E-COMMERCE: RETAIL FORECASTING
Customer
Our client is an online jewelery ecommerce business that offers a remarkably wide selection of diamond, gemstone, and pearl jewelry. They source and manufacture diamond and gemstone jewelry from specialized markets around the world at the best prices.
Challenge
Our client also wanted to build a pipeline that automatically provided weekly forecasts of the next month for each store and product in their inventory.
It was difficult to understand where marketing dollars should be spent.
Also, it was difficult to understand the relationship between actual sales and the marketing dollar spent
Solution
BI Demantra created a cloud based solution using HDInsight with Azure SQL Database, Azure Data Lake, Azure Data Factory, and Azure Machine Learning experiments and notebooks. We blended historical data from various sources and selected a feature set.
Model and predictive algorithm was adjusted several times after the model was trained and results were compared to the actual sales.
Result
- In retail business , accuracy and timely forecasts drive success. Our client was able to see relevant data within a matter of weeks and the model was adjusted after comparing results.
- Additionally, marketing dollars were invested in the right products to follow the retail forecasting predictions.
- ROI was realized as a decreased cost of holding goods and increased values in marketing dollars thus maximizing revenue.