Catering to your customer means knowing their needs in detail. Studying transactions is one way to get those specifics, but for a global consumer paper goods leader, the task of sorting through millions of individual receipts from tens of thousands of retailers was Herculean.
More and more manufacturers have access to Big Data – complex and massive collections of information nearly impossible to process with traditional methods – but few have systems in place to make use of it. Through a six-week proof of concept, BlueGranite showed this American multinational firm how employing Hadoop’s substantial storage capabilities and mammoth processing power can make analyzing Big Data a breeze. A follow-up demonstration project showed the manufacturer some of the many insights predictive analytics offer.
The colossal corporation manufactures goods on a major scale for big box retailers, who in turn sell to individuals. The manufacturer obtains sales data from those retailers. Finding and analyzing the sales trends, item popularity, and other valuable insights buried in that data can be key to boosting the bottom line.
The manufacturer was receiving terabytes of weekly line-item-level transaction data. To put that into context, it would take 212 DVDs just to store a single terabyte. The company’s technology only had the capacity to handle a few weeks’ worth of information at a time. The manufacturer wanted to explore Big Data processing solutions and brought us in to demonstrate.
We took a massive quantity of the manufacturer’s data and loaded it into Microsoft Azure HDInsight – a cloud-based Hadoop platform. We then used Power BI to dispay different ways of consuming that information.
The HDInsight concept removes the need for an expensive Hadoop cluster, and many of the administration requirements. The platform-as-a-service product lets customers easily create clusters to do massive amounts of processing without the large initial capital investment of an on-premises solution.
We coupled that demonstration with a proof of concept showing the manufacturer some of the many benefits of Azure Machine Learning (ML) – a tool designed to be an entry point to analytics. The cloud-based predictive learning studio allows novice users to upload, cleanse, and transform data, and, from there, to create predictive modeling. Azure ML can also reduce the need for complex code and make the jobs of those who write it - statisticians or PhDs - easier.
By using point-of-sale data and product shipment quantities, BlueGranite built a predictive model using Azure ML that allowed the manufacturer to predict when a location might run out of stock and to proactively send more product before that happens.
We collaborated with this company to give a glimpse into a few ways different platforms could benefit its business. If you are ready to explore the best way to harness your organization’s data, we’d love to partner with you! Contact us today for more information on how we might be able to help.