One of a company’s main measures of success is through its customers. Nothing is more rewarding for us than learning about how data analytics experts are implementing a data-driven strategy.
I spoke with a handful of our customers across Government, retail, telecommunications and automotive organizations to find out how things were going and what lessons they had to offer to share with you! I hope to add more to these thoughts as we work with our customers and partners throughout the year.
Kamelia Benchekroun, a big data architect at the Renault Group says succeeding in the implementation of a Big Data strategy requires not only a scalable solution based on open standards, but a shift in the company’s mindset. “Big Data requires a truly disruptive approach requiring teams to be more agile with IT and operations need to work more closely, not in silos. Furthermore, to become truly data driven, it is important to accept the evidence advocating the implementation of one, unique data pool. It has to be based on an open solution and able to integrate data from both inside and outside the company. Once your data lake exists, you can build a catalogue to provide an easy and secure access for data consumers. It really is important to avoid numerous data lakes.”
Zog Gibbens, enterprise architect at Walgreens Boots Alliance suggests that companies need to make every effort to bring the IT and business users together to collaborate on best practice for your deployment. “It isn’t always easy but it’s critical. Equally, while traditional systems are better suited to a ‘waterfall’ approach where technology is introduced all at once, that’s not so true of big data projects. Through the collaboration between IT and the business, you need to focus on the goals and desired outcomes and introduce the technology in phases to reflect those.”
Luke Kay, EDW technical director at the Department for Immigration and Border Protection in Australia suggests keeping it simple for those who are just starting out. “Start with a defined business problem or need in the organization that complements your teams existing skill sets. We had some really good quick wins from providing solutions that our traditional EDW couldn’t solve. Good help is out there, obviously from Hortonworks, but with the pace of change in this area you need to be careful of dated advice on the internet – make sure you look into and ask questions on the Hortonworks Community website or you could be exposed to old information. Obviously no system is completely perfect from the get go, so there will be some level of rework – this may be in the form of bugs, but more likely it’s be based on your experience growing.”
Luca Olivari, Chief Data Officer of Contactlab, which is a leader in solutions for customer engagement, says executive support and trust is mandatory to get the ball rolling. “Data is an asset with implicit economic value so if correctly understood at the board level, the sponsorship should be forthcoming and your path simpler. Once you’re underway with Hadoop, process and store everything. Most of the data you need for actionable intelligence is already there so collect, collect, collect. That done, you must adopt a culture that explores the data without prejudice – look to be exploratory, not explanatory” Olivari thinks of data science as a lab and sets up multiple data ingestion points to explore enrichment algorithms and seek to understand online and offline behaviors.
Dipl. Ing. Dieter Knittel, Senior Project Manager of T-Mobile Austria breaks it down into three key areas – decision-making, culture and data management. “For decision-making you need to have a clear and accurate view of whether your organization sees data as a valuable asset and who your corporate sponsor it – senior management support is critical. For culture, any cultural change takes time and it’s important to identify clearly the competitive advantage of being a data-driven organization that’s no longer just a ‘good to have’ but a ‘must have’ “It seems an obvious point but when creating a big data strategy, it is essential to evaluate data quality,” you need to make sure to set the business towards willingly providing data of a solid standard. ““I’d summarise by advising to take care of governance, processes and roles from day one. Carefully plan your architecture and security concept but use an agile and iterative approach to deliver quick solutions based on current business needs. Do this by identifying specific high value opportunities as quick wins but never, ever lose sign of the big strategic picture. Start quick, think big, scale!”
Dr. Dirk Jungnickel, SVP Business Analytics, du says “My main piece of advice is to urge anyone starting out on an Analytics Programme, to drive an analytics-driven decision culture. Do not underestimate the importance and challenge of this. In many companies, executives may already struggle to properly interpreted descriptive analytics, i.e. reports and dashboards beyond the monthly financials. Making people understand and accept that business decisions nowadays should be based on machine-generated predictions or even be fully automated can be a huge challenge.”
I’m proud we are able to work with the best and the brightest across various industries as well as within the open source community to gather a wealth of knowledge and experience. Once again, customers really speak to who we are as a company and what value we’re able to create together!