Amazon Supply Chain and Logistics

How to use Demand Planning on AWS Supply Chain for better forecast accuracy

Micro- and macroeconomic conditions, coupled with evolving customer demand patterns, exert persistent pressure on supply chain networks. Organizations must strive for greater efficiency and agility and proactively improve supply chain functions to improve performance and meet customer demand. Supply chain management systems include different functions that coordinate resources and business processes. Demand planning plays an important role in effective supply chain management, ensuring timely stock replenishment, enhanced capacity management, and optimal sales and revenue.

We previously described how AWS Supply Chain improves supply chain visibility to increase resiliency. This blog post covers AWS Supply Chain Demand Planning, a purpose-built demand planning module that enables accurate demand forecasts, adjusts to market conditions, and continually learns from changing demand patterns and user inputs to increase plan accuracy.

The demand planning process

Accurate demand forecasts are essential. Poor forecasting can lead to inventory imbalances, causing overstock or stockouts, increased costs, missed sales opportunities, and potentially diminished customer satisfaction. For example, over-forecasting creates excess inventory, reducing available cash flow while increasing storage costs. Under-forecasting leads to products being out of stock, which affects customer experience, customer satisfaction, and customer loss.

Demand planning is pivotal for organizational success. It dictates how, when, and where resources are allocated, ensuring customer demands are met without overextending resources. Through systematic evaluation, forecasting, collaboration, and regular review, it shapes the supply chain strategy and operational efficiency. The typical demand planning process includes a series of steps that can be summarized in the following categories:

  1. Data integration: This is the foundation of demand planning. It involves gathering historical sales data, current order data, inventory levels, and other relevant metrics. Data sources can include enterprise resource planning (ERP) systems, customer relationship management (CRM) tools, and external market intelligence reports. Integrating diverse data streams is critical to ensure a holistic view of demand factors.
  2. Forecasting: After data is collected and integrated, statistical models and algorithms are applied to predict future demand. Forecasting currently combines empirical data with industry knowledge and is where machine learning (ML) offers the greatest potential improvement.
  3. Collaboration: Demand planning involves communication and collaboration among departments such as sales, marketing, finance, and operations. By bringing together insights from different teams, companies can reconcile statistical forecasts with market intelligence, promotional plans, and strategic initiatives. This collaborative approach ensures that the forecast represents a consensus view, thereby increasing its accuracy and acceptability.
  4. Continuous review and adjustment: Demand planning is an iterative process. As actual sales data comes in, it’s compared against forecasts to identify variances. These discrepancies are analyzed to refine forecasting models and adjust future predictions. Regularly reviewing and adjusting the demand plan ensures it remains relevant and reflects the latest market conditions and internal company strategies.

Each of the four categories relies on the others: without accurate data, forecasts are flawed; without collaboration, forecasts lack depth; without continuous review, even the best forecasts become obsolete. This interconnectedness ensures that the process remains dynamic, accurate, and relevant, reflecting both the realities of the market and the aspirations of the business.

AWS Supply Chain provides you the following benefits beyond traditional demand planning:

  • Automation: Demand Planning eliminates numerous, non-value-added manual tasks such as data entry, calculations, and adjustments. This enables faster demand forecast creation time and the reduction of errors caused by manual tasks.
  • Harnessing the power of ML: ML analyzes historical sales and real-time data (such as open orders) to create forecasts and adjust models to improve accuracy. This improves forecast accuracy and reduces the risk of too little inventory (that is, stockouts) or too much inventory (that is, excess inventory). AWS Supply Chain also uses ML for lead-time variability detection to improve supply planning accuracy.
  • Collaborate efficiently: In-application collaboration capabilities to facilitate consensus with other team members. This also improves coordination, enables faster decision making, and reduces the risk of error.

The next section builds on the discussion of the demand planning process by outlining how AWS Supply Chain supports the journey. This section focuses on the first three critical phases of the process (data integration, forecasting, and collaboration) and provides a step-by-step guide for new users. These steps detail how to seamlessly set up AWS Supply Chain, harness its advanced capabilities, and transform your demand planning approach. Future blogs will delve into the fourth phase, continuous review and adjustment, by detailing its integration with AWS Supply Chain based on industry best practices.

AWS Supply Chain Demand Planning prerequisites

  1. You must have an AWS account. If you don’t have an AWS account, you can follow the account creation process in How do I create and activate a new AWS account?
  2. You also need an AWS Supply Chain account. If you are not currently a customer, visit AWS Supply Chain to learn more and get started.

Setting up Demand Planning

  1. The first step is data ingestion into the AWS Supply Chain data lake. The data lake uses ML models to understand, extract, and transform disparate, incompatible data into a unified data model.
  2. As an admin, hydrate the AWS Supply Chain Data Lake with the required data entities for Demand Planning to generate forecasts. The Data fields required by AWS Supply Chain applications in the AWS Supply Chain User Guide provides the full list of required fields to generate forecasts. Note that for more accurate forecasts, you must ensure that the optional entities in the data set are populated. These fields are described in the Data fields optional for AWS Supply Chain applications.
  3. After data is ingested, you must check the user permissions. As an admin, you can add the desired number of users to Demand Planning and then provide the necessary user permissions are managed. New user invites are sent after permissions are set. At the bottom of the screen, admins can select one of four role-based permission levels (Admin, Data Analyst, Inventory Manager, or Planner).The illustration shows use management settings and permission roles.
  4. After user permission roles are selected, choose AWS Supply Chain on the left navigation pane and then choose Get Started to launch the Demand Planning module.
  5. The following screenshot shows the next step in the process where you specify the forecast generation timeframe (called the planning horizon). You enter the forecast interval and length of the interval that you want the application to create a forecast plan for. If you are interested in a monthly forecast for the next six months, select Monthly for the time interval and enter 6 for period.The illustration shows demand planning horizon, i.e. time interval and horizon.
  6. The next step allows you to configure the forecast granularity depending on your preference or forecast need. On this screen, you select the level at which you want to forecast by choosing the hierarchy attributes for site, channel, and customer.The illustration shows demand planning hierarchy.
  7. After the preceding steps are complete and you have set your forecasting scope, you prepare your data so AWS Supply Chain can generate forecasts. You need to configure your dataset by selecting how to handle negative values in your dataset.The illustration shows dataset configuration settings.
  8. The final step is selecting the Amazon Simple Storage Service (Amazon S3) bucket where demand plans will be published. After the location is selected, choose Continue to generate the forecasts, as shown on the following screenshot. The illustration shows enterprise settings and demand plans publishing.

This starts the demand planning process and generates the outputs shown in the following screenshot.

The illustration shows demand planning output, i.e. a demand chart by month.

­Using the interval period you selected in Step 5, the forecast demand is displayed both as a graph and as a table. The graph contains a dashed line that represents the prior year’s demand and a solid line that represents current data. The current demand line includes the historical demand data and the forecast for the remaining months in the interval period you selected in Step 5. This view of historical demand coupled with the prior year’s demand data helps you adjust and finalize the ML-recommended forecast and reach a consensus demand plan that can be used for your downstream system, including supply planning. You can also change the product you want to view from this screen.


Supply chain networks are consistently tested by fluctuating market conditions and customer demands, requiring organizations to be proactive and agile. AWS Supply Chain Demand Planning addresses these challenges, assisting organizations to be not only responsive but also anticipatory in their supply chain strategies.

The demand planning process of AWS Supply Chain bridges the past, present, and future of businesses. By ensuring optimal resource allocation and customer satisfaction, this process involves systematic steps, from data integration to continuous adjustment. This highlights its iterative nature and importance in achieving improved accuracy, reducing inventory imbalances, and providing optimal inventory levels.

Beyond traditional demand planning, AWS Supply Chain introduces automation, eliminating tedious manual tasks such as data entry, calculations, and adjustments. By harnessing the predictive power of ML, it refines forecasting and adapts to real-time conditions, striking a balance between stock availability and efficient cash flow. AWS Supply Chain also emphasizes cross-departmental collaboration and fosters a consensus-driven approach to align various business units towards a common demand forecast. This collaborative effort underscores the significance of interdepartmental harmony in supply chain management.

AWS Supply Chain combines advanced technology with simplicity, providing an innovative upgrade to demand planning. This enhanced system can be used to help increase forecast accuracy, improve operational efficiency, and enable proactive supply chain management, positioning it as a vital strategy for modern businesses.

AWS Supply Chain is available without any up-front licensing fees or long-term commitments. It provides a scalable solution that scales with your needs and Demand Planning is available to all AWS Supply Chain customers. Visit AWS Supply Chain to learn more and get started. You can also visit the AWS Workshop Studio for a self-paced technical overview of creating an instance, ingesting data, navigating the user interface, creating insights, and generating demand plans.

Harold Abell

Harold Abell

Harold Abell a is a Senior Product Manager for AWS Supply Chain. Harold is one of the founding product managers for AWS Supply Chain and involved with the concept and design of the application. Harold has over 10 years of industry experience in supply chain and software development with both General Electric (GE) and Amazon Web Services (AWS). Harold graduated from Brigham Young University with a BS and MS in Manufacturing Engineering and an MBA from the Fuqua School of Business at Duke University. In his spare time, Harold loves all things outdoors, including snow skiing and boating with his wife and three girls.

Vikram Balasubramanian

Vikram Balasubramanian

Vikram Balasubramanian is a Senior Solutions Architect for Supply Chain. In his role, Vikram works closely with supply chain executives to understand their goals and problem areas and align them with best practices in terms of solution. He has over 17 years of experience working with several Fortune 500 companies across different Industry verticals in the supply chain space. Vikram holds an MS in Industrial Engineering from Purdue University. Vikram is based out of North Dallas area.