How do you use data to create realistic scenarios for decision making?

Data is your compass, not a crystal ball. Here is the best way to navigate. Gather your data from diverse sources. It is not just about numbers, but news, social trends, weather, etc. A rich mixture paints a truer picture. Leverage AI to help find trends connecting past events to potential futures. Embrace uncertainties; life's unexpected plot twists. Model risks and market shifts, so you are prepared for storms. Translate the insights into actionable scenarios, empowering your decisions. The future is unwritten, but data can guide your course.

1. Determine the type of data that is needed to inform the decision-making process 2. Gather the necessary data from various sources 3. Factor in errors, inconsistencies, or missing values that need to be addressed before analysis 4. Use statistical techniques - This could include descriptive statistics, data visualization, correlation analysis, regression analysis, clustering, or machine learning algorithms 5. Once you understand the data and have identified relevant patterns and trends, you can use this info to create different scenarios representing possible outcomes 6. Assess the potential impact of each scenario on the decision at hand 7. Continuously monitor the outcomes of your decisions and compare them with the original scenarios

Scenario analysis is a strategic tool that enables executives to navigate uncertainty by evaluating potential impacts of various factors on business goals. By crafting and analyzing diverse scenarios, it illuminates risks and opportunities, guiding data-driven decision-making. This approach enhances risk assessment, strategic agility, and prepares the organization for multiple futures. It’s an essential practice for leaders aiming to become resilient against unpredictable market dynamics.

At Buffalo Solar, I developed a simulator to accurately predict snow losses on solar panels using Marion et al.'s methods. Through MS Excel simulations, I identified optimal panel orientations that significantly enhanced efficiency. This project showcased how targeted data analysis can directly improve sustainability and operational performance.

Simulation has been a very useful tool for decision making, especially when it's difficult to derive a closed form solution to an optimization model that resembles the decision problem at hand. There are three main approaches for simulation modelling in decision making optimization: 1) Agent based modelling 2) Discrete Event Simulation 3) System Dynamics In any of those approaches, you apparently need to model your system in a way that closely resembles the real system under investigation. That's where predictive analytics approaches can help you ensure your models fit your data properly. Simulation itself is prescriptive analytics and as such can help you find near optimal choices for the factors you control, to achieve your goals.

Scenarios often involve complicated relationships and interdependencies between different factors. Using data in simulations can capture these complex interactions and provide a more holistic understanding of how different variables influence each other and contribute to the overall scenario outcomes. Furthermore, this approach allows the impact of variations in key parameters to be assessed. This helps to identify critical influencing factors and understand how changes in these variables can impact the overall scenario and provides insights into the robustness and vulnerability of different strategies.

Leveraging data to simulate scenarios enables precision in forecasting and decision-making. It quantifies impacts, uncovers trends, and validates assumptions, ensuring strategies are rooted in reality. However, the integrity of simulations hinges on the quality and reliability of the data itself. High-quality, accurate data ensures it reflects realistic outcomes, while unreliable data can lead to flawed insights and misguided decisions. Ensuring data integrity, through validation and cleansing processes, supports scenario planning, enhancing the credibility and usefulness of simulation outcomes. This approach enhances objectivity and reliability, empowering leaders to navigate uncertainties with confidence

This is an interesting question. You need to remember that even if you have data, if the data is sketchy, you're going to get a sketchy scenario at the end of your analysis. I'd say that even if the data being used for a scenario-analysis is arbitrary, it must be steeped in some kind of reality, even if the source of the data is qualitative. Heuristics do play a big role in an accurate scenario-forecasting. Sometimes you won't get an official, vetted source for metrics like attrition or growth. And yet, the need to work out what might happen next year is crucial.

Using data for scenario analysis offers validation, quantification, and objectivity. It grounds scenarios in evidence, quantifies effects, and ensures realism. Data-driven simulations handle complexity, enhance precision, and assess robustness. They also improve credibility by reflecting realistic outcomes. Even imperfect data ensures scenarios are rooted in reality, aiding accurate forecasting and decision-making.

Using data to simulate scenarios helps you prepare for different outcomes and make better decisions. Here’s how: Predict Outcomes: Data-driven scenarios show possible futures, helping you understand potential risks and opportunities. Informed Decisions: By seeing how changes in variables affect results, you can choose the best course of action. Resource Planning: Anticipate needs and allocate resources more effectively. Stress Testing: Assess the impact of extreme conditions to ensure your plans are robust. Using data to simulate scenarios gives you a clearer picture of what might happen, so you can act with confidence.

Where data changes and responses to them are time critical (for example offshore energy) it is important to have networks of appropriate sensors continuously monitoring, feeding and alerting changes to data concerning ongoing operations to appropriate decision makers so they can respond swiftly.

When choosing data sources for scenario simulations, it's critical to apply methodologies that incorporates data source correlation and validation techniques: - Assess the credibility, timeliness, and accuracy of each data source, focusing on those with a track record of reliability and relevance - Use statistical methods to analyze the relationships between data from different sources, ensuring they are logically and empirically connected - Implement checks for data consistency across sources, employing anomaly detection and time-series analysis to identify and rectify discrepancies. - Leverage advanced analytics and machine learning algorithms to integrate and harmonize data, enhancing the depth and breadth of your scenarios.

General rule of thumb - your data needs to be from a reliable source. However, again, I feel the need to underline cases where you need to choose data sources but they might not be reliable or up to date. And again, here, heuristics come into place. BUT - in my experience, the important thing is to have a method or formula steeped in common sense and experience in order to produce a number or a metric to use in your scenario analysis that won't get you into trouble as an analyst.

Choosing data sources for scenario analysis requires accuracy, relevance, and timeliness. Prioritize sources aligned with objectives, accessible, and compatible with analysis tools. Consider ethical and legal implications to ensure compliance. By selecting reliable, up-to-date sources, you enhance the quality of analysis, supporting informed decision-making.

First, write down the core question you want to answer via the scenario analysis. Identify what the major factors or variables that will significantly change the answer. Then, set initial values at the most common case or the oblivious scenario. Run the analysis with combinations of values in each factor. Take medicine pricing for example. The core question is "what is the optimal price for the new medicine that can enlarge our market share?" We can use factors, like incident rate, number of patients, recurrence rate, efficacy rate, etc.

To brainstorm drivers for scenario planning or decision-making, try a few key techniques. Start with brainwriting, where everyone silently writes down ideas to avoid groupthink. Use mind mapping to visually connect related ideas, sparking new thoughts. Conduct stakeholder interviews to gather diverse perspectives. PESTEL analysis helps you systematically consider political, economic, social, technological, environmental, and legal drivers. Finally, look at historical trends for insights. Prioritize the most impactful and uncertain drivers for your scenario planning process.

To create scenarios: Define the focal question or issue to address. Identify key drivers influencing the question. Assign values or ranges to each driver based on data and assumptions. Combine drivers and values to generate scenarios. Provide descriptive names or labels for each scenario.

Think of your scenario analysis calculator like a machine that produces something. You have a few knobs to adjust. Depending on how you adjust the knobs, you'll get a slightly different product/output. Now what does your outcome broadly look like? And what are the knobs that you can adjust? And what effect should they have on your output? That in my experience ie the hard part, getting all that data together. Once that's done, you can choose whatever tool you're comfortable with to get the job done.

In addition, I would develop scenarios to create Baseline scenario: This scenario represents the most likely outcome based on current data and trends. Alternative Scenarios: Develop alternative scenarios to explore different potential futures. These might include: Best-case Scenario: Optimistic outcomes, assuming favorable conditions. Worst-case Scenario: Pessimistic outcomes, considering potential risks and challenges. Other What-if Scenarios: Explore specific events or changes, such as regulatory changes, technological advancements, or competitive actions.

Three key recommendations while evaluating scenarios: 1. Focus on probability: Pressure-test the resilience of your plans and explore how changing key assumptions or drivers affects the outcomes of each scenario. 2. The pitfall of precision: While data-driven scenarios provide insights for making business decisions, remember they provide probable, not definite, outcomes. Instead of striving for perfect predictions, focus on understanding the relative impact of different possibilities. 3. The importance of stakeholder involvement: Involving stakeholders in evaluating scenarios is highly recommended as their perspectives and insights can greatly enhance the evaluation process and ensure that various viewpoints are considered.

Ideally would be to Model and Simulate Scenarios Use Statistical and Computational Models: Apply techniques such as regression analysis, Monte Carlo simulations, or decision trees to model the scenarios quantitatively. Assess Assumptions: Clearly outline the assumptions made in each scenario, as these will impact the results. And then evaluate and compare these scenarios.

Start by identifying key variables relevant to the decision. Collect relevant data from reliable sources, analyze it to id patterns, trends, and potential outcomes. Next, create different scenarios that represent possible outcomes based on a range of assumptions and both best and worst case scenarios. Test each one against available data to determine likelihood of each outcome. This can help assess potential risks and rewards associated with each scenario, and determine which option is most realistic and favorable. Ultimately, using data to create realistic scenarios for decision making requires careful analysis, consideration of all available data, and a willingness to adjust assumptions and predictions as new data becomes available.

1. Understand your audience. It is almost certain that the people who are listening to a scenario analysis, don't primarily care about the data/analytics side of things. 2. Understand how your audience want to be talked to. Just a description? Recommendations? COnnection to strategic projects and outcomes? Risk factors? Alternatives? Do that homework. The actual scenarios themselves might not be of that much interest. 3. Decide on the best format - some might want a 100-slide PPT (OK, different discussion 🤣). Others might prefer a memo (Amazon!), and yet in some other situations, you might be asked to stand up and present a summary.