5 min read • published in partnership with ITI Group
Why manufacturing simulations can fail before the model is built
Simulation modelling allows manufacturers to test operational changes before committing resources. But many models fail to influence real decisions because they’re built on vague ambition rather than clearly defined questions, according to Darren Travers, simulation lead at ITI Group.
If you work in manufacturing, you don’t need me to tell you that volatility and uncertainty have become permanent features of doing business. Energy shocks, supply chain disruption, geopolitical instability and tightening environmental controls are piling pressure on businesses still expected to increase throughput, control costs, manage risk and reduce carbon emissions.
This fundamentally changes what “good decision-making” looks like. Manufacturers are used to optimising systems from a position of relative stability. Those days are long gone. They now have to make high-stakes decisions while conditions continuously change around them. Legacy decision tools such as spreadsheets, static forecasts and linear planning models were never designed for that level of instability.
Simulation modelling fills that gap. It offers a way to stress-test decisions before committing capital, disrupting live production or redesigning supply chains. It creates a controlled environment for exploring consequences in advance, and as a result, it provides manufacturers with options – the exact thing many businesses tell us they lack.
The value of simulation modelling, however, depends entirely on how the problem is framed.
The ‘exam question’ determines everything
Many simulation programmes underdeliver for the simple reason that they start in the wrong place. Most models begin with an ambition – a digital replica of a factory is a frequent request we hear from manufacturers. Reproducing operational reality at scale as accurately as possible requires a huge amount of data and effort, and as impressive as the final model looks, it’s often unusable for decision-making.
The problem isn’t technical capability. A fully-realised digital twin of a factory or global supply chain is absolutely possible, but what many overlook is the why. A simulation is only as useful as the decision it is built to support. That’s the ‘exam question.’ Build versus buy or whether production should be kept offshore or reshored are great examples.
Without that clarity, that grounding objective, simulation modelling becomes a technical demonstration rather than a decision tool. One manufacturing team, for example, was preparing to invest heavily in additional machinery to relieve an apparent bottleneck. A simulation of the line showed that the constraint wasn’t machine speed but material flow between production stages. Adjusting buffering and sequencing removed the need for the investment entirely, saving the business several hundred thousand pounds.
The starting assumption (slow machines are causing the bottleneck) was wrong and the answer (investing in a new machine) was both wrong and expensive. This is the central discipline of simulation modelling, and where so many businesses trip up. The model doesn’t define the question; the question defines the model.
The counterintuitive advantage of smaller models
Once the exam question is clear, model design becomes simpler and more effective. In many cases, simulations don’t need to replicate entire factories in full detail. They focus on the specific decision being made in isolation.
A production team testing the benefits of a U-shaped cell layout over a straight-line setup doesn’t need a full supply chain replica, in the same way that a global logistics planner doesn’t need production constraints at machine-level detail. They need enough structure to answer one decision and have confidence in the answer.
A focused approach typically produces faster insight, clearer outputs and more direct alignment between model results and real-world action. It also reduces the maintenance burden that often turns large simulation environments into static assets. Another benefit is that it makes trade-offs visible in a way that traditional planning can’t.
Manufacturing systems rarely have a single optimal configuration. I’m sure you heard people say that ‘you never eliminate a bottleneck; you simply move it somewhere else.’ Increasing throughput may disproportionately increase energy consumption, which impacts operating costs and sustainability targets. A supply chain optimised for cost and speed may be less resilient to shocks. These trade-offs are often invisible until after a change is implemented.
Simulation allows them to be seen in advance. That visibility changes how decisions are made. It moves manufacturers from reacting to understanding consequences before taking action. It also changes the role of planning. Simulation won’t help you predict the future, but it will help you prepare for it – whatever it may bring.
Revealing how manufacturers actually make decisions
An overlooked side benefit of simulation modelling is that it often reveals how decisions are made within the organisation, where assumptions come from and how strongly established beliefs influence operational thinking.
Building an effective simulation model requires manufacturers to formalise how processes actually work. That includes sequencing logic, operational constraints, staffing assumptions, material flow and decision rules that may never have been fully documented before. In many companies, this knowledge only exists in the minds of long-serving, experienced engineers and operators.
This is where simulation can cause friction without adequate guidance and support. Simulation has a habit of challenging assumptions that have gone untested for years. We have even had manufacturers ask for a model to produce a specific answer they wanted, not an accurate answer. That reaction is more common than many companies would admit. It reflects the tension between operational instinct and evidence-based analysis, particularly in industries like manufacturing, where decisions have historically relied on ‘gut feel’ and experience.
Simulation doesn’t remove the value of experience, however. We see the best outcomes coming when operational expertise shapes the assumptions and scenarios being tested. The process works best when experienced engineers use simulation to challenge and refine their thinking rather than validate decisions that have already been made.
There is also a longer-term operational benefit. The process of building simulation models captures institutional knowledge that might otherwise be lost when someone leaves the company or retires. Operational logic and practical engineering understanding become embedded in the model itself, creating a structured record of how systems function at a time when manufacturers face losing their most experienced workers.
Better decisions begin with better questions
In our experience, the manufacturers gaining the most value from simulation modelling begin with a clearly defined challenge or goal and build from there. The quality of that initial question determines the quality of the insight that follows.
At ITI Group, we work with manufacturers to structure simulation modelling around practical operational questions and measurable business outcomes. That may involve saving time, lowering costs, driving efficiency, reducing risk or understanding how new technologies will perform before investment decisions are made.
Our simulation experts can help replicate your current or proposed operations in a risk and disruption-free environment, enabling you to easily and clearly visualise how proposed changes will affect performance. The focus is always on helping identify trade-offs, reduce uncertainty and improve confidence in decision-making under real-world conditions.
Volatility is unlikely to ease in the foreseeable future. The manufacturers that respond most effectively will be those that can evaluate options early, act decisively and adapt faster than the competition when conditions change.
