top of page
Writer's picturePrateek Rastogi

Why Baseline Modeling Matters in Supply Chains

Building a Solid Foundation for Smarter Decision-Making



Imagine making a costly decision that impacts your entire supply chain, only to find out later that it was based on incomplete or incorrect data. This is a common risk for businesses that skip or rush through the crucial step of baseline modeling in supply chain design. While some may view it as a tedious process or try to take shortcuts to deliver results quickly, a poorly constructed model can lead to failed projects, or worse, the wrong recommendations being implemented.


In this blog, I’ll explain why baseline modeling isn’t just a necessary step—it’s a major milestone for any supply chain to achieve. By creating a solid baseline, you're not only validating your data and assumptions but also setting the stage for optimization and continuous improvement.


Understanding Baseline Variations


Let’s break down the different types of baselines used in supply chain modeling:


1. Outside-the-Model Baseline

This baseline involves summarizing what the raw data tells us about what has actually happened in the supply chain. It’s your first step in collecting and dissecting data to uncover key insights. Depending on the scope of your project, you should analyze costs across various components (e.g., sourcing, production, warehousing, transportation) by product groups, customers, geography, time periods, etc. This step helps to validate the accuracy and completeness of your data by comparing it with sources of truth and business knowledge, essentially helping you identify missing or overlooked data.

Example: A manufacturer, while analyzing raw data to create outside-the-model baseline, finds that their warehouse inbound and outbound flows plans were not working since they missed capturing transshipments flow between warehouses.


2. Modeled Baseline

Once you've analyzed the raw data, the next step is to create a baseline model to reflect the current state within your supply chain design tool or mathematical optimization model. Unlike the outside-the-model baseline, which may include exceptions (such as stockouts or unusual shipping routes), the modeled baseline reflects the rules your supply chain should follow under normal operations. In this step, you’ll trim exceptions and create a model that assumes no major disruptions like stockouts, plant shutdowns, or natural disasters.

Example: While building the baseline model, the team realizes that inventory levels exceed warehouse capacities in certain periods. This occurred because they applied a simplistic inventory turn assumption at an aggregate product level, whereas a specific business line had a much faster turnover rate during peak season.


3. Optimized Baseline

After building your baseline model, the optimized baseline comes into play. An optimized baseline estimates what costs should be with your existing supply chain and compares them to what they really are, allowing you to find potential improvements without even changing your supply chain. While the structure of your supply chain remains the same (all nodes serving their intended roles), aspects like sourcing, production, storage, and product flows can be optimized for better performance. Depending on the scope, you may choose to optimize only certain areas, such as warehouse-to-customer assignments, while keeping other parts of the supply chain the same as in the baseline.

Example: After optimizing the baseline, the team found they could reduce warehouse storage costs by 14% simply by changing the product flow plan, without negatively impacting other costs.


Benefits of Baseline Modeling


So, what do these steps achieve for your supply chain?


1. Validate Data

You don’t want to make recommendations based on inaccurate or incomplete data. By analyzing data both outside and within your supply chain design software, you can ensure that the data is clean, complete, and aligned with the scope of your project.


2. Validate Business Assumptions

The baseline model is crucial for testing the assumptions your project relies on. Often, you may gather assumptions that seem valid but may not hold up under scrutiny. For instance, in a recent fleet optimization project, initial data gathering suggested that order deliveries couldn’t happen after 5 PM. However, after modeling the baseline, we discovered that some deliveries were feasible until 9 PM for key customers. Without validating this assumption, we could have constrained the model incorrectly.


In network design models, we often aggregate data (e.g., product groups), which can cause the loss of some important nuances. For example, in one scenario, aggregating all products as ‘fast-moving goods’ neglected the specific handling costs for certain temperature-sensitive items. By closing the gap between the modeled baseline and the outside-the-model baseline, you can ensure that you're not missing important rules or details in your scenario analysis while doing aggregations or filling data gaps with assumptions.


3. Assess the Impact of Exceptions

By comparing the outside-the-model baseline with the modeled baseline, you can measure the impact of operational exceptions. For example, frequent stockouts might lead to increased transportation costs as orders are fulfilled from secondary locations. Understanding the cost of these exceptions can surface the need for further advanced analytics, such as rebalancing inventory within your network or optimizing safety stocks using stochastic modeling.


4. Digital Twin

The baseline model is your first step toward building a true digital twin of your supply chain that can help you make better decisions. While there’s always an urge to capture every detail of your supply chain operations in a highly granular digital twin model, this often creates a 'black box' that becomes difficult to understand and manage. A well-constructed baseline model, however, allows you to test key decisions and their impact on KPIs without the hefty investment. Once your baseline is set, you can explore more detailed digital twin simulations going forward.


5. A Step Toward Agile Supply Chains

Supply chains are inherently dynamic and must adapt to uncertainties in both your business and those of connected partners (including all tiers of suppliers and downstream customers). Maintaining a baseline model allows for quick evaluation of different what-if scenarios and enables data-driven decision-making even during disruptions. By regularly updating this model, you’ll be able to assess strategic, tactical, and operational decisions with agility, helping you respond effectively to changes and uncertainties. In short, achieving an agile supply chain begins with agile decision-making rooted in a robust baseline model.


Best Practices for Baseline Modeling



Continuously monitor and bridge gaps between the outside-the-model baseline, the modeled baseline, and the optimized baseline


Here are some key practices to help you create a successful baseline model:


1. Create a Source of Truth for Your Analytics

Supply chain design projects often span the entire supply chain, so it’s critical to invest in collecting and cleaning data that will be useful not just for one project, but for ongoing analytics. Setting up a reliable database will enable future analytics projects and data-driven decision-making.


2. Automate Baseline Setup

Automate the process of creating your baseline model—from collecting and cleaning data to filling gaps, running the model, and generating insights. This way, you can keep an ongoing view of your supply chain, track costs of day-to-day exceptions against standard policies, and spot any deviations from expected results. In addition, comparing your baseline with an optimized baseline helps you catch inefficiencies early, before they impact profits. Once your baseline is set, aim to model key decisions regularly through quick optimization scenarios. Without automation, even a basic scenario test can turn into a lengthy project.


3. Enhance Data Models with AI/ML

Data coming from ERP systems often have gaps for modeling purposes. By integrating AI and machine learning techniques, you can identify data gaps and inaccuracies and correct them automatically. This ensures that your baseline model is built on the most accurate and complete data possible.


Takeaway


Establishing a baseline model is not just a step in the process, it’s a critical feature of your supply chain design project. Here are the key takeaways:


  1. Baseline modeling is a must-have, not a nice-to-have: Don’t skip this step, it’s fundamental to the success of your projects.


  2. Start with a foundational digital twin model based on optimization: All supply chains can benefit from a baseline model that provides valuable insights into KPIs, tradeoffs, and decision-making at a fraction of the cost and time of a full digital twin solution.


  3. Automation is key: By automating your baseline model, you lay the groundwork for a true digital journey that will continuously improve your supply chain.


Baseline modeling is not optional!


In short, setting up a baseline model is the first step in creating a smart, data-driven supply chain. Once you’ve established this, you’ll be well-equipped to explore more advanced optimization scenarios and other supply chain analytics.


This is one of my recommendations for supply chain transformation. Check out a related blog on Six Recommendations for Effective Supply Chain Transformation.


bottom of page