5 ingredients of a successful predictive simulation project

A plant manager uses Twinn Witness predictive simulation software while overseeing manufacturing operations


Duncan Brown is a Senior Business Development Manager with Twinn, by Royal HaskoningDHV. He’s experienced in helping clients make better decisions through the use of predictive simulation and digital twins. 

You have the software and know how to use it. So you’re poised for predictive simulation success, right?

Not quite.

Those are 2 key elements, but unless you have these other 5 ingredients, the technology and training will only get you so far. Let’s look at the essential people and processes for succeeding with predictive simulation.

1. It’s a team effort – involve the right people

Complicated organisational structures – together with siloed data and decision-making – mean it’s tempting to take a narrow view of a simulation project. Involve the minimum number of people so you don’t wind up in decision paralysis.

But there’s major pitfall to that approach. When you execute a simulation project in isolation, you limit value and return on investment because you don’t have the right inputs or the follow-through.

Stakeholder engagement

You need to seek out the right people in different parts of the organisation – be it design engineers, programme managers, operations managers, production managers, supply chain managers, finance and more. Any area that could be a potential beneficiary of or contributor towards predictive simulation should be engaged.

A stakeholder engagement matrix like the one to the right is useful for identifying and managing the various people you involve.

Through this engagement process, you generate enthusiasm for predictive simulation – those ‘light bulb’ moments where people understand how their lives will improve.

A stakeholder engagement matrix which helps plot the needs and requirements of project team members

The old ‘show don’t tell’ motto comes into play here. In our experience, stakeholder workshops facilitated using a visual interactive simulation model bring the potential to life. This means you aren’t just reviewing PowerPoint plans, looking at static images or playing videos (although these all have value). Rather, you’re enabling people to explore how a process works and encouraging them to identify suggestions for improving it collaboratively. Importantly, you’re also showing them how to achieve it.

Model development

Likewise, the modelling itself shouldn’t be one person’s responsibility. Ideally, your predictive simulation team should consist of:

  • Domain experts who understand the intricacies of the system you're modelling
  • Data scientists proficient in handling and processing data
  • Experts who can develop the simulation software, supporting in-house programmers
  • Validation experts who can assess the model’s accuracy

Effective communication and collaboration are crucial. Regular meetings, brainstorming sessions and a shared vision keep everyone on the same page.

2. Begin with the end in mind – clearly define the goals

With stakeholder engagement and enthusiasm buoying the project, the next key ingredient for a successful simulation project is having clear objectives and defined scope.

A common pitfall we see is starting with the software rather than the goals and scoping study. This puts you in a risky situation where you can easily over-engineer the solution or create something that’s too simplistic.


You can’t build a house without a strong foundation, and goals are that foundation in a simulation context. Goals should be specific, measurable, achievable, relevant and time-bound (SMART), so people are aligned around a shared vision.

Scoping study

The scoping study enables you to define the appropriate level of model detail, so you can structure your work and create a realistic design and plan for delivering value. This phase is designed to bring out all the business questions the model needs to answer, with stakeholder interviews helping define the short- and long-term goals related to those questions.

This then determines what processes need to be modelled, what level of detail needs to be reflected with predictive simulation, what data is required and what reports will look like. It also addresses future uses.

When we support organisations through this phase, we often get feedback saying that, through the detailed goal-setting and scoping exercise, people learned more about their processes than they ever did before. This then generates even more enthusiasm about the benefits of predictive simulation.

3. Rubbish in, rubbish out – ensure the right data, at the right level

Data is the fuel that powers simulations. Just as a car won't run smoothly with low-quality fuel, a simulation project will struggle without high-quality data.

Data identification

Start by identifying the data sources relevant to your project. Depending on how complex the simulation is, this may include real-world observations, historical data and/or experimental data. Ensure data is comprehensive and representative of the system you're simulating.

Data pre-processing

Data accuracy and reliability are paramount. Data identification can be an exercise in of itself, but pre-processing is an essential step.
It’s often necessary to clean and transform raw data into a usable format. The better your data quality, the more accurate and reliable your simulation will be – and the more likely it is to achieve its goals.

4. Testing times – verify and validate

Once the model is developed to a certain point, it’s time to test and validate it – to ensure the results are trustworthy and that it accurately answers the questions defined in the scoping study. Rigorous testing gives confidence in the model itself, and that the data is feeding it effectively.

Validation vs verification

Validation and verification are the quality control checks that guarantee the simulation’s reliability.

  • Verification checks that the simulation software behaves as intended
  • Validation involves comparing the simulation's output to real-world data to ensure it represents the system being modelled

These steps may involve sensitivity analysis, calibration and rigorous testing under various scenarios. They are critical to identifying errors, refining the model and building faith in the simulation's results.

Be thorough

When you’re just shy of the finish line, it’s tempting to rush the testing process. Don’t. It's like stress-testing a bridge to ensure it won't collapse under pressure.
Start planning validation and verification early. Involve all key stakeholders, including reporting these efforts to people in the ‘Monitor’ and ‘Keep informed’ quadrants of your stakeholder matrix. This will help deliver widespread confidence in the model so people are more likely to accept its results.

When you’re testing, review the entire process – not just the model itself. It’s not just about debugging. You need to understand the model’s suitability and ability to answer the business questions in the scoping study.

5. Sweat your asset – ask ‘what-if?’ to gain insight and foresight

Once the model is fully tested and validated, it’s time to use it. The experimentation phase is when you run ‘what-if’ scenarios to understand the effect of different decisions and trade-offs.

Thorough experimentation and Design of Experiments (DoE) are the backbone of successful simulation projects. They serve as the litmus test for the accuracy and reliability of simulated models. By systematically varying input factors and observing their impact on outcomes, DoE allows you to uncover hidden relationships, optimise processes and validate the simulation's predictive capabilities. Twinn Witness predictive simulation software, for instance, has advanced experimentation and optimisation functionality that simplifies this process and the process of interpreting results, so you gain the deep insight you need.

Without rigorous experimentation, simulations risk becoming mere theoretical constructs divorced from reality. In essence, experimentation and DoE bridge the gap between simulation and real-world applicability, ensuring insights can drive informed decisions, enhance efficiency and, ultimately, lead to tangible, impactful results.

Now your predictive simulation project has ‘wow’ factor

When your simulation project has these 5 ingredients, it has longevity. You end up in a position where your model:

  • Enables experimentation with immediate business questions
  • Also provides a foundation for ongoing, evidence-based decision-making that maximises efficiency and return on investments

John Ladbrook at Ford summed this up nicely. He said:
“Success is marked by the growing use of simulation in plants and, most of all, in other departmental meetings where we frequently hear ‘what does the simulation say?’ It delivers real results, based on real data, that everyone believes in.”

Read more about Ford’s predictive simulation experience here.

And contact me to learn how these ingredients can deliver value for your predictive simulation project.

Do you want to know more or have a question? - Contact our experts!

Do you want to know moreor have a question?

Contact our experts!