Process lock-in: Are your forecasts accurate?

Forecasting methodologies in the manufacturing environment

Majority of manufacturers are poor at forecasting, surprisingly, this doesn’t put them off, they suffer from “epistemic arrogance” caused by a misplaced sense of confidence based on models which appear scientific. However, some forecasts are accurate. They know exactly the amount of exploitable hours in a given day, month or year. Such forecasts are nonetheless necessary as they help manufacturers make decisions about resourcing their operations for the future.

Changes in the manufacturing industry happen frequently, and understanding the pace of these changes is crucial for effective business planning. The first step is determining how far into the future a manufacturer needs to forecast. This depends on the available options and decisions, which in turn dictate lead times and the planning of short-term, medium-term, and long-term actions along with their respective timescales.

Relying on simplistic extrapolation techniques in manufacturing often leads to errors. This is because many contextual variables can significantly impact the industry. Therefore, two forecasting techniques are particularly relevant: qualitative and quantitative. While no single approach guarantees an accurate forecast, combining qualitative insights from experts with quantitative predictive models can yield highly effective results.

Qualitative Forecasting Methods: Empowering People

Qualitative forecasting involves gathering and evaluating judgments, opinions, and past performances from experts to make predictions. This approach not only yields valuable insights but also empowers people by leveraging their experiences and expertise. Various manufacturing excellence functions such as group activities, autonomous maintenance, people development, office and administration, and overall leadership excellence can significantly benefit from qualitative forecasting. Here are three main techniques:

1. Panel Approach The Panel Approach brings together diverse teams for regular discussions, fostering a collaborative environment that enhances the reliability of forecasts. This method is particularly effective in group activities and autonomous maintenance, where different viewpoints can provide a comprehensive understanding of potential changes and improvements.

• Technique: Cross-functional focus groups.

• Meeting Frequency: Daily/weekly or quarterly/biannual.

• Application: Suitable for short-term and medium-term forecasts.

• Advantage: Multiple perspectives enhance reliability.

• Challenge: Reaching consensus can be difficult. 2. Delphi Method The Delphi Method uses a series of questionnaires to gather expert opinions without the influence of group dynamics. This technique is ideal for areas like people development and office administration, where unbiased, expert-driven insights can guide medium-term planning and improvements.

• Technique: Expert questionnaires.

• Meeting Frequency: Quarterly/biannual.

• Application: Suitable for medium-term forecasts.

• Advantage: Reduces bias from face-to-face meetings.

• Challenge: Constructing appropriate questionnaires and selecting experts. 3. Scenario Planning Scenario Planning explores various future possibilities, making it perfect for overall leadership excellence. By considering multiple long-term scenarios, leaders can develop robust strategies that are resilient to different potential futures.

• Technique: Long-range projections.

• Meeting Frequency: Annual/strategic planning.

• Application: Suitable for long-term forecasts.

• Advantage: Considers a range of future scenarios.

• Challenge: Not focused on consensus but on exploring options.

Quantitative Forecasting Methods: Process Lock-in and Equipment Optimization

Quantitative forecasting uses mathematical models to predict future behavior, focusing on two critical areas: process lock-in and equipment optimization. By analyzing historical data and identifying patterns, businesses can make informed decisions to improve efficiency and effectiveness. Here are the two main quantitative forecasting methods:

1. Time Series Analysis Time Series Analysis is highly effective for short-term and medium-term forecasts in process lock-in areas such as 6S (Safety, Sort, Shine, Set-in Place, Systemize, and Standardize), Visual Management, Focused Improvement, Safety, Health and Environment (SHE), and Overall Process Effectiveness (OPE). By analyzing past trends, businesses can optimize these processes and maintain high standards of operational excellence.

• Technique: Moving averages and exponential smoothing.

• Meeting Frequency: Daily/weekly.

• Application: Suitable for short-term and medium-term forecasts.

• Advantage: Effective for predicting based on past behavior.

• Challenge: Ignores causal variables. 2. Causal Modelling Causal Modelling is ideal for long-term forecasts, particularly in equipment optimization areas such as equipment and product management, planned maintenance, quality maintenance, energy and water management, and Overall Equipment Effectiveness (OEE). This method helps businesses understand the interrelationships between various factors, allowing for more strategic planning and resource allocation.

• Technique: Regression models and complex networks.

• Meeting Frequency: Continuous improvement cycles.

• Application: Suitable for long-term forecasts.

• Advantage: Accounts for relationships between variables.

• Challenge: Complexity and assumptions of models.

The Power of Integrated Forecasting

Combining qualitative and quantitative methods enhances forecasting accuracy, blending expert judgments with predictive models. This integrated approach significantly improves various manufacturing functions, providing a roadmap for immediate improvements and a resilient framework for future growth. Techniques like Time Series Analysis and Causal Modelling, when applied to process lock-in and equipment optimization, maintain safety, efficiency, and sustainability standards.

Charting a Path Forward

By embracing both qualitative and quantitative forecasting methods, manufacturers can overcome the pitfalls of over-reliance on simplistic models and epistemic arrogance. Instead, they can harness the power of expert insights and robust data analysis to navigate the complexities of the manufacturing landscape with confidence.

These integrated forecasting techniques empower organizations to drive continuous improvement, make informed decisions, and strategically plan for the future. In doing so, manufacturers can achieve sustained operational excellence and remain agile in an ever-evolving industry.