A global delivery company sought to digitally transform their existing revenue and delivery forecasting process, which was heavily reliant on Excel-based macros. The data was gathered from transactional sources, aggregated, with a quick round standardization and enrichment and then forecasts were based on the standardized data. While the manual process was accurate, it lacked granularity (forecast at a customer level) and the flexibility to simulate different scenarios.
Stratacent implemented a solution that automates the sourcing of data, created rules for data standardization and enrichment and finally used the SAS forecast server to create granular customer-based forecast data. The forecasted data is then fed to their financial planning system for downstream consumption.
The company’s talented team of statisticians created a manual Excel-based solution that provided a very accurate forecasting data. However, the team lacked the technical aptitude to scale the solution up to a granular scale, run it in an autonomous mode or even forecast data based on various simulated scenarios such as a customer sale that drives up delivery.
Another challenge was data management. Besides being a manual process to source the data and standardize/enrich the data, the team was missing the proper definition for the source data. This often resulted in the need to manually correct data before and after the forecast process.
The team was so busy with this process that was run every other week that they didn’t have the time to innovate, create new products or bring efficiencies in the current process.
Stratacent was able to intervene with a highly effective solutions team. The SAS-based solution was designed with proper data flow and architecture. This solution architecture was devised from the ground up, with the company’s interests kept in mind throughout the process.
Our team started on two fronts. The first one was the forecast models. Using the sample data, the team immediately started working on creating models. The models were then tested with the actuals and iteratively retrained to come up with the accurate forecast data at a granular customer level. The second front was the data and automation. The team started working with their in-house system integration team, infrastructure team, and other vendor teams to define a data architecture. After business teams came up with the rules for data standardization and normalization, the rules were implemented using an ETL tool and the data was loaded into the SAS forecast server.
Our agile team worked with the customer to set up the project charter, vision document, various epics, and user stories.
Being cognizant that most models, if not operationalized, end up on the shelf, our team worked on creating the wrapper shell script to deploy the models in production. This wrapper script allowed us to create an end-to-end automated solution.
Our consultants also acted as educators for the company’s employees. Stratacent deployed SAS education to properly train the employees on SAS products. The team also helped them learn how to properly use their new system and walked them through the path to migration on the new system, devising better solutions every step of the way toward the final goal.
Technologies Used: SAS 9.4, SAS Forecast Server, Linux, Teradata, Oracle and Shell Scripts
The result was an end-to-end automated solution that created the forecast data at the customer level and fed it into their financial planning application. Furthermore, with the staff properly trained into using the SAS forecast studio, that they leveraged the forecast models in a simulation lab where the team can do ad-hoc research with various scenarios.