integration:

1. What are the dependencies? (TIP: Beware of “roll-ups.”)
2. What are the key variables and how do they influence not just the end project value but also all the other variables within the model? (TIP: Use modeling tools that allow you to visualize your critical integration issues.)

Proper tools: pick the right tool for each job

The E&P industry has worked for years to develop software tools to help integrate various subject matter estimates into a unified model of the business plan for oil and gas assets. But software is only part of the solution. You also need multi-disciplined teams and effective workflow processes. Decision science and related tools are also critical to answering some of the questions posed in this article.

Business simulation is the process of building a model for a real project and using that model to test assumptions and develop insights about the project. As this article briefly describes, an important element of this process is the proper gathering of information about the project, especially about the interaction of all the key variables. A computational tool for simulation could be a simple spreadsheet or another means of organizing and using the data.3

Business simulation modeling and workflow tools, such as PetroVR Toolsuite from Caesar Systems, used in the integration case study above, can provide the E&P industry a unique platform for developing complete project frameworks—frameworks that develop integrated analysis and result in more reliable forecasts.

The question to test for effective business simulation:

Do our workflow and modeling tools have the following important attributes?

• Decision science capabilities: risk analysis and integration;
• Flexibility: a process and a tool that are adaptive to change;
• Transparency: clear presentation of assumptions and results.

Conclusion

The pursuit of more accurate, credible, and dependable forecasting – especially in today’s dynamic and complex upstream oil and gas environment – is no fool’s journey. Success depends on asking the right questions and implementing decision and risk analysis work flows along with enhanced software evaluation tools. Truly integrated business simulation exists at the interface between good decision making and forecasting processes and the accurate and comprehensive modeling of complex systems. OGFJ

References

1. Extracted and condensed from “Managing frames to make better decisions” by Paul J. H. Schoemaker and J. Edward Russo, contained in Wharton on Making Decisions, edited by Stephen J. Hoch, Howard C. Kunreuther, and Robert E. Gunther, published by John Wiley & Sons, 2001.
2. David Skinner, Introduction to Decision Analysis, published by Probabilistic Publishing, 1999. ISBN 0-9647939-3-0.
3. William Gray, Troy Hoefer, Andrea Chiappe, Victor Koosh, “A Probabilistic Approach to Shale Gas Economics,” SPE paper no. 108053, presented at the SPE Hydrocarbon Economics and Evaluation Symposium, Dallas, TX, 1-3 April 2007.

About the authors

Donald Zmick [[email protected]] is vice president of client services for Caesar Systems, which serves the worldwide petroleum industry with integrated business simulation software tools, such as PetroVR tool suite, as well as with training, support and on-site consulting. He has more than 20 years’ experience in the oil and gas exploration and production industry. Zmick earned a bachelor of science, magna cum laude, from Texas Tech University and a master of business administration with an accounting concentration and finance specialization from the University of Chicago. He is a member of the Society of Petroleum Engineers.

Alexander Jok [[email protected]] is a senior consultant for Decision Strategies, a strategy consulting company. He uses his 13 years of financial and operational experience to focus on the appraisal of upstream oil and gas business opportunities, with an emphasis on the evaluation of multi-party economic/field development planning. Jok earned a bachelor of science in physics from Indiana University and a master of business administration from the University of Phoenix.