During the holiday season, the ability to analyze and react to data effectively can be the deciding factor between having your shelves stocked or scrambling with inventory transfer issues. Historic data can provide the basis for true statistical forecasting, accounting for seasonality, and other retail trends. Cleansing the data will provide a consistent baseline and serve as a foundation for analysis, which can be presented using best-in-class dashboards to empower your business. Using Anaplan, retailers now can sense trends and pivot accordingly to make the holidays as successful as possible!
“The earliest GPS was essentially a digital map that led drivers through a predetermined route faster and with greater ease than a paper map. But as early GPS could not account for traffic jams, road hazards, or other variables, it did not warn drivers to pivot or course-correct to account for the unplanned. Similarly, many companies are currently making major investments in digitizing and automating their supply chains to make them better informed, more frictionless, more cost-efficient, and hence more capable.” Excerpted from “Why Supply Chains Must Pivot,” MIT Sloan Management Review.
In some ways, holidays are the most reliable times of the year. They take place the same time every year, they are focused on the same type of celebration—whether it be mothers, fathers, seasons or moon cycles—and they involve the same groups of people. Conversely, they are the definition of chaos for most retailers when it comes to the forecasting process. In the period leading up to a holiday, supply chain demand can increase exponentially, creating broad shifts in timetables and priorities.
The ability to sense and pivot to meet changes in demand is critical. By using historical data and a robust statistical forecasting model, Anaplan can help transform the holidays into something to look forward to for retailers.
Here's a look at how a multinational beauty company leveraged Anaplan and data best practices to optimize the supply chain forecast across more than 3,000 retail outlets.
How Did Last Year's Data Impact This Year's Decision?
The company had a wealth of prior holiday data, but its challenge was to mine that data for insights on how to adjust its supply chain to achieve a sustainable advantage for the coming year. To create a global forecast that was both accurate and valuable, we used a statistical forecasting model built in Anaplan to forecast key Mother’s Day SKUs across more than 3,000 stores.
The model analyzed the data and applied 30 statistical forecasting methods to identify the best-fit method that would produce the most accurate forecast. Because the Mother’s Day SKU data set was very seasonal, methods that performed better with seasonality—such as additive and multiplicative decomposition—were often identified as the best fit based on error percentage.
Once the demand forecast was produced, an inventory deployment model was constructed to integrate with the demand forecast. The inventory deployment model used the demand forecast—along with safety stock and cycle stock calculations—to identify a right-sized inventory for each store based on the remaining demand. The model then used supply chain assumptions, such as lead times from distribution centers to stores and minimum order quantities, to create inventory deployment recommendations in support of the right-sized inventory value.
As part of this initiative, we conducted a study to understand the financial impact of using the inventory deployment recommendations from Anaplan, as opposed to the actual purchase orders that were submitted by the company’s retail outlets. The study estimated that the Anaplan-generated inventory deployment recommendations offered an opportunity to improve net revenue by between 7 percent and 16 percent based on a reduction of stock-outs and excess inventory levels.
All retailers have data and most have too much of it. Having data that is clean and mapped correctly can facilitate a smoother forecasting process. This can provide the clarity and normalcy needed to get through the sometimes-volatile holiday season.
For new SKUs with no history, we used Anaplan to create a mapping to similar SKUs to construct a data set on which a forecast could be built. This allowed us to accurately build out a fully loaded forecast utilizing all the SKUs available. In addition to utilizing SKU mapping, we used Anaplan to normalize lost sales due to stock-outs to create a more reliable data set.
The Benefits of Building Forecasting Dashboards
To harness the power, flexibility, and scalability of Anaplan, we created a set of dashboards using the new user interface. The dashboards provided key insights that allowed our client to easily understand the recommended shipment allocations and supporting data, as well as how the current season was trending weekly compared to previous years' sales metrics. These insights gave them the ability to sense in-store demand more accurately and pivot their replenishment orders accordingly.
Data is Your Saving Grace
Like most holidays, the past can be a blessing and a curse. Having historic data in your model can provide your organization with the ability to sense and pivot to accommodate increased demand and address inventory issues. Normalizing your data to create a consistent foundation and using Anaplan’s statistical models to accurately analyze your data, can provide your organization the competitive advantage needed to survive the upcoming holidays.
What are some other methods you use to make your seasonal forecast more insightful? Leave a tip below and let’s discuss it!
Brian Gallagher is a Manager and Anaplan Solution Architect at Cervello responsible for helping clients design & deliver Anaplan models. He is also a business process subject matter expert with experience re-designing processes based on industry standards.
Brian has over eight years of finance & technology experience, the last four of which have been focused on process re-design & delivering Anaplan solutions. He has served as the implementation lead on multiple enterprise-level clients. Brian earned a B.S. in Finance from Fairfield University with a minor in Mathematics. He has also earned his MBA from the University of Massachusetts.