BI in Action

Sponsored by:

Become a BI in Action Sponsor June 20 & 21 2007!

• Agenda
• Sponsors
• Partners
• Blog
• About
• Contact Us

BI in Action Blog

« November 2007 | Main | January 2008 »

December 31, 2007 BI, Delivered from the Cloud A new report in Knowledge Management confirms a trend that began emerging over the past year -- business analytics delivered via Software as a Service. "More and more business analytics software providers are moving to address increasing market demand for software that is updated frequently, hosted off-site and purchased on a subscription basis," KMWorld notes. Factors propelling this trend include corporate budget constraints on traditional software, limited development resources for in-house BI software, and increasing maturity of BI-over-the-wire solutions. The report notes that recent market consolidation -- such as the Oracle-Hyperion, SAP-Business Objects, and IBM-Cognos acquisitions -- is spurring the proliferation of business analytics solutions, which will grow the market, and thus, the opportunities for SaaS-delivered services (which could come from these large vendors as well.) The authors note that they don't expect the whole BI world to move to the Cloud anytime soon, however. As they put it, the shift will be gradual, and both SaaS and traditional, on-premises, licensed software will co-exist within enterprises for a long time to come. A few months back, while preparing a report for Database Trends & Applications, I spoke with Tracy Trawick, consumer insights manager for Hamilton Beach, which employs such a hybrid approach to BI, getting the best of both worlds of SaaS and on-site. Hamilton Beach's market research department primarily relies on on-site software, but lately has begun tapping into SPSS's SaaS-based platform to cover a growing workload that end-users simply don't have time to sort through. "All the data manipulation on the back end is supported by SPSS," Tracy said. "We're using it for very template-based iterative projects so we don't use up a lot of programming time. I don't have the time to learn the logic and the programming tools." Tracy uses the SPSS SaaS platform to create and launch consumer market research surveys, and manipulate the data as it comes in. However, she still employs SPSS on-site software for more complicated projects. In line with the Hamilton-Beach example, I also spoke with Oliver Halter, partner with Diamond Management & Technology Consultants, who concurred that "companies currently tend to use SaaS on the fringe of their operations." Oftentimes, "this also means that the SaaS partner provides an additional service that makes it worthwhile for the company to outsource the software and function. Companies tend to have an easier time switching to SaaS if the data managed in these systems is 'non core' and is not considered a competitive advantage." Moving to a SaaS model has its share of challenges, however. Dilip Wagle, a partner with McKinsey & Company, told me that integration remains a big challenge, especially for companies seeking to leverage solutions from the Cloud. Dilip cautioned, for example, that "counter to the promise, the deployment and integration effort is not zero. Most effort is centered on integration and customization effort of single SaaS applications, especially in the enterprise space. Moreover, as companies increasingly move to more SaaS applications, integration challenges will manifest themselves in terms of integration with other SaaS applications." Data security and availability is also another challenge, and perhaps the achilles' heel of SaaS arrangements. What happens if the application provider goes out of business, or has a major network outage? SaaS customers need to make sure they have robust backup systems in place. Security raises more red flags. As Dilip warned, "pure 'in-the-cloud' services imply that all the customers data are essentially stored off-premise in a data center owned or contracted by the service provider. In the event of data center failure on the part of the service vendor, the customer has no recourse but to hope that data were appropriately backed up and managed in a secure fashion. The problem can be exacerbated if the SaaS vendor in turn, outsources back-end data center operations to yet another third party. This can complicate accountability and liability in the event of failure or security breaches." Posted by joemckendrick in BI Vendor Watch • Business Intelligence  |  Permalink  | Comments (0)  | TrackBacks (0) December 30, 2007 Rumors of BI's 'Death' are Greatly Exaggerated In a 2007 year-end wrap-up, InfoWorld declared BI to be "dead," citing it as it's ninth-most underreported story of 2007. Hmmm. An entire industry and technology sector disappears under our noses, and it only rates number nine as a news item? Actually, it appears the author, Bill Snyder, was being somewhat facetious, mocking the pundits that predicted the end of the market as we know it, resulting from the mega-acquisitions of leading vendors that took place over the past year -- Business Objects being gobbled up by SAP, Hyperion by Oracle, and Cognos now falling into IBM's orbit. But, keep in mind that there is still a huge market remaining with players such as SAS, SPSS, MicroStrategy, and Actuate. Oh, and what's the name of that vendor out in Washington state again? I think they have something or other linked to their database product. As Rob Tholemeier, a former industry analyst turned private investor, put it, the acquisitions of the big players creates space for other companies to flourish. The same thing happened to the database market, he notes: “There are more database companies around now than when Informix was purchased.” There's going to be plenty of competition and solutions for a long time in this market, with enough room for everyone. Long live BI. Posted by joemckendrick in BI Vendor Watch • Business Intelligence • Microsoft  |  Permalink  | Comments (1)  | TrackBacks (0) Enter Data Warehousing 2.0 Enterprise data warehousing is supposed to reduce the incidents of stovepiped BI approaches seen in spreadsheets and spreadmarts. But many EDW sites may get so complex and far from its original design that end-users end up running back to spreadsheets, warns Dan Linstedt. Dan recently pointed to the opportunities and challenges with the next generation of data warehousing, which he refers to as "Enterprise Data Warehousing 2.0." (Just as there's now a trend identified as "BI 2.0".) The problem with enterprise data warehouses as they have existed is that they have been expensive to build, usually requiring expensive consultants. The issues aren't necessarily seen in the first iteration, but later on as more business units want to get on board. As Dan explains it, the company selects "star-schema" modeling as the way to build its EDW. "Then, they select conformed dimensions, and shared fact tables. The first implementation costs the business 90 days and maybe 5 consultants, and maybe $250,000 USD. If your lucky, it might be $150,000." So far, so good, Dan says. "The business unit that this is built for becomes very happy, with quick delivery, apparently low cost, and super fast access to dimensional information that meets their business needs... But then, reality sets in... Other business units see this success, and want 'one of their own' built." No biggie. "Building a second or even a third star schema and then federating these together doesn't seem to be such a big deal," Dan says. "The cost may increase only slightly to maybe $180k or $275k, and the number of days to implement may increase only slightly to maybe 110-120 days." However, by its fifth or sixth iteration, things are getting too complex, and the original designs of the EDW become lost or distorted, Dan explains. "IT (because of business needs) takes existing dimensions and begins to add different and loosely affiliated information to the same 'dimension,' thus, apparently attempting to 'conform' it." As this process continues, and IT gets in to the 5th or 6th project, "the conformity of the dimensions becomes lost in the fray," he says. "Too many different kinds of data are added to the dimension 'to conform it to the enterprise.' which distorts it's original purpose." In fact, if done improperly, "each time IT increases the size of this monster, it always creeps in to higher cost, and longer implementation timeframes." Agility is lost, and "a simple 'change' that the business has to make (that used to cost $150k and take 90 days) now costs well in to the $350k range and takes six months or more. What was a conformed dimension now becomes a "deformed" dimension, and has trouble meeting the business needs." As a result, the business users do their own workarounds of this clumsy enterprise beast -- which means going back to relying on spreadsheets. DW 2.0, Dan relates, "comes with the standard definitions that the industry has lacked over the years, finally and at last we have standards, definitions, and frameworks to follow." Plus, an essential piece of DW 2.0, Dan believes, is putting the right data model in place. Dan is also the designer and advocate of the "Data Vault" modeling approach, also known as "Common Foundational Integrated Data Model Architecture." (Data Vault sounds better). The Data Vault modeling architecture is a "hybrid architecture consisting of the best of breed data modeling techniques used in both third normal form, and Star Schema - except it is a foundationally based architecture with standards, which if adhered to can steer your enterprise common data model in the right direction." The Data Vault model has been under development since 1990, and has been available for free since 2000. Data Warehouse guru Bill Inmon said that “The Data Vault is the optimal choice for modeling the EDW in the DW 2.0 framework." Dan provides a technical overview of Data Vault here. Posted by joemckendrick in Data Management  |  Permalink  | Comments (0)  | TrackBacks (0) December 11, 2007 Top 10 BI Trends to Watch - 2008 I just published my list of the "Top 10 Data Management Trends to Watch" for the upcoming year in Database Trends, and thought I would share them with you here. 1 - More industry consolida­tion, acquisitions. The year 2007 saw a number of acquisi­tions and mergers among prominent vendors, including IBM’s announced acquisition of Cognos, SAP’s acquisition of Business Objects and Oracle’s acquisition of Hyperion. Expect more of the same in 2008. 2 - Finally, BI with enterprise reach will become more of a reality. “Democratic” BI - accessible to executives and employees throughout the organization, rather than being limited to analysts - has long been an unfulfilled pipe dream within the industry. However, in 2008, this vision may finally begin to be realized. 3 - Business intelligence moves toward performance manage­ment and complex analytics. As BI moves out to the mass­es, there will be increased scrutiny and leveraging of the metrics that drive the business itself. 4 - Business intelligence and analytics will move to real time. In 2007, the industry witnessed a dramatic shift away from historical data analysis to the embracing of real-time analytics. New initiatives such as SOA are also facilitating this movement to real-time BI. 5 - SOA and integration will grow as enterprise IT activities. The need for speed of delivery has put SOA front and center on everyone's mind. 6 - Mashups and Web 2.0 will rise within enterprise walls. Due to the rapid pro­liferation of consumer-facing mashup applications, the demand for customized, collaborative-oriented applications is making its way into the enterprise. 7 - More companies will embrace soft­ware- as-a-service (SaaS). Many observers consider the SaaS phenome­non as a vital piece of Web 2.0. 8 - Managing more with less, and align­ing more with the business. Managing more with less is a trend that never goes out of style, and this will especially be the case in 2008. 9 - More server and storage virtualiza­tion. Virtualization is hot for a number of reasons, and extends across a number of parts of the enterprise. Virtualization will also be increasingly used to manage the storage piece of the enterprise. 10 - Data and IT governance will become urgent requirements. Data governance has typically been limited to policies and enforcement, but the emphasis will broaden over the coming year, shaping IT infrastructures. Posted by joemckendrick in Data Management  |  Permalink  | Comments (0)  | TrackBacks (0) Best Practices for Turning Useless Data into Actionable Information ebizQ colleague James Taylor points to a new report out of Forrester (available to ebizQ members for premium download here) that discusses Business Intelligence strategies organizations need to follow in the years ahead. In the report, Forrester analyst Boris Evelson not only provides guidance on BI strategy, but also examines the meaning behind BI. James quotes Evelson's definition of BI: "A set of processes and technologies that transform raw, meaningless data into useful and actionable information." BI has certainly had its issues, as Evelson explains. "Even after decades of BI usage, most organizations find best practices elusive," he observes. "The crowded market segment of BI often gets mislabeled, can be very complex, and constantly changes. The BI market pulses with alternative approaches, contradictory technologies, and hundreds of vendors that constantly move to consolidate, innovate, and leapfrog one another. Enterprises at the beginning of a perilous BI journey need guidance and help untangling the web of BI complexities." BI doesn't sound like much fun, does it? To smooth things a bit, or at least relieve such angst, Evelson compiled a list of best practices in which to get BI started or kick-started within organizations, some of which are highlighted below. His advice overall: "Don't treat BI strategies as a project -- treat BI as a journey." Pick a senior executive sponsor who understands that measurement is management: "Look for an executive who has the entire picture of the enterprise objectives, goals, and strategies, and who has the know-how to link and translate them into the key performance indicators (KPIs) necessary to support these objectives." Put data governance and data stewardship in place: No effort will succeed without common data definitions across all business units. And, as noted above, measurement is vital, Evelson says. "Start small and pick the KPIs that are essential to operate the business. Pick only 10 to 20 KPIs and build the initial standards and governance with them in mind. Come up with standardized definitions of quality, sources of truth for the data, and business rules to calculate these KPIs/KPMs (key performance measures)." Conduct a "current state analysis:" Make sure that this analysis "includes all components of the BI stack and all of the processes and organizational structures surrounding your existing BI implementations. Both business and IT stakeholders have to be involved in this initiative." Define a logical and physical data architecture: Decide whether you will be deploying data marts, a data warehouse, or "semantic layers" (or EII, enterprise information integration architecture) that accesses operational data sources. Know they user: This sounds simple enough, but most organizations have a wide variety of user types that will have different expectations of the BI system. Evelson points out that there are typically three broad classes of users: strategic, tactical, and operational. For example, strategic users make very few decisions, but the decisions these users make have profound consequences for the business." These users "favor aggregated data that extends across business domains — they ask the big questions and need the big picture to do so, too." Operational users, on the other hand, "work on the frontlines — for example, the hundreds or perhaps thousands of representatives in call centers. These users don’t need cross-domain data, just data within their own set of applications. They don’t need the big picture, just a very granular view of many small pictures." Evelson also adds that most BI applications and data will require at least five terabytes' worth of storage space over the next five years. Lastly, addressing the biggest stick in many BI pundits' craws, Evelson advises not trying to fight BI-by-spreadsheet. "Excel’s local control, flexibility, and complexity are perfect matches for data-intensive analysts who like to run every scenario with constantly changing data. No BI tool is going to replace Excel entirely; the trick is a coexistence strategy between IT and the business." Or, as my colleague Michael Dortch (of RFG) so aptly put it a few months back: "Spreadsheets: can't work with 'em; can't work without 'em!" Posted by joemckendrick in Business Intelligence  |  Permalink  | Comments (0)  | TrackBacks (0)