Data warehouse initiatives traditionally have centered on integrating and provisioning available business data. For some time now, architectural focus has been on the processing of data — its extraction, cleansing, transformation, and consolidation — and the deployment of data analysis tools (reporting, OLAP, data mining, etc.).
Sure, there's been some debate about the best data-warehousing approach — consider Kimball vs. Inmon.1 And certain issues, such as master data integration, have recently resurged in interest, though they've arguably always been important and, in the case of business intelligence (BI), often were dealt with in the ETL context. For the most part, though, architectural principles in data warehousing haven't really changed that much in the last 25 years.
That is, until now. With the recent move toward an enterprise service-oriented architecture (enterprise SOA) approach to platform design, BI practitioners now have the ability to switch from a data-centric perspective to a process-centric one — making business intelligence an essential and inherent value-add of day-to-day business activities.
Are BI Users Ready for This Shift?
The good news is that most BI projects to date have proven to be successful. Commonly cited benefits
of BI implementations — garnered directly from BI customers — include increased transparency, more effective collaboration, and better business alignment. The number of companies deploying BI as a dedicated IT discipline has also been growing steadily. There are, for example, over 13,000 installations worldwide of SAP's BI solutions today (including SAP NetWeaver BI and SAP BW), up by 20% year-over-year. This expanded BI footprint has also furthered the development of industry-wide BI best practices.
Still, customers do not yet use BI to its full potential. Analyst research has shown that BI utilization is currently only at 30% of where it could be today. This means that in any given organization, 7 out of 10 potential beneficiaries of BI miss out entirely.
What is the reason for this BI underutilization? Given IT's steadfast efforts, it isn't a question of data availability. We can't look to tool complexity anymore, either; with such strong enhancements on the vendor side over the last several years, BI tools out there generally do a very good job. No, the issue now lies with providing context-relevant and timely information to users. It is about bridging that "last mile" between a BI platform and the information user — the gap that has historically proven so hard to fill and that has strangled many well-intentioned information deployment efforts caught up in custom coding or infrastructure restrictions.
After all, if information is not available where
and when users need it, it is often worthless — busy businesspeople simply don't have time to hunt for data. So what is the right context for analytics? For starters, it's the business users' work environment. To be more exact, it's the business processes that those individuals already use, and it's the devices and technologies they are familiar with.
Many leading companies have recognized this. As a result, we are now seeing the shift from
traditionally data-centric BI deployments toward process-centric ones, many of which incorporate embedded, in-process analytics that leverage business intelligence in the context of business needs.
Some more good news: SAP customers do not have to wait to get started with process-centric BI. The tools and technologies to take business intelligence to the next level, namely SAP's business process
platform and analytics offerings, are already available today. But before we look at these offerings in more detail, let's first understand — from an architectural perspective — how process and business intelligence can work together for more effective, broader BI use.
Process-centric BI means that business intelligence information is injected into users' daily tasks in multiple, pervasive ways. Consider a purchasing manager selecting a supplier, a cost center manager approving an expense, or a call center agent responding to a customer request — in each scenario, immediate access to all relevant information is critical.
Where Will the Intersection of Process and BI Make the Biggest Impact?
The combination of business processes and enterprise information brings some of the biggest benefits in three distinct scenarios:
1. Embedding Analytics Within a Process
When users consult their enterprise systems to help them make decisions, they often have to leave the transactional system they're working in to access analytic applications. This context switch between transactions and reporting not only costs time, but also discourages users from accessing these systems unless it's absolutely necessary.
Only seamless integration between transactional and analytic applications can provide necessary information the moment a decision must be made. When these two systems are cobbled together, any boundaries between the transactional and analytical worlds can be removed — a customer's history and credit worthiness can guide the release of a sales order; a supplier rating can optimize a procurement
manager's supplier evaluation; a budget forecast can help a marketing director evaluate a campaign. In some cases, analytics may even automate the decision making process entirely, something that research firm IDC has coined "intelligent process automation."2
|The question is no longer "What data do we have?" but "How can our data support, optimize, and innovate tasks and business processes?"
2. Using Analytics and Ad Hoc Collaboration for Process Innovation
Not all business processes are predefined, nor are they executed according to plan or schedule. In fact, many activities that businesspeople perform, though still within the context of a process, are highly unstructured, collaborative, and event-initiated.
Consider a shipment that is severely delayed because of a parts shortage. The shipping manager must immediately deal with this unexpected event. He needs to pull in key stakeholders to decide on a revised workflow on the fly. To execute this ad hoc process, analytics merely embedded into predefined work patterns would not be sufficient. A process platform — which, for SAP customers, is enabled by SAP NetWeaver — must be in place to provide the infrastructure for ad hoc collaboration. This infrastructure includes tools such as guided procedures and collaboration rooms that are easy to use and effective in problem-resolution situations.3
3. Using Analytics to Improve the Process Itself
You can only improve what you measure. The cycle time of the procurement process, the fulfillment rate for on-time deliveries, the impact of the bullwhip effect in a supply chain — these are all measurements with two things in common: a time component and a multi-step component. It's only by measuring multiple process instances and analyzing those instances over time and across steps that you can develop new insight into process execution.
When analytics aggregate hundreds of thousands of similar process instances, business users can discern valuable information like average cycle times and the root cause of exceptions, enabling benchmarking and more effective process execution. This takes process-centric BI to the next level, using analytics to evaluate business processes and determine if they're contributing to a company's success. What then are the underlying technologies that can support these intersections of process and BI?
The SAP Technologies for Process-Relevant BI
SAP NetWeaver, if combined with best-practice, SOA-based design principles, enables IT to evolve the
traditional data warehouse-centric BI architecture toward a more process-centric approach.
This transformation is made possible by the
combination of five market and technology trends:
1. Changes in Information Infrastructure
Data warehouses are evolving from central data repositories into transactional data hubs — commonly referred to as the "corporate memory" — that are complemented by real-time operational data. The data warehouse within SAP NetWeaver is a prime example. Master data management systems — such as SAP NetWeaver MDM, which holds the single version of truth for key master data — are becoming an integral part of that corporate memory as well. It's those data hubs, together with common metadata repositories, that are the fixed stars in an agile and self-renovating information landscape. Federation engines provide the semantic alignment, while data quality services assure the management and control of information quality from a business perspective.
|A process-intelligent infrastructure is the linchpin for BI practitioners and business users to optimize and innovate their
business processes for maximum value.
2. Advances in Process-Management Tools
Some of today's business process management (BPM) engines are enabled by business process
execution language (BPEL). These BPEL-enabled engines, such as the one integral to the process integration capabilities of SAP NetWeaver, not
only provide a new level of flexibility in process
definition and execution, but deliver an often overlooked benefit — the immediate measurability of the process execution itself. Every process instance has a unique process identifier attached, which enables the effective tracking, monitoring, and measurement of business process execution.
3. The Arrival of SOA and Web 2.0
Both the move toward service-oriented architecture and the onset of new, collaborative Web 2.0 technologies are changing the game for enterprise software users:
- SAP NetWeaver, for example, incorporates collaborative technologies to enable new user experiences, all supported from one platform and equipped with security, management, and governance.
- SAP NetWeaver Visual Composer provides composite mash-up environments, pulling from services defined within an enterprise repository and
eliminating the boundaries between transactions, analytics, and devices.
- Queries are taking on the shapes of applets like Yahoo! Widgets and Google Gadgets.
- Users can subscribe to alerts through Microsoft Outlook.
- Web browsers deliver rich desktop experiences via AJAX-type, Adobe Flex interfaces.
- Search technologies enable the immediate
pinpointing of relevant queries.
It's a new technological era, one in which collaboration tools foster immediate issue resolution and the rapid transformation of information into knowledge.
4. In-Memory Analytical Processing
Moore's Law has provided 64-bit technology just in time.4 If not for enhanced in-memory analytical processing capabilities, it would be impossible to enable an enterprise search or deliver relevant information to the right user in a personal, pointed fashion.
When BI is injected into operational processes, perhaps even optimizing transactional execution, it needs to fulfill the same strict, real-time requirements of an operational, mission-critical system. That's why SAP has been collaborating with technology titans like Intel to harness the power of the latest multi-core, 64-bit processors and to break down the traditional shackles of database-centric information design.5 SAP NetWeaver BI Accelerator breaks new ground here by providing instantaneous query response times, even as the number of underlying transactional records grows into the billions or more.6
5. Evolving Industry Best Practices
No two industries are the same. Retail has different process requirements than the pharmaceutical industry, and process needs in telecommunications look nothing like those in consumer products. A robust technical infrastructure alone cannot meet all industry needs. True business intelligence comes only when industry best practices, garnered from hands-on experience, are embedded into the software.
SAP has the experience here, having worked with companies across a wide spectrum of industries for over 30 years. By extension, SAP experts can have meaningful discussions with customers, helping them develop a comprehensive approach to process-centric BI. SAP also delivers prebuilt, industry-specific BI models that are packed with content, including critical KPIs and metrics.
A Hands-On Look at How Process and BI Come Together
The SAP Discovery System for enterprise SOA, built and delivered jointly by SAP and HP, is an ideal environment to experience how process and analytics come together via a business process platform. Delivered in a fully configured and preintegrated environment, the SAP Discovery System for enterprise SOA offers an out-of-the-box starting point for unleashing the power of enterprise SOA — a faster, more flexible way to adapt and innovate business processes.
The HP-powered SAP Discovery System leverages industry-leading technology specifically designed to optimize SAP NetWeaver. Running on low-cost HP ProLiant DL380 G4 servers with innovative 64-bit Intel Xeon processors, this system can be immediately deployed to give you a prototyping, development, and learning environment. Customers can test-drive a sample sales and procurement composite application using enterprise services that are built on the SAP NetWeaver platform and that link SAP and non-SAP systems. For ordering information, please visit www.hp.com/go/server/sap.
Recommended Next Steps for Process-Centric BI
With the availability of new business process and analytic technologies, BI has reached a turning point. BI deployments — and the analytic insight they
contain — can now be embedded into business processes. This evolution, and the integration of process and BI, makes it possible for customers to fully unleash the power of BI and gain more value out of their business intelligence investments.
That said, certain things need to come together for process-centric BI to work:
- Customers must understand their processes, and the standards that govern those processes, extremely well. To get started, we recommend embedding analytics into one or two of your most critical business processes — ones that you already understand thoroughly. Any improvements here will likely make the biggest impact for your company, and will help you build a business case and secure management buy-in for your BI project.
- To optimize your existing business processes, you'll need a state-of-the-art business process platform that also integrates master data — a prerequisite for tying together analytics and business process steps. If you haven't already, consider moving to
SAP NetWeaver 7.0, which provides the required technological advancements in one integrated platform. To experiment with this platform, consider deploying the SAP Discovery System for enterprise SOA (see sidebar), which provides SAP NetWeaver 7.0 technology, such as Business Intelligence, Master Data Management, Process Integration, and Portal, as well as SAP ERP 6.0. The SAP Discovery System also comes with detailed documentation and preconfigured business scenarios. Customers can, for instance, use the included portal-based SAP NetWeaver Visual Composer to orchestrate business processes with high analytic content.
- Finally, you've got to be able to analyze huge amounts of data in near-real time for in-process analytics and BI to be truly effective. To increase query-processing speed, consider implementing SAP NetWeaver BI Accelerator, which is designed to make lightning-fast response times a reality. You can implement it without having to reengineer any existing data models, enabling you to move gradually toward enterprise SOA.
Many companies have started putting process-centric BI in place; some are already seeing the results. It's a simple formula, really: adding integra-
ted BI and analytics into business processes can equate to a higher return on your business intelligence investment. And with market-leading tools and technologies from SAP, you can get started with this BI transformation today.
The Reporting and Analytics 2007 conference in Orlando, October 15-17,
2007, for in-depth coverage of SAP
analytic tools and applications (www.reporting2007.com)
"Analytics: From a Niche World to a
Central Role" by Neil Raden (SAP NetWeaver Magazine, Winter 2007, www.NetWeaverMagazine.com)
"Intelligent Process Automation: Let Analytics Drive Your Workflow" by Henry Morris (SAP NetWeaver Magazine, Fall 2006, www.NetWeaverMagazine.com)
Lothar Schubert is Director of SAP NetWeaver Solution Marketing, focusing on BI and information management. Lothar has over 10 years of industry experience in
the enterprise application space, having also assumed roles in product management, consulting, and project management. Based out of Palo Alto, California, Lothar holds a graduate degree in physics, as well as an MBA from UCLA. He can be reached at firstname.lastname@example.org.
Jens Doerpmund is a Senior Solution Architect in HP's consulting organization where he leads the SAP NetWeaver BI team.
For the last nine years, Jens has been focusing on the design and implementation of enterprise
data warehousing architectures, predominantly in SAP environments. Jens has a master's degree in advanced computer science from the University of Manchester, UK. He can be reached at email@example.com.