Executive Guide to Forecasting in Manufacturing, Part 1: Why Process Matters

Forecasting is a recurring frustration for many manufacturing executives. Forecast meetings often feel like negotiations - sales pushes numbers up, finance pushes them down, and operations sits in the middle. The result is a compromise that serves politics more than performance. Unsurprisingly, trust in the forecast process erodes, and executives default to gut instinct.

Yet every critical decision - production planning, capacity investments, and inventory allocation - depends on having a reliable view of the future. As Demand Forecasting for Executives and Professionals by Stephan Kolassa, Bahman Rostami-Tabar, and Enno Siemsen reminds us, the question is not whether a forecast is "good" or "bad." The real test is the quality of the process behind it. Or as the book puts it: "Every manager involved in forecasting must accept that there are no good or bad forecasts. There are only good or bad ways of creating or using forecasts." For enterprise manufacturers, this shift in mindset can unlock significant supply chain improvements.

Key Insights from the Book

Forecasting often falters due to scattered data, complex models, and weak cross-functional alignment. Too often, forecasts are misused as targets or plans rather than objective predictions of the future, leading to political compromises and lost executive trust.

But eliminating forecasting isn’t an option. Even make-to-order businesses rely on implicit forecasts to plan raw materials, component purchases, and workforce needs. And since supply chain lead times are often longer than what customers are willing to wait, pure make-to-order is rarely feasible. Someone still needs to forecast and hold inventory. If that process fails, the whole supply chain feels the cost.

The book stresses two essential truths:

• Process over outcomes. A forecast can miss actual demand and still be “good” if it used all available information. A forecast that happens to match reality but ignored key signals is “bad” because the process was flawed. As the authors note, betting on the right roulette number does not make it a good decision - it just means you got lucky. The same is true for forecasts.
• Forecastability matters. Some demand series are inherently harder to predict than others due to noise, volatility, and lack of signal. Repeated misses do not always mean the process is broken - sometimes the underlying demand is simply unpredictable. What matters is whether the process improves accuracy compared to basic methods.

Another critical point is the organizational challenge. Improving forecasting is rarely about just buying new IT systems or hiring more analysts. The real difficulty lies in aligning diverse stakeholders, overcoming silos, and ensuring clear communication across functions. Without this, even sophisticated tools underperform.

Finally, the book warns against over-reliance on point forecasts. A single number creates an illusion of certainty. Effective processes communicate uncertainty through ranges or probability distributions, helping leaders balance risks in decisions.

Quantics Perspective

In enterprise manufacturing, forecasting is not optional - it is foundational. Yet the real challenge lies in managing complexity:

• Different business needs, different granularity. Business units, regions, and plants often require forecasts at different levels of detail and over different horizons. Finance may need category-level projections, while production and logistics require SKU-level inputs tied to plant capacity or regional demand. In B2B settings, bulk orders create intermittent demand patterns that are difficult to predict at a granular level and make it harder to judge forecast quality.
• Anecdotal vs. evidence-based improvement. Many organizations believe that adding external signals - like market indicators or customer insights - could improve forecasts. But too often these remain anecdotal ideas rather than validated, evidence-based improvements.
• Sales representative input. Sales teams often hold valuable market knowledge, but consistently capturing and structuring this input in large organizations is notoriously difficult.

This is where Quantics goes far beyond generating a "best possible forecast." Our solution supports the entire forecasting process, enabling executives to manage complexity with clarity and control:

• Benchmarking every forecast against statistical baselines to evaluate quality in context.
• Incorporating both internal and external demand signals, with built-in checks for correlation, causality, and time lags.
• Moving beyond point forecasts by providing probabilistic forecasts and scenario planning.
• Tracking multiple stages of each forecast run, separating true forecasting from target-setting in multi-step S&OP processes.
• Automating collaborative workflows so large teams can contribute efficiently while reducing manual effort.
• Delivering transparency by logging overrides, documenting reasoning, and measuring the added value of every adjustment.

For executives, this means greater clarity, efficiency, and trust in the forecasting process - and, most importantly, a direct line of sight from forecasts to better, faster business decisions.

Practical Takeaways

Executives can strengthen their forecasting effectiveness by:

• Separating forecasts from targets and budgets. Forecasts should be objective inputs, not negotiation outcomes.
• Focusing on process quality, not forecast precision. Judge forecasts by how well they use all available information, not by how closely they match actual demand in hindsight.
• Recognizing forecastability limits. Some time series are inherently harder to predict - the goal is to improve accuracy relative to other methods, not to chase perfection. Just as a race car might reach 350 km/h on a smooth track but struggle on a muddy trail, some demand series simply won’t allow high accuracy. Results must be interpreted in context, using benchmarks to judge whether a forecast is good or bad for the situation at hand.
• Communicating uncertainty clearly. Replace single-number forecasts with ranges or distributions to support risk-based decisions.
• Aligning cross-functional teams. Manage forecasting as an organizational challenge, breaking silos and ensuring transparency.
• Improving process before technology. Fixing the process often yields faster ROI than just betting on new systems.
• Leveraging dedicated software for enterprise manufacturing. Platforms like Quantics go beyond delivering the best possible point and probabilistic forecasts - they strengthen the process, enforce consistency, and scale demand forecasting and planning across complex organizations.

In Conclusion

Forecasting in enterprise manufacturing will always be challenging - some time series will never be easy to predict. But by focusing on process quality, acknowledging the limits of forecastability, and addressing the organizational challenge of cross-functional alignment, executives can transform forecasting from a political burden into a source of resilience and efficiency.

At Quantics, we believe the future of forecasting is not just about better algorithms - it is about better processes, smarter collaboration, and transparent decision-making. In our next post, we will explore the forecasting workflow in detail - showing how executives can structure steps to connect data, expertise, and decisions in a repeatable and efficient way.‍

Disclaimer

In this post, we share highlights from the book “Demand Forecasting for Executives and Professionals” by Stephan Kolassa, Bahman Rostami-Tabar, and Enno Siemsen, together with our own reflections on how these ideas apply to today’s manufacturing supply chains. This is not a replacement for the book, but rather a guide to spark thought and discussion. For a deeper dive, we encourage you to explore the book itself - available here.

‍