Enterprise Data Warehouse Optimization

  • Challenge
    Data volumes are growing at least 60% each year. And using big data analytics, retailers can increase operating margin by more than 60%1. But the traditional enterprise data warehouses these retailers use for analytics cannot keep up. For a leading retailer, a new analysis was taking 12 weeks—much too long for the business to wait. In addition, IT faced purchasing additional capacity for the data warehouse that could cost an additional $1M or more.

    In order to create faster analyses and avoid high hardware costs, the retailer needed to offload data processing to Hadoop. The retailer has pricing, transactional and operational data that changes on a daily, often hourly basis. In addition, this data was spread across devices, mainframe, Oracle Exadata, Netezza, and Teradata. In order to do that, they needed a product, Datameer, both business users and IT could use to quickly analyze large amounts of data.

    Business users need data to understand seasonality of products and competitive pricing in order to target customers better. In a pricing analysis project, the company used Datameer to join data on the most commonly clicked on item categories (on the retailer’s website) with historical data on items purchased during previous holiday seasons. As a result of the analysis, the retailer was able to change their website to highlight the most popular items for the holiday season—in time to convert increased holiday traffic into revenue.

    IT needs to monitor devices and networks in order to prevent outages at peak times. Using pre-built analytic functions, the retailer used Datameer to identify patterns in the device logs coming from the WLAN controllers, mobile devices, routers and firewall devices in the stores. Through the analysis of the impact of devices on the network, bandwidth consumption and crashes of applications on the devices, the company was able to understand what contributed to network performance. As a result, they proactively identified and then reduced the number of network failures by 30%. For this project, the retailer saved one resource and 5 days of requirements gathering time.

    • Reduced analysis time from 12 weeks to 3 days
    • Reduced number of network failures by 30%

    Big Data Processing

  • Reduced network
    failure by 30%

  • +1.800.874.0569