The Outlook
VOLUME 2 ISSUE 3  
HOME
The Wave Front
Business Process Management – Demystifying the Challenges
by Edgard Capdevielle

Organizations in today's environments are limiting their spending to urgent projects with quick ROI and significant impact. These projects primarily include: security, storage, cost reduction initiatives, and integration of business processes. Organizations have overspent on software purchases and they need to implement and integrate those technologies, reaping the benefits before focusing attention on new initiatives. In today's environment, IT buyers are looking to make sense of their different systems, speed up their processes, cut out the fat, and automate manual process integration and exception-handling that were not delivered by their software vendors.

So much attention is now focused on orchestrating and integrating existing systems that many analysts believe that Business Process Management (BPM) is becoming the killer app that unleashes value from existing software assets. This can be accomplished by dynamically capturing and managing business logic with fully integrated application services. This article attempts to bring clarity to the BPM space and point out a few areas of innovation that we, as early-stage venture investors, are exposed to.

From Processes to Process Management

In a business context, the concept of Process Management refers to digitizing and supporting enterprise activity with software. More specifically, we refer to the following capabilities:

1. Design: Create, modify, and reuse processes.
2. Monitoring: Read and analyze metrics.
3. Operation: Execute and integrate processes.
4. Optimization: Provide enhancements and manage side-effects.
5. Automation: Leverage rules and intelligence to take unnecessary human labor out of the equation.

By definition, all enterprise software applications will have some process management functionality, but in this article we will refer primarily to BPM as pure-play applications that deal with managing business processes generically. Let's now identify the types of processes that are usually found in an enterprise and are candidates for automation and management. Processes are often classified according to four key criteria:

1. Time: the period of time they take from start to finish.
2. Complexity: the number of people and systems involved, inside and outside of the company. It also refers to the level of accuracy imposed in this process and the number of exceptions and side-effects that may occur.
3. Criticality: the magnitude of costs and benefits that the process affects.
4. Frequency: how many times per period does the process need to be executed.

In this multi-dimensional spectrum, we have processes that can be as simple as consolidating sales data and as complex as re-engineering product production. Depending on where processes fit in our four-criterion map, they could be easier or harder to digitize and manage by an enterprise application. Now that we have defined a framework to understand the diversity of business processes, we can explore the evolution of process automation and integration efforts in the enterprise.

BPM Market and Evolution

It is very hard to identify the birth of BPM as a software category since processes and applications have always been closely associated; but it can be traced to the first-generation of workflow solutions first offered over a decade ago. Workflow technology has moved from simply providing routing of work and documents between people to automating the intra and inter-company routing of work steps between the involved resources (people and systems). Implementing workflow used to entail hard wiring of the processes involved, which were automatically outdated at deployment due to the complexity and changing nature of people and business. Workflow solutions tend to be "static" and have a hard time modeling dynamic business processes and the exceptions that take place. Traditional workflow engines also lack the necessary metrics for process analysis and generally address them as an afterthought.

Managing processes and data consistently across applications has also driven Enterprise Application Integration (EAI) efforts. The first generation of EAI products emerged shortly after, allowing more flexible, scaleable and manageable application integration, but largely lacking BPM functionality. These early integration platforms were somewhat proprietary, costly, and point-to-point, proving to be prohibitive for common adoption. For example, if a company spent significant resources integrating its many systems and wanted to add a new sales support application it was still required to hardwire the new application to each of the systems in place. This made integration efforts increasingly complex and expensive. A major EAI technology breakthrough emerged in the mid-90s with the wide adoption of Message-Oriented-Middleware (MOM). MOM is a common bus or connection point that leverages third-party message queuing services to offer scaleable integration for high-volume transaction systems and distributed locations. This technology brought the necessary standards for solving the transport problem- getting messages from one application to another in a reliable, simple fashion. With this technology, applications needed to be integrated to the bus once, and not to every other application in the enterprise. MOM, however, still lacked basic functionality that had to be hard-coded, such as data translation and BPM capabilities.

Over the last 12 months, Web Services technology has started to emerge as the method of choice for application integration. This technology, when applied to legacy applications, encapsulates specific functionality (like adding employees to the corporate employee database) and exposes it to anyone who is authorized to use it through open standard methods. This newly defined application service, which is still part of a legacy application, can now be used as described by the standards that are part of Web Services. New applications being developed today include web services or other standard connectivity mechanism. The philosophy behind web services is that integration of applications is no longer a second thought that comes after the systems are in place, but is embedded in each application. The software is now responsible for connecting to its environment through standards. Web Services is not a magic wand that can be applied to all functions in an application and it is still in its infancy, but it is proving to be the next-generation integration solution. By itself, web services technology does not address BPM, but it may be an enabler because processes need to dynamically use a combination of services that can easily come together through web services technology.

Today, most integration solutions have added some BPM functionality to their suites. In addition to standard message queuing, EAI suites provide standard business protocols for data translation, security and transaction integrity; which are all necessary for BPM to make sense. As the focus shifts from application to processes, BPM applications are required to have a repository of processes and a semantic understanding of the enterprise and its resources. Deployment and monitoring of processes still involves a mix of workflow, integrated systems, and a common business understanding of the process and capabilities of available resources-this last component is still delivered through manual hard coding. BPM and its enabling technology has advanced tremendously in the past few years as described in this section, but in order to approach BPM holistically one would need to use best-of-breed BPM technologies.

A recent and significant step forward is the publication of the Business Process Modeling Language (BPML) by coalition industry groups. BPML simplifies the identification of application functionality as process objects that can be aggregated into larger processes and reused where appropriate. The Outlook Ventures team believes that the introduction of BPML is a disruptive breakthrough that for the first time allows processes to be treated as objects, with real attributes and behavior. We think that BPML is to process management what Object Oriented Programming was to software development. For example, a process like "adding a new sales person to a company" cuts across many systems (sales, HR, and finance) that can be integrated through EAI technology, but the process itself can now be treated, changed, analyzed and managed as a process object consisting of integrated functionality that resides in 3 different systems. We expect to see many innovations in this area.

Another breakthrough coming from a different angle is driven by a different standard called the Web Service Choreography Interface (WSCI). WSCI is also supported by most industry leaders and is a standard that deals with coordinating and orchestrating web services that form business processes. Very similar to BPML, WSCI is the web services approach to BPM.

To this day, however, most IT buyers and business professionals admit that less than 25% of their business processes are digitized, much less managed or measured. Let's shift focus to how IT buyers are looking at BPM technology.

Buying BPM Today

IT buyers today are faced with a variety of options when looking to implement a BPM solution. Software vendors are entering the market from their own segments, including workflow, integration, transaction, collaboration, etc. The common element is that organizations are taking a process-centric view of the world and vendors will need to leave their document-centric, application-centric, or transaction-centric views and buy into the trend that next generation BPM solutions will work at a higher level, where process logic is truly separated from application logic.

In the current business environment, IT buyers are not willing to implement a solution that would require a major overhaul of their infrastructure and our assumption is that selling holistic BPM solutions only happens in scenarios that include systems and applications that are in place today. Based on this assumption, IT buyers need to ensure application integration from the new CRM or eCommerce application to the outdated legacy mainframe. Comprehensive integration of systems generally addresses the first issue that a BPM implementation needs to tackle: the platform. Most EAI initiatives require a standard communication platform that can be leveraged in BPM implementations. These platforms include Microsoft's .Net and J2EE, and can be complemented with web services - a platform neutral technology.

Perhaps the most important aspect of purchasing BPM software is justifying the investment and working through ROI calculations. In today's environment and focusing on the cost side of the equation, BPM solutions need to be implemented quickly and be very flexible, supporting all enterprise processes. Fast implementations tend to have less variance and contain costs better. Piecemeal implementations are obviously preferred over big-bang projects and inclusion of current systems and processes is preferred over replacement of legacy infrastructure. On the benefits side, BPM brings visibility and improvement to business execution. Ultimately, the flexibility to cost-effectively automate all processes drives competitive differentiation. ROI calculations can be much more extensive and quantitative, but outside of the scope of this article. As most executives know, quantitative ROI calculations are as accurate as Chicago weather reports.

The next step is the selection of a BPM vendor, which is not a trivial task and must be closely aligned with each company's business goals. Platform vendors such as BEA, IBM, Microsoft, and SUN are adding BPM modules to their run-time platforms. Their solutions sit on top of the integration modules and just below their portal modules. Integration vendors like IONA, Tibco, Sterling Commerce, webMethods, and Staffware are morphing their integration offerings to add a more process-centric approach. Companies like Handysoft, Fujitsu/iFlow, Adobe/Accelio, and Ultimus are trying to make their workflow and collaboration products more aligned with corporate applications. It is worth noting that today, the most widely-used workflow tool is still plain e-mail.

As mentioned before, there are thousands of business processes not supported by any established category of software and are put in a generic bucket called Collaboration. Collaborative Applications that deal with processes include Zaplet, Lotus/Domino, and Collabranet to name a few. These tools do not track or measure processes and have avoided integration by not dealing directly with enterprise applications (using e-mail or a stand-alone portal as a medium for collaboration). These players may decide to come up with holistic approaches to BPM but are not there yet.

There are also a new set of players in the BPM space. Movaris, Intalio, Savvion, FuegoTech, and Q-Link are smaller private companies introducing products that separate business and process logic from third-party applications, creating and managing process objects. This allows them to provide process modeling, monitoring, and analysis capabilities. However, these players depend on the integration of applications that is outside their scope. Movaris and FuegoTech may be the only exceptions among these players because they do have connectors to popular applications to facilitate integration. In the new-entrant category there is also a set of players that come from web services management and may adopt a process-centric view. These include Actional, AmberPoint, Talking Blocks, and WestGlobal.

Even with all these options and approaches to BPM, the market is still lagging and adoption is not widespread. With research showing that 75% of the processes in an enterprise are still manual, then the majority of the value provided by process automation and management has not been realized. (See Figure 1)

Figure 1: Automation of Enterprise Processes

Processes

The working assumption is that automating processes adds value to the enterprise by reducing cycle time as well as bringing the necessary metrics and visibility. So, why are the majority of processes not automated and managed?

We believe that there are two main inhibitors to automating and managing the remaining processes that are slowing down BPM adoption:

1) The marginal benefit of automating an additional process starts to be offset by the cost associated with automation. That is, there is little incentive for IT buyers to support and integrate remaining processes with today's cost structure for automation and integration. Even if the process is simple, it may not be worth it to codify it in a workflow engine rule and integrate it with the rest of the automated processes. Today's budget scrutiny and ROI expectations emphasize this inhibitor further.

2) The more critical factor is that many processes in fact are just too dynamic and complex to be fully described in a workflow engine during setup. Some processes change constantly, have random exceptions, and evolve over time, needing on-the-fly re-configuration. For example, the process for handling product returns at Palm is very complex: it varies by vendor, product, price, part damaged, and type of damage. It just can't be codified, but still needs to be monitored and optimized.

The consequence is that about 3/4 of a company's processes are still a set of manual, time-consuming tasks. Therefore, there is little visibility into what the majority of the knowledge workers are doing and no way to optimize or learn. Today's BPM vendors address narrow aspects of a complete solution and we believe that a company implementing a BPM solution needs to consider a best-of-breed approach. After exploring the challenges of buying BPM technology, let's explore some of the more innovative approaches to tackling these issues.

What's next? Opportunities in BPM

As mentioned before, web services technology is still in its infancy and is by no means a mature technology. But as it matures we believe that orchestration of web services will be a valid component of implementing BPM functionality. The value proposition is simply to reduce the cost of integration, tackling the cost inhibitor mentioned above. In fact that's the approach that Intalio, FuegoTech and Q-Link have taken. In today's enterprise, there are few or no web services to integrate into processes, but as web services technology is adopted, service-based BPM could become a reality.

We also see a new breed of players approaching the market with a process-driven integration approach. Specifically, Agentis and Celcorp are using agent-based technology to watch and learn how enterprise users navigate through applications. They then create the functionality and integration to support navigation through cross-application processes. This type of solution is also attempting to tackle the cost inhibitor mentioned above.

A new private company, nSite Software (an Outlook Ventures portfolio company) is offering a next-generation BPM product that tackles both inhibitors at the same time: implementation cost, and the ability to manage dynamic processes. Their platform reflects the reality of constant changes in business processes by empowering users to create their own processes on-the-fly, and to facilitate changes to those processes in real-time, as exceptions necessitate. Visibility, metrics, and real-time processing of exceptions are the guiding principals of the nSite Process Server, which was built specifically to bring the benefits of process automation to the large constituent of exception-based manual process in today's business environment.

The market dynamics of related or tangential markets are also a big part of the increased visibility of BPM. We are aware of the increasing commoditization of application server platforms and we think that integration software will follow the same path. Our position is that as application and integration platforms become standard and widely deployed this new process-centric trend will emerge strongly, allowing for software-based innovations to reshuffle power and presence in this industry.

As we continue to witness the evolution of this space we are keeping an eye open for new developments and investment opportunities that solve actual business problems. We welcome your feedback on this article and are interested in learning about companies that are doing something innovative in this space.


This article was written by Edgard Capdevielle, Principal at Outlook Ventures, with valuable input from the Outlook Ventures CIO Council, the Outlook Ventures team, William Heppner (Experio), Reed Scott (CEO, Pakana Corp), and Oliver Muoto (VP Business Development, Epicentric).

References:

  • Delphi Group Whitepaper: BPM2002
  • Several articles from the EAI Journal
  • InformationWeek: October 14 issue
  • Several articles from eWeek Magazine
  • Several Gartner reports
  • Several Hurwitz Group reports
  • Several articles by Howard Smith of CSC
  • Information from BPMI.org


  • Published by Outlook™ Ventures
    Copyright © 2002 Outlook Ventures. All rights reserved.