Ride the Next Wave of Flexible 'BI Workspaces'Ride the Next Wave of Flexible 'BI Workspaces'

Much has been written in recent years about the need to spread BI to the masses — embedding intelligence within applications, simplifying interfaces, improving access to reports and so on. A new report from Forrester Research suggests that the next wave of productivity lies in providing flexible "BI Workspaces" to the oranization's few analysts and power users. Overcoming the classic tech struggle between autonomy and centralized control, BI Workspaces are said to offer a best-of-both-worlds environment in which experts can analyze and model vetted data without facing IT constraints.

Just what are "BI Workspaces"? The Forrester Report, "BI Workspaces: BI Without Borders," defines them as "a data exploration environment where a power user can analyze production, clean data with near complete freedom to modify data models, enrich data sets, and run the analysis whenever necessary, without much dependency on IT and production environment restrictions." The primary examples of these environments are multidimensional OLAP (MOLAP) slice-and-dice analytics, BI and data warehousing delivered Software as a Service (SaaS) style, and in-memory analytics platforms. This article explains why report author Boris Evelson writes that one or more of these capabilities must now be on the list for any leading-edge BI environment.

Why Your Arms Are Tied

The quest for the proverbial "single version of the truth" and various compliance requirements have in many ways restricted the flexibility of reporting and analysis, according to Forrester. As an example, corporate security, data privacy and regulatory requirements are often an obstacle to open and timely access to information. What's more, despite ever-faster-and-more-powerful processing capacities, analyst and power-user demands still take a back seat to day-to-day transactional loads. Perhaps most restrictive is the architecture of many BI environments, which creates a "cascading effect of interdependent components" (see chart at right).

Cascading Effects of Independent BI Components
(click image for larger view)

"Most BI environment architectures use several components (sometimes more than 30!), such as data discovery, transformation and integration, data modeling, online analytical processing (OLAP), reporting, dashboards, and alerts, and involve as many steps for raw data to reach its final destination in a report or a dashboard," writes Evelson. "One simple change to even a single source data element may result in a few changes to extract, transform, load (ETL) and data cleansing jobs, which may turn into several data model changes in operational data store (ODS), data warehouse (DW), and data marts; this in turn affects dozens of metrics and measures that could be referenced in hundreds of queries, reports, and dashboards."

All too often, power users, analysts and business users alike have attempted to work around these restrictions by resorting to spreadsheets, but this approach no longer works. "Aside from the lack of controls and huge operational risk, spreadsheets are just not powerful enough to analyze terabytes of data involved in any modern, competitive decision-making," Evelson writes.

Three Varieties of BI Workplaces

The beauty of BI Workspaces is that they let power users work with controlled production data without having to wait for IT and without having to port said data into a spreadsheet doomed to be disconnected if not downright dangerous. Here's how Forrester sizes up three leading approaches to providing BI Workspaces:

MOLAP. Newer than relational online analytical processing (ROLAP) and hybrid OLAP (HOLAP - combining MOLAP and ROLAP), MOLAP offers desktop-based slice-and-dice analytics. With leading examples including IBM Cognos PowerPlay and Oracle (Hyperion) Essbase, MOLAP lets analysts "create PC-based cubes using any data model that represents the best fit for the latest business requirements," writes Evelson. Thus, an analyst can use MOLAP to explore what-if scenarios such as deferring capital spending and halting discretionary spending in the face of an economic downturn and reduced sales.

BI SaaS and DW SaaS. The downside of MOLAP is that it can't always scale, so many companies are turning to SaaS-based BI and data warehousing to quickly gain access to flexible browser-based reporting and analytical tools. SaaS-based BI services such as InetSoft and Panorama analytics for Google Apps are a good fit for "smaller enterprises [or] departmental use cases where data sets are relatively small," according to Forrester. SaaS-based DW, in contrast, is geared to multigigabyte and terabyte data sets. The report cites the following example related to the current U.S. subprime credit crisis: "A number of financial services firms are loading large amounts of loan-level mortgage data into DW SaaS from providers such as 1010data [and Vertica] so they can analyze the patterns of historical prepayment, default, delinquency, and loss severity rates."

In-Memory Analytics. MOLAP and SaaS-based BI offer flexibility, but they still require data modeling steps. In contrast, in-memory technologies let you perform calculations and aggregations at RAM speeds with little data preparation. Thus, "in-memory models from vendors like InetSoft, QlikTech, and TIBCO Spotfire do not require a distinction between facts and dimensions — any element can be instantaneously used in either capacity," writes Evelson.

Recommendations

Based on the findings of the report, "BI Workspaces: BI Without Borders" offered four recommendations:

Include BI workspace functionality in overall BI requirements, but make sure end users have adequate training, use case examples and documentation.

Support BI workspace users by providing clean and timely data, and make sure users understand the implications of not using clean, secure or nonstandardized data sets.

Understand the implications of using a different technology/approach for BI workspaces. Some vendors provide both traditional BI and BI workspace applications using different technologies, so make sure you understand the risks and costs of supporting both environments.

Consider risks associated with SaaS-based BI/DW. "Due diligence is key," says Forrester. "For example, ask whether the SaaS vendor ensures data security by providing you with the results of its SAS 70 Type II Audit, and whether your hosted data is backed up in case of equipment failure or for disaster recovery purposes. "

The complete "BI Workspaces: BI Without Borders" report, which includes a matrix of Workspace approach definitions, strengths, cautions and vendors, as well as supporting case examples, is available as a free download through this link (registration required).