Prompted by budgetary pressures in a tough economy and the increased complexity of traditional business intelligence...
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
(BI) deployments, manufacturers are taking a heightened interest in Software as a Service (SaaS) BI technology as a more cost-effective, lower-maintenance alternative to traditional analytics software.
SaaS -- a paid subscription delivery model in which a vendor hosts, operates and manages a software service for clients -- has become a popular option for enterprise applications such as CRM and ERP, primarily because of the cost benefits and fast-start approach to deployment.
Despite these advantages, companies have been slow to embrace SaaS BI tools, in part because there are fewer well-established vendor options to choose from and because BI as a discipline doesn't translate as naturally to the SaaS model.
"BI is a later port [to the SaaS model] because it's a riskier port," said William McKnight, president of McKnight Consulting Group, which specializes in business intelligence technology. "BI is wide open -- it means access to the right information at the right time with the right level of quality. It's hard to draw a common line around all that and say what the right solution will be for everybody."
More offerings from SaaS BI vendors
Recognizing a growing opportunity, however, established BI vendors and newer startups have flooded the market with a variety of offerings of SaaS BI for manufacturing, from off-the-shelf analytical tools and reporting capabilities to specialized, industry-specific analytical applications, in addition to full-blown BI and data management suites. SAP, for example, offers a trio of SaaS offerings for business intelligence, resulting from its acquisition of BusinessObjects. Newcomers such as PivotLink offer analytical BI tuned for a variety of industries, while companies like Kognitio serve up a full data warehouse SaaS solution.
To fully capitalize on the promise of SaaS, companies should look for BI solutions that have been rearchitected to fully exploit this new software delivery model and are not merely direct ports of traditional technology, according to Dyke Hensen, PivotLink's chief marketing officer.
SaaS BI, Hensen maintains, should adhere to the following tenets:
- It should be easy to integrate multiple data sources across an organization.
- There should be tools for aggregating and slicing and dicing data without a requirement for pre-defined modeling.
- Users should be able to self-provision the software, including creating their own reports and dashboards, without assistance from IT staff.
"Initial SaaS applications for BI were geared to the super-duper analyst or IT department, where you still had to preassemble a database and manage the performance of the application," Hensen explained. "You can't take the old way of doing it and assume a successful BI implementation. To deliver BI in a SaaS environment, you almost have to forget about the way you did it in the past."
Weighing pros and cons of SaaS BI
Despite the promised advantages of SaaS BI, there are tradeoffs and hurdles to its adoption. For small and medium-sized companies, where budgets and IT resources are scarce, SaaS BI remains a compelling alternative. Yet for large companies committed to the idea of building, maintaining and leveraging a massive data warehouse for a variety of analytics applications, SaaS BI may not be the best fit.
For one thing, despite the advances in Internet bandwidth, performance of a large-scale data warehouse in the cloud could be an issue, as could security for certain types of businesses (government, for example) and certain instances of data, according to Colin White, founder of BI Research, a consultancy specializing in data integration and data management, and co-author of the report Pay As You Go: Software-as-a-Service, Business Intelligence and Data Management.
"We've still not cracked all the issues," White said. "There are questions around how quickly you can deploy a virtual image or optimize particular workloads and [whether] the solution supports the scalability and performance you want."
Doing BI in a SaaS model can also circumvent the benefits of a centralized data warehouse architecture, since data is often supplied directly to the SaaS vendor without going through the established extraction, loading and translation processes typically associated with enterprise data warehouses, noted Claudia Imhoff, president and founder of consultancy Intelligent Solutions and co-author of the "Pay As You Go" report.
"Part of the beauty of business intelligence is consolidating data in a data warehouse and reusing it over and over again for various applications," Imhoff explained. "When you do SaaS, you fracture that architecture … and you can lose that consistency and reliability you get from a data warehouse." To avoid these issues, she recommends running data through the data warehouse before shipping it off to a SaaS vendor, or creating the appropriate workflows, using service-oriented architecture (SOA), to directly integrate the two processes.
Companies evaluating SaaS BI also need to consider security -- the top SaaS BI requirement cited by 90% of respondents to the "Pay As You Go" survey -- as well as ease of maintenance and customization requirements. While BI applications have traditionally been heavily customized, SaaS BI relies on line-of-business users to do any customization. This means that organizations implementing SaaS BI retain flexibility for modifying things like formats and field names, but they will have to give up some major application customization, Imhoff said.
In addition, understanding the roadmap of your SaaS vendor will ensure that you align with a vendor that best matches your evolving business requirements, as will looking for vendors that meet security certifications like SAS 70 Compliance.
Preparing for SaaS BI market changes
Finally, companies evaluating SaaS BI need to consider the viability of their potential vendor, since many of these companies are small and highly specialized and so face a greater risk of going out of business. For example LucidERA, an upstart in SaaS BI, closed its doors in June, and competitors PivotLink and Birst, which specialize in on-demand sales, financial and marketing analytics, both sprang into action with programs designed to switch LucidERA customers to their respective SaaS BI platforms.
Experts like Imhoff and White expect SaaS BI to stick around for the duration, even with such hiccups, and to become a viable way to jumpstart BI deployments. Said Imhoff: "Any time the economy is tight, companies are looking for alternatives to satisfy their needs."
About the author: Beth Stackpole is a freelance writer who has covered manufacturing techniques and manufacturing technology extensively.