A global consumer packaged goods company launched a brand new product in their portfolio and wanted to ensure a successful product launch to lead to long-term adoption. Traditionally, the company created high-level forecasts for similar product launches, but they wanted to push the envelope and see if a more granular and dynamic forecast could be created to avoid potential supply chain disruptions.
Because of this, the CPG company wanted to create a week-store-product level forecast with essentially no historical data.
In just 24 weeks, phData was able to build an innovative forecasting model that incorporated trends in overall consumer behavior and sales trends in related categories to generate an accurate and granular forecast.Â
Because of the low granularity of this solution, big data processing techniques were leveraged on top of the Snowflake Data Cloud to generate the model output. The business users consumed the data in a clean Tableau dashboard.
The CPG company chose phData to build this solution for two main reasons: expertise in creating complex forecasting models and extensive CPG domain expertise. Not only did the team have strong technical skills to create an accurate forecast, but they also had a deep understanding of how CPG organizations function in terms of product launches.
phData was able to build a forecasting model with an aggregated forecast accuracy of approximately 90 percent.Â
The forecasting model was created to update weekly, incorporating new data from the product launch and adjusting accordingly to capture real trends post-launch. The CPG organization was extremely impressed with this level of precision, which allowed them to better manage their production and distribution of the new product.Â
The overall distribution execution of the new product was at an all-time high compared to similar product launches and the forecasting output was a key component for that.
Learn how phData can help your organization with Machine Learning.