Question and Answer with the Experts
A business intelligence or data warehouse implementation can be a formidable undertaking. Leading business intelligence and data warehousing solution providers share their answers to the questions they hear often from industry professionals. Mark Hammond, anindependent consultant, provides his analyst viewpoint for each Q&A.
A business intelligence or data warehouse implementation can be aformidable undertaking. In these pages, leading business intelligence anddata warehousing solution providers share their answers to the questionsthey hear often from industry professionals. Mark Hammond, anindependent consultant, provides his analyst viewpoint for each Q&A.
Most of the big BI providershave been bought by eitherERP/CRM vendors or platformvendors. What impact does thishave on the BI landscape?
The most obvious impact is that anyonewho has aligned strategically with theseproviders and is considering BI solutionsshould first look at the offering of thatvendor. That does not mean the choiceis a given, but now it must be seriouslyconsidered. The potential synergy is toocompelling to overlook. However, evenif the software is virtually free, there willstill be a substantial adoption cost, andin most environments there will still bemany applications outside of the platformto consider. Now, however, given therobust nature of the BI offerings, applicationshave strong viability beyond usewith the platform solution.
BI market consolidation has triggered plentyof hand wringing among customers. On thepositive side, mega-vendors are positioningcomprehensive BI stacks that can reducethe difficulties of dealing with multiple vendors.Nevertheless, questions over licensingcosts, product overlap, support quality, andcontinued product development are allissues that organizations are being forced toexamine in plotting out the next five yearsof BI strategy—especially organizationsthat, for instance, may run Business Objects(acquired by SAP) in a non-SAP environment.In the overall landscape, look for consolidationto create a vacuum of sorts that startupsand open source BI providers seek to fill.
What is the benefit ofMicrosoft’s acquisition ofDATAllegro?
With this acquisition, the SQL Serverproduct line will span data warehousesof all sizes. Microsoft has made significantinvestments into data warehousescalability in SQL Server 2005 andSQL Server 2008. The acquisition ofDATAllegro extends that and allows SQLServer data warehouses to scale out topetabytes. Customers will benefit throughcontinued innovation, product integration,and low total cost of ownership.
In the market at large, Microsoft’sDATAllegro acquisition adds momentum tothe trends towards shared-nothing, massivelyparallel architectures that can support100s of terabytes and intense concurrentuser activity. The acquisition gives Microsofta card to play in the high-end DW market,which it has been effectively shut out of withSMP-based SQL Server systems not wellsuited for high-end scalability. And it’s fuelingspeculation that Oracle, SAP, and others maybe on the prowl for DATAllegro competitors.Though it will take Microsoft time to transitionfrom DATAllegro’s open source Ingres/Linuxarchitecture to the SQL Server platform, organizationscan expect to enjoy another optionfor high-end data warehousing.
What’s the biggest barrier todeploying BI to the extendedenterprise?
From our experience, the single biggestissue for this audience is usability. Yourchallenge is that you will likely not havedirect authority to enforce best practicesor ongoing education. To make BI accessibleto casual users inside and outsideof your organization, you need a robuststrategy to address two key obstacles: 1)Users won’t readily learn (or remember)how to use overly complex tools; and 2)users won’t understand the structure andnomenclature of the data. For truly pervasiveBI, it just has to be easy.
Multiple challenges confront any organizationlooking to deploy BI to the extendedenterprise. Ownership of BI systems andstrategies needs to be addressed decisivelyif a BI solution is to transcend organizationalor geographic boundaries. Collaborationamong business and IT department headsis essential to achieve buy-in and eliminateconfusion. For IT, sound architecturalstrategy, open systems, and finesse arerequired to effectively transition from siloedinfrastructures towards an integrated enterprisesolution. User engagement, training,and feedback are critical to meeting userrequirements and broadening adoption.
What specific tools andapplications connect users tothe three levels of BI—strategic,analytical, and operational?
Strategic BI applications include strategymaps, scorecards, and dashboards thatpresent and track critical metrics suchas customer satisfaction scores, marketshare, profit margins, or overhead costs.By closely monitoring those factors, companiescan immediately detect problemsand take corrective action. Analytic BIapplications such as OLAP tools, predictiveanalytics, and ad hoc queries helpdetermine the cause of a major problem.For example, if profits are declining,is it because of low sales or increasingexpenses? Operational BI applicationsmonitor business activities as they areexecuted (such as entering orders, restockinginventory, fielding customer inquiries),enabling rank and file workers to managethese activities or respond to issues.
As BI implementations mature, organizationsare looking to build on connecting users tostrategic, analytic, and operational BI systemsby connecting the systems themselves.The three levels, though distinct, should notbe viewed as silos but as focus areas thatcan complement each other with mix-andmatchfunctionality. For instance, substantialvalue can be derived by equipping an executivedashboard (the strategic level) with adhoc query and predictive analytics (the analyticlevel) to enable executives to conductroot cause analysis and forecast trends. Useof open standards and a service-orientedarchitecture can supply a common platformfor pervasive and richly functional BI.
How do you think consolidationin the BI industry will affectthe market?
With Oracle’s acquisition of HyperionSolutions last year, customers no longermust choose between a vendor offering“best of breed” products or an enterpriseapplication vendor offering more integratedsolutions. Today, customers canhave the best of both worlds. Our customerssee tremendous value in having theability to partner with a financially strongvendor who offers market-leading, hotpluggableBI solutions that can integratewith enterprise applications, middleware,and database technologies. Other majorenterprise software vendors are nowattempting to follow Oracle’s strategy.We can expect to see an increasing trendamong buyers toward consolidating theBI technologies used within their companiesand aligning BI investments withtheir choice of strategic vendors.
Despite customer questions and concernsover BI market consolidation, it’s unlikelyto affect the continued brisk growth in theoverall BI market. Gartner, for instance,forecasts an 8.1 percent compound annualgrowth rate in the BI platform market, reaching$7.7 billion in 2012, as BI remains a toppriority across industries and organizationsprogress towards enterprise-level maturity.The question is with which vendorscustomers will be spending that money. Consolidationis forcing customers to diligentlysize up their BI options from a strategicperspective and ask hard questions (amidconsiderable hype and FUD) over how tobest capitalize on the business value of data.
What are the benefits of anopen systems BI solution vs.a closed systems solution?
In today’s BI marketplace, companiesmust choose between open systems solutionsfrom independent BI providers andclosed system, all-in-one-stack solutionsfrom conglomerate vendors. With opensystems BI, companies are not lockedinto a single vendor and can freely assembletechnologies best suited for theirorganization. Freedom of choice gives theBI purchasers greater leverage for betterpricing and the flexibility to changesoftware components as needs change.Additionally, open systems vendors aresolely focused on BI and meeting customers’requirements through ongoingtechnical innovation.
Viewed another way, the issue facing organizationsin the market for BI technology iswhether to opt for solutions from independentpure-play BI vendors or mega-vendorBI providers. The spin from both campsis naturally quite different. IndependentBI vendors tout a singular BI focus andgreater interoperability in heterogeneousenvironments. Mega-vendors promote anend-to-end platform and the manageabilitybenefits of dealing with a single vendor. Inthe wake of industry consolidation andthe diminished population of pure-play BIvendors, customers are carefully weighingthese and other factors in setting directionfor vendor and product selection from bothstrategic and tactical perspectives.
Are your customer programsleaving money on the table?
Marketers today are challenged withbreaking through the clutter and gettingthe attention of customers. No longer aretraditional outbound campaigns sufficient—marketers must be able to evolvetheir marketing to engage with customerson the “inbound” or when they choose tocontact the organization. To make theseunpredictable interactions more predictable,marketers need to equip themselveswith the ability to strategically map deliveryof the right message at the right timeacross all marketing and service channels.Portrait’s integrated solutions leverageinnovative yet easy-to-use analytics toconduct two-way intelligent dialog withindividual customers and build strongerand more profitable customer relationshipswith them while maximizing ROI.
In CRM, there’s the traditional integrationchallenge and the new Web 2.0 challenge.The traditional challenge is to tightly integratecustomer data across Web, email,store, and other channels. After substantialinvestment in CRM, organizations are turningattention to such granular technologicalchallenges as creating a single view by reconcilingdifferent email addresses, names,orders and other dimension that a singlecustomer may have logged across multiplechannels. The new Web 2.0 challengeis building Web-based communities thatenable customer feedback and collaborationin forums, comment fields, wikis and othersocial networking tools. Web 2.0 is swiftlybecoming de rigueur for marketers to buildcustomer loyalty.
As data volumes continue togrow, uploading data to the datawarehouse has become muchtoo time-consuming. Is there away to speed this process?
The process of uploading data can besimplified by using a data managementsolution with changed data capture(CDC). CDC, or delta processing, isa technique in which an older file anda new file are compared to identifychanges that have occurred. Only thedata that has changed is updated in themaster file. This type of processing willhelp reduce the amount of data that willneed to be uploaded to a data warehouse,improving performance.
Though it’s been around a while, changeddata capture (CDC) has found its sweetspot with the advent of the real-time datawarehouse. Available as a built-in or optionalcomponent in leading ETL tools, CDC candrastically reduce the volume of data to bemoved into a warehouse (or to an operationalapplication or other system, for thatmatter). Naturally, if data loads are reducedby, e.g., 85 percent, load times are reducedcommensurately, making multiple intradayloads practical. The additional benefits ofminimizing impact on operational applications,bandwidth, and I/O make CDC wortha close look by any organization that wantsto freshen its warehouse data to enhance itsvalue to the business.
Open source data integration ismaking the headlines, but is itready for the enterprise?
Beyond the technical aspects (opensource data integration offers today thesame level of features and performanceas proprietary vendors), the true answerlies in the ability of commercial opensource vendors to meet the demandsof enterprise customers. These vendorscontrol the road map of their products,provide enterprise-grade technical supportand IP indemnification, and reducethe dependency of their customers on asingle mega-vendor that controls theirentire IT stack and budget. Open sourceis today a true alternative to proprietarydata integration and many companiesof all sizes are proving it every day bymaking the switch.
Open source data integration has been ona fast track in 2008. For instance, Talendin June introduced change data capturetechnology into its flagship product, enablingacceleration of processing times by movingonly data changed since the last load, animportant feature for enterprises with largedata sets and time sensitivity. While thesteady maturation of enterprise features inopen source data integration is promptinginterest, change won’t happen overnight.History shows that open source initiallytends to be adopted tactically and incrementallybefore moving into the mainstream, andsigns point to continued growth in the opensource data integration market in the nextseveral years.
Back to Table of Contents