Predictive Analytics Case Experience

Forecasting Tomorrow's Gasoline Price Today


In many cities around the world, retail gasoline prices follow a regular price cycle in which there is a long series of small price decreases interrupted by occasional, large and sudden price increases. Dr. Noel is an expert on retail gasoline price cycles and has published several research studies that show how one can use econometrics and predictive analytics to forecast gasoline price increases before they occur. Dr. Noel designed and estimated the effectiveness of a variety of data-based prediction techniques, including techniques based on calendar patterns, cycle-length patterns, price-cost margin patterns, and cross-sectional price min-max patterns. Dr. Noel then simplified these techniques into simple purchase timing rules of thumb that interested consumers could use to save money on gasoline. Dr. Noel's techniques enable these consumers to systematically buy at or near the "troughs" of the price cycle, at below-average prices, again and again.

Shortly after Dr. Noel released his study on how to predict gasoline prices for Australian cities, the Australian Competition and Consumer Commission (ACCC) updated their consumer gasoline pricing page to begin offering predictions on when gasoline prices were likely to rise. Updated three times a week, the ACCC website now recommends on a given day that consumers either "avoid" buying now (when prices are likely to be at their highest), "delay" buying until a bit later (when prices are lower but likely to still fall) or buy now (when prices are predicted to be close to the troughs). These recommendations are consistent with the "spike-and-wait" and "position-based" purchase timing strategies, as those terms are defined and described by Dr. Noel in his published work.

Back to: