The SCM capstone Forecasting Drilling Bits Demand: A New Horizon for Oil & Gas Supply Chains was authored by Thiago Pinheiro Faury and supervised by Dr. Ilya Jackson (ilyajack@mit.edu). For more information on this research, please contact the thesis supervisor.
In the fast-paced world of energy-exploration technology, companies are constantly on the lookout for innovative strategies to help them stay ahead of market dynamics. For one global energy company, the key to unlocking efficiency was transforming how it predicts the demand for drilling bits—essential tools in oil and gas exploration.
Transforming forecasting: From manual to automated excellence
Demand for drilling bits is complex to forecast due to their variety in design and the need for application-specific manufacturing methods, ranging from build-to-order to engineered-to-order. Each bit type requires detailed planning to align with market demand—a challenging task when relying on manual methods across 30 geographic units. Manual forecasts typically cover up to three months, but this period is inadequate for manufacturing certain types of drilling bits and planning long-term capacity and production schedules.
Traditionally, the company relied on manual and descriptive forecasting methods. This approach often led to inaccurate forecasts, excess inventory, increased costs, and sometimes a dip in market share due to unmet customer demands. However, a shift toward automation and predictive analytics has revolutionized their forecasting, resulting in more accurate predictions and optimized operations.
A Leap in Accuracy
By adopting automated causal and time-series models, the company significantly enhanced its forecasting accuracy. With the new models, the global mean absolute percentage error (MAPE) rates have dropped by 65%. This leap in accuracy isn’t just a number—it’s a game-changer for managing inventory and optimizing resource allocation.
More Results
Enhanced data visualization through Excel pivot tables with conditional formatting helped highlight performance trends across models and geographic units. Findings revealed that traditional time-series models, such as Croston, Theta, and FourTheta generally outperformed more complex causal models, particularly in geographic units with specific business models like high tier and high volume, tied to consistent activity patterns.
While automated forecasts offered considerable global advantages, especially for activity, their effectiveness varied locally (i.e., at a geographic unit level), underscoring the need for tailored approaches in different market contexts.
For more about this capstone project, and to see the full results of this research, visit the Supply Chain Management Review online at SCMR.com.
Sponsor a Capstone Project
Collaborate with seasoned MIT SCM students and researchers to tackle real-world supply chain challenges head-on. Through a 10-month partnership, your team gains actionable insights and fresh solutions as student teams apply analytical skills and effective strategies to your current projects. Joint teams that include MIT SCM students and MIT CTL faculty work on real-world problems. If you are interested in learning more about sponsoring a capstone project for the upcoming academic year, please email scm-capstone@mit.edu to connect with a member of our team.
Contact Us