Boris Evelson, VP, Principal Analyst, Forrester Research
Migrating Business Intelligence and Analytics Platform to Cloud
It’s vital to understand that data visualization is not a separate market or platform other than a few open source data visualization components often used in custom software development but rather a feature or capability of most modern business intelligence and analytics platforms, tools and applications. Typical considerations to consider before migrating business intelligence (BI) and analytics platforms to the cloud include:
-Cloud Benefits/Improved Business Agility: The ability to have BI applications up and running very quickly, as opposed to having to wait weeks or even months for hardware and software to be installed, is the key factor. But, it's also important to have new features and functions available to business users immediately, without any need for support, because new capabilities often support business innovation. Other highly rated benefits include improved collaboration, the opportunity to make BI capabilities available without having the required skills available in-house, and the ability to support large numbers of mobile and remote users
-Security and Compliance: In many ways, it would be concerning security issues weren’t top of mind with cloud migration, as companies rightly need to be concerned about the security of their data and compliance with applicable laws and regulations. That said, many businesses will quietly admit that they have greater confidence in the security capabilities of a top-tier cloud provider, as their own organizations would likely not be able to afford to employ the same level of expertise.
"Data visualization is a key step that businesses should think about when turning data into actionable insights and results"
-Benefits and Concerns of Cost and Pricing Models: Many see lower overall cost as a benefit, even though it isn't as high on the priority list as the advantages associated with agility and speed. Another factor in favor of cloud BI is the ability to substitute upfront cost with monthly payments. At the same time, cloud BI customers are concerned that the total cost of a cloud solution will actually be higher than a traditional software license plus maintenance after a number of years. However, it's hard to come up with a true cost comparison that takes into account all factors, including hardware, software, and total operational cost as well as business benefits such as speed and agility. Many also feel that pricing is unclear or complicated.
Data Modeling for Analytics and Visualization
Analyzing and visualizing data staged in Big Data hubs or lakes like Hadoop require a completely different approach. All DBMS (including RDBMS) have one common feature: a catalog. The catalog, or data model or schema, instantly makes a list of tables, entities and attributes available for analysis and visualization using a BI tool. Hadoop, however is not a DBMS; it’s a file system. While one can build and enrich Hadoop HCatalog, it’s not a required process and the richness of metadata available in HCatalog varies from organization to organization.
To make Hadoop-based data ready to be explored and analyzed, Forrester recommends a six step approach: ingest, discover, enrich, explore, move, and analyze. This approach will likely require different technologies and vendor platforms than BI used to analyze/visualize data in RDBMS.
Steps for Prospering with Data Visualization
Data visualization is a key step that businesses should think about when turning data into actionable insights and results, but the process differs per organization. There are four key steps that business leaders should follow to foster innovation and organizational growth:
1.Start with the Business Requirement: Start the data visualization process with a simple, yet not always obvious, question: do you even need data visualization? Simply looking up a customer balance or the number of widgets that your company sold yesterday does not require data visualization; a simple query that returns a number will suffice. However, if you're looking for a trend or correlation, analyzing a metric by various dimensions, or trying to convince an audience of your story, effective data visualizations can help you get better insights from the data. Make sure you select the right type of visualization to support your business requirement.
2.Refine the Business Requirements before Designing: Identify your business goals and objectives, define metrics and KPIs, map data visualization to specific audience types – for example, executives versus data scientists. Also consider different decision categories, such as strategic, operational, tactical, and finally, design for the right delivery channels, i.e. mobile.
3.Apply Data Visualization Design Best Practices: Apply best practices to the key components of data visualization: data and chart types and scales; color, text, and labels; layout and background; and interactive elements, including navigation, animation, and icons.
4.Use Your Data to Tell a Story: As the final step in the data visualization design process, decide if you need to organize the data charts, dashboards, scorecards, and infographics in a sequence that tells a compelling story for a presentation to senior executives, the board of directors, partners, and investors, for example. While many executives and analysts still prefer to use productivity applications like PowerPoint for presentations, an increasing number of BI vendors now offer interactive storyboarding functionality out of the box. These BI platforms contain multiple storyboard templates, which may include a simple sequence of data visualizations; a sequence with interludes or deviations into supporting subjects; inputs that lead to a conclusion; one cause with many effects; or timelines.
Impact of BI
BI technology is indeed not inexpensive. While some BI vendors are beginning to charge $10-$20 per month per user for BI “starter packs,” most large BI deployments still carry a hefty $2,500 per user license price tag. The good news is that an increasing number of BI applications do see significant ROI: 54 percent report ROIs between 10 percent and 199 percent within two years, and 48 percent achieve that ROI within two years.