| | March 20179CIOReviewprioritizations. Second, a prioritized and focused effort tends to result in shorter time to value, which will serve as a proof point for the future expansion. Third, it is specific enough to establish a baseline, measure the pain point, and thereby demonstrate the ROI (Return on Investment).Three Important Steps It's a tricky balance to start small, yet have the foresight to:· Build out the infrastructure for reuse· Plant the seeds early for metrics and measurement; and· Keep it open and easy to use in order to maximize "user freedom." A common infrastructure will make it possible to reuse and expand into multiple use cases with higher investment efficiency. While many large organizations tend to start their data and analytics projects from the common infrastructure, I strongly recommend starting with a real business problem, but then keep in mind the extensibility and flexibility of the architecture so that you can reuse and expand.Metrics and measurement are eve-rything in terms of ROI, and are also are the key to avoid wasting large exercises of data analytics. Organizations may be inclined to delay the investment in this area, treating this as "internal" or "technical debt." The reality is just the opposite. In our discussions with customers and industry analysts, the earlier you can establish and track metrics, the higher chance for a successful project as you now have the ability to measure and to pivot based on the measurement outcome. Lastly, the fear of losing control is common in organizations, leading to bottlenecks in each step of the data analytics operations. When users can't easily use the data analytics capability or cannot count on a reasonable turnaround time, they walk away to seek alternatives. This leads to "shadow IT," reducing the value of the entire data analytics effort across the organization. Better that experts assist users in effectively taking advantage of data analytics.Speeding Time to ValueWith organizations having to identify a real business problem and taking care of infrastructure, metrics and users, they should leverage cloud computing and open source technology to get started, so that they don't have to create everything from scratch. In other words, once your organization has identified a specific business problem that needs to be solved via data analytics, focus on the data and the problem statement, not the infrastructure and full-blown production environment. Cloud service gives you the freedom to do this.These days it's very cost-efficient to get started, shielding you from the complexity of infrastructure, at least temporarily. Furthermore, you can lean on open source technology in the data analytics space to get to some quick time-to-value, whether it's an open source visualization tool, or open source machine learning algorithms.We are now in an exciting time for data analytics indeed. Howev-er, the excitement doesn't come without its associated challeng-es. The key is to start small with a real business problem and leverage cloud computing and open source technology to great advantage. The key is to start small with a real business problem and leverage cloud computing and open source technology to great advantageJin Zhang
< Page 8 | Page 10 >