Three Steps to a Risked Revenue Forecast

Abstract

Management For many reasons, different assets have different levels of risk/uncertainty associated with a decline curve forecast. Failing to quantify these differences in uncertainty is the equivalent of disregarding a significant piece of the business equation because one must include risk and uncertainty when forecasting production. Similarly, failing to quantify the risk associated with a product price forecast—or arbitrarily assuming “high,” “medium,” and “low” cases—is also ignoring a piece of the equation that is not only scientifically/statistically possible to assess, but also quick and easy to calculate. Moreover, to understand and plan for the business situation based on decline curve forecasts, one needs to combine the uncertainties of both the decline curve forecast and the product price forecast. Using low-cost personal computer software, one can quantify the uncertainty spread surrounding the situation to determine the risk associated with a revenue forecast. Because decline curves represent physical systems from the reservoir through the sales meter or stock tank, different decline curves have differing amounts of volatility. Hence, assuming the physical system around the asset does not change, one can systematically quantify the differing degrees of certainty in the forecast (Figs. 1a and 1b). Having quantified the uncertainty associated with the production forecast (Step 1), the professional or manager then turns to the uncertainty associated with product prices (Step 2). Previous analysis and research indicate that oil and natural gas markets are “efficient,” i.e., prices reflect all available public information almost instantaneously and price movements follow a “random walk” model (Figs. 2a and 2b). The random walk premise also suggests that the best forecast of tomorrow’s price is today’s price. Therefore, using a given starting price, one can easily quantify the product price uncertainty going forward. Price volatility ebbs and flows over time. However, looking at historical price volatility will give one a reasonable estimate of the future volatility that one’s organization should be prepared to accommodate. Hence, rather than assuming a price and basing the business model on it, one can proceed with a known price—today’s price—and proceed with a price volatility estimate that is reasonable and scientifically/statistically derived. Having quantified the uncertainty associated with the production (Step 1) and with product prices (Step 2), it is a matter of using statistical analysis to quantify the uncertainty associated with forecasting revenue (Step 3), which is the product of the two. Doing this will yield a revenue forecast that may be alarming when one sees the uncertainty.