Buyer’s Guide: Business Intelligence in the Cloud
Updated · Mar 22, 2012
Business intelligence (BI) vendors are involved in a headlong rush to provide their software in the cloud. The drivers include anywhere access, improved scalability, reduced implementation time and lower upfront costs. But users have concerns about vendor lock-in, performance issues, security and even software ownership.
So should companies dump their existing on-premise BI software and move to the cloud? We asked some experts to give users sensible advice, as well as an overview of the market. In addition, we picked a few interesting cloud BI offerings from literally scores now on the market.
“Most BI vendors are incorporating cloud offerings in 2011/2012, so a list loses its value,” said Barry Cousins, senior research snalyst at Info-Tech Research Group. “Large vendors have taken notice of Birst because of its flexible ‘On-Premises/On-Demand’ model that allows customers to easily shift components of the solution across the firewall. Also, the big vendors are trying to emulate QlikView’s flexible data access for end users, ease of use, and HTML5-based mobile UI (user interface).”
Flexibility, he believes, is the key to cloud success. That means making it easy to transition parts of the business intelligence solution between on-premises and on-demand, in either direction at any time, backed up by friction-free licensing models. In addition, users want less rigid data models, allowing the user to blend structured corporate data with external unstructured data.
He doesn’t advise companies to leap into the cloud with the expectation that they will save lots of time and money right away. He advises them to get ready for a financial outlay and to commit time to the project in order to do it right.
“Organizations should not make this decision simply to avoid an overburdened IT department because the governance and process employed by IT is critical to the result,” Cousins said. “Instead, view cloud-based BI as a potential improvement to a purely on-premises model.”
Among the keys to success, he added, are thorough preparation in terms of resolving data quality issues, proper training of personnel and a good plan to be able to distribute business intelligence throughout the end user’s job experience with integration into CRM/ERP systems, rich email distribution and mobile device support.
The addition of cloud-based platform components creates added complexity for operations, Cousins said. Thus, decision makers should demand a rigorous design and testing period before go-live.
What sorts of questions should users ask of incumbent vendors if they are looking at an upgrade of an aging system? Cousins advised users to look beyond features and cost (though they remain important) and begin by asking for case studies from customers implementing a similar architecture.
“Focus on the architectural flexibility your organization may need over the coming years,” he said.
Here’s a short list of business intelligence vendors with cloud offerings:
QlikTech is one of the vendors Cousins called out as an innovator in cloud business intelligence. Donald Farmer, QlikTech’s vice president of Product Management, said many providers of cloud-based applications are adding monitoring, reporting and analysis to their services and using its QlikView product to drive the analytics of their hosted services.
“QlikView is regularly deployed on small departmental servers and managed by the department’s informal IT ‘whiz kid’ rather than a formalized IT role,” said Farmer. “But these small on-premises deployments have the advantage that they can grow easily to a full IT-managed solution.”
Vendor hype to the contrary, he added that cloud-based business intelligence is developing slowly. The reason: Business intelligence adds the most value when it integrates data from multiple, complex systems and gives users the opportunity to explore a complete view of the business. This is difficult to do on the cloud, because it remains a challenge to bring all the data together from many systems, integrate it, cleanse it, match it and conform it.
“Real BI is growing slowly on the cloud and cloud solutions are very limited,” said Farmer.
One of his pet peeves is solutions that call for the management of additional servers. He urges users to look for solutions with excellent manageability and a small footprint. When selecting products, the best approach is to follow the data.
“Think first about your data sources and assess how much pre-processing of data is needed before it is ready for analysis,” he said. “Typically this may be 70 to 80 percent of the entire effort. If this must happen on premises, ask yourself how much you will really save – either financially or operationally – hosting the final 20 percent on the cloud?”
Jaspersoft carved out a reputation as one of the early adopters of open source technology for business intelligence. Now it claims it offers the industry’s first business intelligence for platform-as-a-service (PaaS).
“We are working with all the major PaaS providers to ensure our BI platform is available within these new cloud-based development and deployment environments,” said Karl Van den Berg, vice president of Product and Alliances, Jaspersoft.
Partnerships include Red Hat, VMware’s CloudFoundry PaaS environment and others. Van den Berg calls attention to the fine points when selecting cloud business intelligence – understanding the differences between software-as-a-service (SaaS) BI, BI for SaaS and BI for PaaS.
For example, SaaS business intelligence, which he defines as hosting a business intelligence platform or application in the cloud and delivering business user functionality on demand, requires significant customization and users with requisite technical skills and a business problem that allows cloud-based consumption to be a better alternative. Business intelligence for SaaS, on the other hand, requires building BI functionality into a SaaS-based application for the purpose of delivering specific, data-driven functionality within that application.
Business intelligence for PaaS is a set of BI components, built from web-based and open standards, available as a set of reporting and analytic technologies that enable a developer to quickly assemble a modern, data-driven application and then deploy and manage it in the cloud. This approach, he said, creates greater agility, reduces infrastructure and start-up costs and delivers faster time-to-value through a better user experience.
Jaspersoft offers all three flavors.
Van den Berg advises users not to try and do too much, too soon. Think big, but start small. Starting small means beginning with a core use case or set of users that will benefit most from cloud business intelligence and then growing the deployment over time to more users and more use cases as needs evolve. Thinking big means you need to have the end goal of a broader deployment in mind. Avoid selecting a cloud business intelligence offering that is not going to scale, for example, in terms of cost or in terms of the variety of data sources (like NoSQL) that you may have to support in the future.
Drew Robb is a writer who has been writing about IT, engineering, and other topics. Originating from Scotland, he currently resides in Florida. Highly skilled in rapid prototyping innovative and reliable systems. He has been an editor and professional writer full-time for more than 20 years. He works as a freelancer at Enterprise Apps Today, CIO Insight and other IT publications. He is also an editor-in chief of an international engineering journal. He enjoys solving data problems and learning abstractions that will allow for better infrastructure.