From 2011 – 2018 the founders of BIEXA worked at Dagrofa aps, as CIO and Development manager where they participated in the following case:
Dagrofa is the 3’rd largest food retailer in Denmark with 1000+ selling points and approximately 500 shops in controlled banners and an annual turnover of approximately 3.5 Billion USD.
During 2016 a BigData and Datawarehouse strategy was developed. The ambition was to significantly reduce the Total Cost of Ownership (TCO) of the existing TeraData, Oracle and Cognos platforms and at the same time to create a foundation for delivering more value to business.
The main selection criteria for the latter was real-time processing capabilities, a Machine Learning tool set rated amongst best from Gartner and an architecture and platform with enough computational power to deliver true self service to business units.
During the selection process the following platforms were evaluated:
- Oracle Datawarehouse and Exadata
- Google Cloud Platform
Option 1. and 2. were deselected due to a really high TCO and significant upfront investments
Option 3. was deselected due scarce available competencies.
Option 4 was selected for Proof of Concept.
Following a successful POC it was decided initially move the TeraData Data warehouse with really large amounts of Point of Sales data and to build a market basket analysis solution on top. As part of the move to the Google Cloud Platform, it was furthermore decided to optimize the structure of the Data warehouse and move from an OLAP based model to Data Lake based structure.
The project lasted approximately 8 months from POC to full production and decommissioning of the old platforms and it delivered the following results:
- Cost savings in excess of 85% compared to the existing Entreprise platforms and estimated 50% compared to the evaluated Hadoop alternative
- Real time processing and extreme query speed, from days and hours to minutes and seconds (data sets with several terabyte and billions of rows)
- Self service analytics for end users based on Qlik Sense ODAG (On Demand App Generation) in a seamless integration to the Google BigQuery backend
- During development (and in all future projects at Dagrofa) the platform delivers an unmatched scalability and gives an unprecedented agility in development and implementation of new solutions
- The solution was established on the larges dataset in Dagrofa (XX TB of Data), and has proven that it is fully capable for the consolidation of the remaining Business Intelligence platforms (Oracle, Cognos, Microsoft).
- During late 2017 a near real-time fraud Machine Learning solution was developed in a really short time span to POC the platforms abilities in this area