or Build Buy: Maximizing the Value of Performance Management Applications
A U T H O R S : Mark Hammond, Guy Weismantel C O N T R I B U T O R S : Paul Clark, Davythe Dicochea, Elaine Goldstone, Jennifer Meegan A U D I E N C E : IT managers considering a performance management application project
Introduction Section Heading
Several years ago, an organization that wanted to provide business users with a rich, metrics-driven analytic environment in a personalized dashboard had little choice but to roll up its sleeves and start building. Enterprising developers could hard-code a performance management application from scratch, or customize a generic query and reporting tool with metrics and functionality specific to such business areas as sales, marketing, supply chain, or finance. As many organizations discovered, this approach left something to be desired. More than a few saw these ambitious projects come in over budget, behind schedule, and lacking the ease of use and analytic reach they promised to deliver to a data-hungry business audience. But when an IT department attempts to build the plane while its being flown, its not unusual to suffer such missteps. Times have changed. For one thing, economic pressures have made it more imperative than ever to give business users the high-level metrics and deep analytic insight they need to optimize performance, react swiftly to changing conditions, and make informed, proactive decisions. And organizations now have a compelling alternative to the laborintensive process of building performance management applications. Over the past couple of years, business intelligence (BI) vendors have brought to market packaged applications that deliver key metrics, ad hoc query, powerful analytic engines, best practices, and an intuitive user interfaceeliminating the need for the customers to build these products from scratch. As well explore in this paper, organizations are increasingly abandoning the old customcoding approach to analytics and are taking advantage of packaged performance management applications to realize: Cost reductions of 15-25% Faster time to deployment Greater return on investment (ROI) Greater extensibility and less maintenance A strategic platform for integrated enterprise analytics
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
1
Performance Management Applications: Engineered for Business Insight
Performance management applications build on the generic reporting, and query and analysis capabilities of conventional BI tools. With conventional BI, IT staff and business users typically collaborate to generate reports, devise ad hoc queries, and customize the BI tool to a particular business function. Packaged applications provide the same ad hoc query and reporting, but go a step further by providing specialized, prebuilt reports and best-practice metrics specific to business functions. Over several years, they have matured to offer organizations a potent means of achieving rapid and precise insights into enterprise data. Moreover, they serve as strategic building blocks with which an organization can incrementally develop an analytic infrastructure that spans multiple business units and processes. Packaged applications typically rely on a common data model that enables them to snap together for cross-functional analysis that can enhance performance across the enterprise. This framework serves to synchronize among interdependent processes, so that sales data analysis contributes to optimal manufacturing, inventory, and logistics, or product marketing is fine-tuned for different geographic locations based on sales by region, demographics, and profitability. As business analytics continue to mature, many organizations have in their sights this type of next-generation business intelligence infrastructure. Key characteristics of a packaged performance management application include: Specific business functions: Tailored for customers, marketing, products/service, supply chain, finance, human resources Powerful analytic engines: Supports dynamic data segmentation, metric and key performance indicator (KPI) tracking, rules-based alerting, predictive analysis, and statistical process control Metrics and KPIs: Delivers dozens of prebuilt metrics in visually inviting, personalized dashboards Guided analysis: Helps business users explore information to identify root causes and take corrective action Best practices: Reflects industry, domain, and customer experience and best practices developed over the years Alerting: Provides customizable thresholds and exception reporting for timely notification of business anomalies
2
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
P E R F O R M A N C E M A N A G E M E N T A P P L I CAT I O N S : E N G I N E E R E D F O R B U S I N E S S I N S I G H T
High Visibility to Get from Data to Decision
Performance management applications focus on business context and business relevance. They provide an environment rich with metrics, reports, queries, and terminology familiar to users in a specific business unit. For instance, a performance management application used by a customer relationship management (CRM) specialist will feature data points and analytic roadmaps on sales by customer, sales by region, time of year, churn rate, customer profitability, and so on. By offering fast and easy visibility into data, performance management applications help decision makers zero in on metrics that matter, identify and correct problem areas, and capitalize on opportunities. The structured analytic environment provides a common framework for systematic analysis and performance improvement. They drive the generic capabilities of query and reporting to a next level. Performance management applications are not exactly new. For years, IT departments have customized BI tools with functionality sought by business users, and have handcoded these applications from scratch. Packaged performance management applications offer an attractive alternative to labor-intensive development, usually with functionality that an IT department would be hard-pressed to match.
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
3
The Evolution Towards Packaged Performance Management Applications
The maturation of performance management applications parallels that of operational applications in the 1990s. For years, companies had built homegrown transactional systems to support finance, supply chain, sales, and other processes. In time, many abandoned custom development in favor of packaged operational applications from SAP, Oracle, PeopleSoft, and others. Of course, companies that adopted these software packages tailored the packaged application to suit their needs. That same trend is playing out in the performance management applications arena. As of late 2002, nearly two-thirds of U.S. organizations were building these applications rather than buying them, according to a survey of 578 BI professionals by The Data Warehousing Institute.1 But as TDWI noted, that is rapidly changing. Sixty-two percent of TDWIs respondents reported they would be interested or very interested in deploying a packaged analytic solution in 2004, up markedly from 47% in 2002. For its part, Business Objects has enjoyed rapid uptake of Sixty-two percent of TDWIs its suite of performance management applications respondents reported they would be focused on customers, products, and services, products interested or very interested in and services, supply chain, human resources, and finance. deploying a packaged analytic Additionally, Business Objects recently launched seven solution in 2004, up markedly new analytic solutions, focused on distinct business issues from 47% in 2002. with a vertical focus on retailers and CPG manufacturers, and two finance related solutions, focused on cash management and Sarbanes-Oxley. More than 150 organizations worldwide have licensed these product since 2002-03, including GE Power Systems, Unisys, Ford Motor Co., Ingram Micro, Sysco Foods, Qwest Communications, and Nieman Marcus. Our success mirrors forecasts by IT analyst firm IDC that the worldwide analytic applications market will grow from $2.5 billion to more than $6 billion in 2005. A principal reason why organizations opted to build analytics is because many of the packaged products lacked functionality that business users sought. In TDWIs survey, missing functionality was the top reason for not buying a packaged application, cited by 52% of respondents. That is changing, however, as vendors continue to improve the products. Many organizations recognize that the packaged performance management applications market is just a few years old, and that applications will grow more robust and refined over time.
1 The Rise of Analytic Applications: Build or Buy?The Data Warehousing Institute, October 2002.
4
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
T H E E VO L U T I O N TOWA R D S PAC K AG E D P E R F O R M A N C E M A N AG E M E N T A P P L I CAT I O N S
Business Objects believes that five key factors will drive the evolution from build to buy: Faster time to deployment Lower costs, greater return on investment (ROI) Embedded industry best practices Superior business usability and analytic depth Extensibility across the enterprise Lets take a look at these factors, and how they influence decisions on whether to build or buy performance management applications in your organization.
Faster Time to Deployment
Customer experience in multiple industries and research by independent experts has shown that organizations can more rapidly implement a packaged performance management application than build their ownoften by a significant margin. A packaged performance management application will typically have as much as 60% of the functionality that an organization desires built in. In contrast, customizing a generic BI tool or building an analytic tool from scratch poses a significant development challenge to an IT staff, and diverts both IT and business personnel from other tasks. Designing and developing a custom analytic project takes the average organization 16 weeks, opposed to three weeks to refine a packaged application to specific requirements, according to estimates by Giga Information Group, a Forrester Research subsidiary.2 Companies put a premium Eighty-one percent of respondents to on rapid deployment. Eighty-one percent of respondents TDWIs survey cited faster time to to TDWIs survey cited faster time to deployment as the deployment as the greatest advantage greatest advantage offered by a packaged application, offered by a packaged application, followed by minimizing use of IT resources (72%) and followed by minimizing use of IT providing needed functionality (70%). resources (72%) and providing needed functionality (70%). Nearly half (45%) of the TDWI survey pool reported deploying packaged performance management applications in six months or less; for custom-built applications, the time to deployment averaged nine months. Custom-building a performance management application involves gathering business user requirements, building in the desired functionality, and revising it based on user feedback. The back-and-forth between IT and business is often a healthy exchange, but
2 Analytic Application: Buy It or Build It? Giga Information Group, June 2000.
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
5
T H E E VO L U T I O N TOWA R D S PAC K AG E D P E R F O R M A N C E M A N AG E M E N T A P P L I CAT I O N S
also saps the productivity of both units. Thats an important consideration in smaller organizations, or those challenged by a downturn in economic fortunes. Similarly, an IT organization would have to devote enormous resources to developing the sophisticated engines that drive leading packaged performance management applications. BI vendors have invested heavily in underlying engines that, for instance, use set-based technology to segment large data volumes and offer users an easy, visual means of executing what is, under the hood, a complex query.
Lower Costs, Greater ROI
Its self evident that a well-executed performance management application implementation will deliver substantial return on investment to organizations in virtually any industry. BI technologies have generated millions of dollars in payback by helping organizations reduce costs, drive revenue, and improve customer and supplier relationships. But which approach yields the greatest ROIbuild or buy? IDC tackled this question in an October 2003 report, The Financial Impact of Business Analytics. Based on a study of 43 organizations in North America and Europe, IDC found that buy delivers a substantially larger ROI than build:
Median Build Buy
Five-Year Investment $2,088,860 $1,807,656
Median Return 104% 140%
As IDCs research shows, the buy approach is also less costly than build, by a factor of 15%. Part of that savings is due to faster time to deployment; avoiding the lengthier design, develop, and deployment cycles required to build an application can reduce IT costs by tens or hundreds of thousands of dollars. A number of soft factors figure into cost and ROI analyses that are not easily quantified. For instance, its difficult for an organization to predict the extent of maintenance, troubleshooting, user training and technical help, and customized documentation that may be required to support a custom-built application. Packaged performance management applications offer a clear advantage in those regards. The buyer organization enjoys access to the vendors service and support for maintenance and troubleshooting, as well as training on how to use the application for maximum advantage.
6
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
T H E E VO L U T I O N TOWA R D S PAC K AG E D P E R F O R M A N C E M A N AG E M E N T A P P L I CAT I O N S
Embedded industry best practices
A compelling feature of a packaged performance management application is that it has embedded best practices and domain knowledge specific to a business function or vertical. This built-in expertise is designed to provide users with a powerful, intuitive decision-making environment. By instantiating these features into a packaged application, vendors such as Business Objects take advantage of the wealth of knowledge developed over many years by leading practitioners in CRM, supply chain management (SCM), finance, human resources, and other areas. Business Objects relies on several sources of intellectual capital: Our customers: In developing our applications, we have interviewed and surveyed hundreds of IT and business representatives to help determine what functionality, metrics, and reports are of most value. This process has provided invaluable feedback and has helped us develop applications that answer the most compelling business, the most compelling business questions, out of the box. Our workforce: As Business Objects moved into the performance management applications arena, we made a concerted effort to hire business analysts skilled in particular areas. These analysts work with our development teams to incorporate their subject matter expertise into our products. Industry organizations: Business Objects applications reflect best practices developed by trade organizations and experts in many industries. For instance, BusinessObjects Supply Chain Intelligence features many of the best practices adopted by the Supply Chain Council, a consortium of more than 700 member companies. The resources that a software vendor of Business Objects size is able to commit to performance management applications are clearly far larger than those available to the development team and business analysts of an individual organization.
Superior Business Usability and Analytic Depth
For the business user, a superior performance management application will be distinguished by four key characteristics: visual presentation, metrics and alerts, guided analysis, and predictive analytics. Each of these characteristics contributes to the applications facility for top-down analysisranging from high-level questions monitored by metrics, to exploratory analysis of complex data.
Visual Presentation
A key to the value of a performance management application is its ease of use, with visual presentation that is readily and understood easily navigated by the user. But the
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
7
T H E E VO L U T I O N TOWA R D S PAC K AG E D P E R F O R M A N C E M A N AG E M E N T A P P L I CAT I O N S
developers job in fashioning an interface to business user specifics is not unlike a police sketch artist drawing a suspects face based on a witness description. Theres a lot of trial and error, erasing and highlighting, and the results are not always perfect.
Each analytic is presented in chart, graph, and textual format to support the analytical preferences of different individuals.
Packaged applications accelerate the business requirements gathering process by providing a ready-made interface from which developers and business users can work. In many cases, a packaged solution may require only some tweaking before it meets the approval of the business audience. In addition, the development facility bundled within a packaged application allows easy customization by individuals or roles. Roles-based presentation enables delivery of specified data by managerial authority, departments, geographic location, and other criteria.
Metrics and Alerts
At the highest level, the performance management application endeavors to make it easy for a business user to continually monitor, measure, and improve business performance. This is best achieved through metrics and alerts. By providing intelligent snapshots of critical information distilled from a large reservoir of enterprise data, metrics help put the decision-maker in control. Alerts push to the fore anomalies and exceptions that require immediate action. Building metrics and alert capabilities into a custom application requires an underlying engine that processes source data based on performance business rules. Developing the business rules and data mechanisms behind each metric can involve hundreds of hours of development time.
8
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
T H E E VO L U T I O N TOWA R D S PAC K AG E D P E R F O R M A N C E M A N AG E M E N T A P P L I CAT I O N S
With a packaged performance management application, organizations may choose from 40 or more prebuilt metrics. Not every metric will be applicable to the organization, but those not used can be easily discarded. Those that are used may be customized in far less time than it would take to build the metric from scratch.
Guided Analysis
Beyond metrics, a performance management application should allow users to sort through data to make an intelligent decision. With an application customized around a BI tool, this may mean hit-or-miss ad hoc queries. Applications built from scratch may feature little more than static report.
Guided analysis built into the analytical interface provides users with a roadmap to discovering root causes of business issues.
A sophisticated packaged application will feature guided analysis. This technology provides a navigational path to non-technical users, with prompts and predefined query selection to help isolate the root cause of issues and take corrective action. For instance, guided analysis can help trace product quality problems that have led to a rash of customer complaints about a defective component from a particular supplier.
Predictive Analytics
Performance management applications excel when they allow business users to extrapolate from historical data trends likely to arise in the coming weeks and months. This predictive capability is a step beyond the rear-view mirror approach of generic BI tools and helps position an organization as a proactive competitor in its market. Some large organizations will take a best-of-breed approach to predictive analytics by coupling custom-built applications with third-party modeling, forecasting, and statistical
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
9
T H E E VO L U T I O N TOWA R D S PAC K AG E D P E R F O R M A N C E M A N AG E M E N T A P P L I CAT I O N S
analysis engines. They face a considerable development challenge, however, and this technique is beyond the reach of smaller organizations with resources constraints. The best applications obviate the need for that extra development. Business Objects, for instance, offers a predictive analysis engine, providing customers with powerful behindthe-scenes technology to anticipate and prepare for future conditions.
Extensibility Across the Enterprise
Organizations that implement performance management applications often do so to meet a tactical needto reduce procurement costs though supplier consolidation, or accelerate revenue cycle quote-to-cash processes. To maximize the value of performance management applications, however, its important that these tactical goals align within larger strategic framework. In time, many organizations will adopt a common analytic framework atop which applications cooperate across functional areas. This model of integrated enterprise analytics helps organizations identify and optimize interdependencies among unitshow a sales uptake affects procurement, manufacturing, and logistics. Organizations that tailor a BI tool with analytics or build an application from scratch should be careful to ensure that the software can evolve to fit within a strategic analytic framework. Otherwise, IT staffs run the risk of building incompatible stovepipes that require further customization. Packaged applications offer greater extensibility than custom solutions. With Business Objects, modules and analytic solutions may be deployed individually and automatically snap together for cross-functional analysis atop a common framework. They may be licensed with optional connectors to enterprise applications from SAP, Oracle, and Siebel, to work with the Business Objects data integration products. A further risk associated with custom-building performance management applications is the loss of institutional knowledge that can occur when developers change departments or leave the company. This can present problems in strategically extending the applications, or implementing upgrades and patches. With a standards-based packaged application, this risk is virtually eliminated, and customers are assured of regular upgrades that reflect the continuing research and development efforts of the vendor.
10
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
Case Study: After Two Builds, Multinational Distributor Opts for the Buy Approach
After two years of costly and time-consuming effort to build from scratch a pair of digital dashboards for sales personnel and external vendors, a multinational worldwide products distributor had enough of the build approach. When it came time to deploy a third dashboard for B2B customer analysis, this multi-billion dollar company opted to buy a packaged performance management application. BusinessObjects Customer Analytics has been rolled out to allow some 300 managers and CRM specialists to track and analyze B2B customer activity through a personalized interface rich with key metrics, alert capabilities, and analytic drill-through to explore details behind customer activity. With it, the distributor expects to realize faster time to deployment, lower cost, and a strategic analytic framework that can extend across multiple departments to deliver a single view of the business. According to an IT manager at this company, The reason we went with BusinessObjects Customer Analytic is we asked ourselves, Do we really want to invent these things from scratch every time, or do we want to go down the fast path with an application thats already halfway there? The companys initial sales and vendor dashboards were developed on a Microsoft portal platform using COM Objects, XML, XSL style sheets, XML Query, and other technologies that presented a learning curve to the development team. Ongoing maintenance and patches also proved a challenge. And because the dashboards were built from scratch, developers spent an inordinate amount of time gathering and revising requirements from business users.
Reducing Cost and Time to Deployment
The IT manager estimates that it took the equivalent of three years of a single individuals time for his internal developers and external consultants to develop the sales dashboard, rolled out in August 2001. Roughly the equivalent of two years of a single individuals time went into the vendor dashboard before its debut in August 2002. With a third dashboard based on BusinessObjects Customer Analytics, the manager expected to reduce development time by at least 25%, if not more. Besides time and cost savings, BusinessObjects Customer Analytics promises to deliver greater business value through robust reporting and analysis. With Business Objects, I can construct a query and drill through to the source data and find my answer, the
B U I L D O R B U Y : M A X I M I Z I N G T H E V A L U E O F P E R F O R M A N C E M A N A G E M E N T A P P L I C AT I O N S
11
manager said. Thats difficult with the Microsoft dashboards we builtwe dont have ad hoc query capabilities. A key benefit is faster, less costly development of widgets, mini-applications that present data from transactional sources to the desktop. Gathering requirements and tailoring widgets to user preferences proved time consuming, and some custom-built widgets got little use. Developing a widget from scratch could take hundreds of hours. Thats reduced to 40 hours or less with Business Objects. With Business Objects, we can use widgets out of the box and develop and customize around them, the IT manager said. We believe it will be easier using Business Objects because of its report-like development as opposed to hard coding. Two years ago, this company had little choice but to hard-code its dashboards. Today, with an 18-mod...