For companies that depend on an accurate forecast to achieve financial goals and operational efficiency, adding Anaplan PlanIQ™ to a current Anaplan model proves to be more efficient, effective and accurate when it comes to making forecast predictions on multiple levels.
Accelytics’ client is a successful retailer of premium beds and mattresses that is a long-time customer of Anaplan. They have implemented many successful use cases on the Anaplan platform that span across several important business planning processes.
Several of the use cases rely on the Anaplan statistical forecasting model as a fundamental input to a collaborative planning process. When Anaplan began offering their customers access to Anaplan PlanIQ™, a forecasting technology that Amazon uses for its enormous retail operations, the client was intrigued to pilot a trial.
The promise of Anaplan PlanIQ™ is that it will allow companies to improve the accuracy and quality of their statistical forecasting, while at the same time, reducing the time and effort required to get those forecasts. The retailer can benefit greatly from this technology if it can be realized in their current real-world environment, and can also be deployed in their use cases to replace the existing Anaplan statistical modeling functionality.
This report details the trial goals, assumptions and compelling results.
TRIAL OUTLINE
The project team designed a study that will test forecast accuracy from 3 different forecasting levels: topline sales, market level and product class. Several historical time periods were defined to help test the effects to the statistical forecasts that may be related to changes caused by the COVID-19 pandemic. All the forecasting model options that are available in Anaplan PlanIQ™ will be tested and compared to previous forecasts and actuals.
The Anaplan PlanIQ™ system uses cutting edge technology in the various models including ML (machine learning) and AI (artificial intelligence) developed and used by Amazon. Many of the models can also utilize related causal data to improve the statistical forecast accuracy. The client identified several causal data streams to test including marketing spend, website clicks data, promotional calendar, and open store days.
TRIAL GOALS
RESULTS
The trial results were very positive. The project team has found that setting up forecasts in Anaplan PlanIQ™ is significantly easier than doing the same in the classic Anaplan statistical models. We have also found that the accuracy of the forecast, especially on a daily level, which is an unusual requirement unique to the retailer, is extremely accurate.
1. The retailer found that the Anaplan PlanIQ™ application itself is easier to use than traditional forecasting methods in Anaplan due to requiring less manual intervention.
2. Anaplan PlanIQ™ is more accurate on the following levels:
3. Anaplan PlanIQ™ has useful functionality not available in regular Anaplan forecasting, including:
4. Anaplan PlanIQ™ showed high accuracy with analyzing the effects of holidays and the promotional calendar. The graph below showcases the forecasting results for the Labor Day Sale.
5. Anaplan PlanIQ™ was used to test and compare results during the COVID-19 pandemic. The system proved to be effective in determining demand shifts during economic events like the pandemic.
Due to the success of the trial, the client has implemented Anaplan PlanIQ™ into their current model to reap the benefits outlined in this report.
To learn more about Accelytics and Anaplan PlanIQ™, contact: [email protected].