Case Study

Client Background Client is a US based 3 billion USD plus footwear designer and distribution brand with 140 company owned stores and selling in 50 countries.  Business Objective Given the disruption caused by Covid pandemic it’s imperative for the customer to break the data silos and incorporate relevant data sources for their demand and supply planning. It’s further required to break away from the manual process of planning by various stakeholders like demand planners, supply planners, S&OP teams etc. To achieve this customer is looking to implement AI/ML driven integrated business planning solutions for supply and demand planning. Solution Valiance in its role as an implementation partner for IBP solution is responsible for ensuring that datasets required to drive various IBP modules, namely demand planning, supply planning, control tower etc are properly understood; documented and mapped. Thereafter source datasets need to be picked from sources like SFTP or via API calls, cleaned, transformed and fed into the destination system. To be able to achieve the above, our team of data architects and data engineers worked with the business and technology team to propose a data architecture based on MSBI technologies. MSBI stack was selected considering the existing landscape of technologies. Team narrowed down the data mapping, data models for staging and data load frequencies for different datasets. Over the course of next 6 months we implemented SSIS routines for demand planning addressing below scenarios Data transformations considering downstream needs. Duplicate records and bad data quality Recovery from job failures Notifications for success and failure Audit logs for traceability SSIS data pipeline processed 50 million & 1 million records for historical and incremental runs respectively. ❝ Valiance’s leadership team has been engaged right alongside their resources making a positive difference in a very difficult project. They are going above and beyond to make the project successful, and I appreciate their contributions very much. Team is knowledgeable and dedicated to the customer’s success -SVP for Apprarel and Fashion, Global supply chain ISV Outcome Successful data integration for demand planning ensures successful demand forecasts across various product categories, styles, geographies and SKU’s. This will further enable proper supply chain management, sourcing of materials, production planning and distribution. At Valiance we have developed precise understanding of relevant data for AFL companies w.r.t demand planning, supply planning, S&OP & IBP.  This capability, combined with our ability to correctly analyze all the required data as a result of our expertise in data science enables us to help such clients increase demand forecast accuracies that directly result in better sourcing precision, and more optimized manufacturing & distribution.

Client Background The client is one of the largest brewing companies in the United States and employs over 30,000 people. It was the world’s largest brewing company based on revenue, but third in brewing volume, before its acquisition. The division operates 12 breweries in the United States and 17 others overseas. Business Objective The client wanted to replace the legacy APO system and leverage machine learning forecast capability which can automatically enrich the forecasts with external drivers and minimize the manual enrichments. Solution Our data science team worked with client and ISV to acquire following datasets needed for forecasting on google cloud platform. Orders: Data at a weekly level for each of 10,000 SKU’s Macro-Economic Drivers: Consumer Price Index, LCU per USD, Retail Sales Index (USD), Industrial Production Index and Inverse Exchange Rate. Weather Drivers: Weighted Precipitation, Snow, Weighted Temperature, Max Weighted Temperature and Min Weighted Temperature. Marketing Spend Drivers: Digital Media Buying, Brand Events, Traditional Media Buying and Brand Promotion. Promotions: We have two types of promotions in Canada, price drop promotions and goodies promotion.   Google Data proc was used to process and prepare the data needed for model training along with Vertex AI for model training. Highly level summary of modelling process we followed is indicated below Finally demand forecasting capability we developed had the Operational forecast – weekly forecast. The level of forecast was at Item, sales channel and monthly level. This has been implemented for Orders, Shipments and POS for the regions USA, Mexico and Canada. Outcome Forecasted values were generated in supply chain solution to provide granular view at Region, province, item and customer level. We were able to achieve accuracy levels for 90% percent for 12 weeks forward forecast on 20% percent of SKU’s that drive roughly 80% of customer business. For remaining accuracy levels were 65%.

Client Background Client is a Mexico based retail giant which operates largest department stores in Mexico. It is a USD 4.5 billion company having 35,000 employees. Business Objective Implementation of a sophisticated AI/ML driven IBP platform for Merchandise Financial Planning, Assortment Planning, Replenishment, Allocation, and Demand Planning to achieve business goals. This implementation will help to achieve. Annual projections to sales LP, turnover margin. Determine OTB. Synchronize with daily & weekly goals for salespeople. Reconcile with global plan and channel mix. Reconcile with assortment. Solution Valiance in its role as an implementation partner for IBP solution is responsible for ensuring that datasets required to drive various IBP modules, namely Merchandise Financial Planning (MFP),Assortment Planning (AP) etc are properly understood; documented and business points are mapped. Thereafter source datasets need to be picked from sources like Google Cloud Storage(GCS) with the use of gsUTIL commands written inside batch files , validated, cleaned, transformed and fed into the destination system. To be able to achieve the above, our team of data architects and data engineers worked with the business and technology team to propose a data architecture based on MSBI technologies specially SSIS. Team completed the data mapping, data models for staging and data load frequencies for different datasets. Over the course of next 12 months we implemented SSIS routines for MFP, AP addressing below scenarios Data fetching from Google storage as a UTF-8 compatible text files and dumping into staging tables. Data transformations between staging and output tables making sure correct Spanish data to be pushed. Duplicate records and bad data quality. Recovery from job failures. Notification email for success and failure. Audit logs for traceability. Outcome Successful implementation of data integration will serve as the foundation for the second phase of the project where the process will be enhanced with analytics which will improve forecast accuracy, increase inventory turns, reduce standing orders and lost sales. At Valiance we have developed precise understanding of relevant data for AFL companies w.r.t demand planning, supply planning, S&OP & IBP.  This capability, combined with our ability to correctly analyze all the required data because of our expertise in data science enables us to help such clients increase demand forecast accuracies that directly result in better sourcing precision, and more optimized manufacturing & distribution.

Client Background Client is a global well-being company retailing, distributing and manufacturing a portfolio of leading international and home-grown brands across sport, food and health sectors. Business Objective The client has both seasonal and core lines, with overall SKUs being more than 200,000 and over 500 retail stores. This SKUs also include products which are launched as New Product Initiatives This in turn, resulted in a non-traditional forecasting technique (machine learning)., having the ability to incorporate multiple external variables*, such as google mobility, weather events, covid, promotional flags, floating calendar etc. Solution The project was executed in 3 phases: Phase 1: It consisted of the Data Collection and Harmonization – Phase 2:  It focused on the Exploratory Data Analysis & Segmentation –   Phase 3: ML model building/ iteration followed by model training & validation The third and the final phase constituted of building and validating the ML models. Following modeling techniques were used: Linear Regression: These models are easier to interpret and debug, so it is a good starting point. XGBoost: This is a tree-based approach and uses boosting technique. Boosting is a homogeneous weak learners’ model. learners learn sequentially and adaptively to improve model predictions of a learning algorithm. Random Forest: It uses bagging technique. It is also homogeneous weak learners’ model that learns from each other independently in parallel and combines them for determining the model average. Gradient Boosting Light Gradient Boosted Machine Model validation in forecasting: The dataset is divided into training and validation data sets. We go backward for some time period (say weeks) and try to forecast for the same time period using historical sales. Post which, we evaluate the model with the help of Actual sales vs Forecast Sales for the Validation data set Below are some external variables that were taken into consideration: Macro-economic: GDP, Inflation, Industrial Production, Unemployment rate, Inverse exchange rate. Holiday Information: Festivals, National Holiday, other holiday Floating Calendar: Includes special event like sports league Promotional Details: Includes discounts and offers on Items Item level Information: SKU, style color, sub-class, category; item attributes: polo grey color Weather Information: Temperature, Wind speed, Humidity, Rain Marketing Spend: It considers expenses through various ways of marketing, for example digital media, traditional media , Brand events. Mobility Data from Google/Apple: Traffic of mobile phones over different places like parks, residential areas etc. Events happening (Sports) Special days: Thanksgiving Day, Christmas, new year, black Friday etc. Product Attributes:  Size, color, design etc. Sales channel Store format Covid Data: Total confirmed cases/tested cases for any region Other specific variables: holiday calendar, promotions, markdown, Mobility Data from Google/Apple: Traffic of mobile phones over different places like parks, residential areas etc. Outcome 80% improvement in forecast accuracy 8X reduction in manual intervention 40% reduction in inventory value