BPM End to end, step by step

Eddie Page, the purchasing manager of Eastman Chemicals, had a process problem. When people ordered materials or parts through the company’s e-procurement system, the price agreed with suppliers did not always agree with those that had been pre-programmed into the SAP financial system.

So the purchasers were instructed to do the kind of thing that users of enterprise applications do the world over. They logged onto the financial systems and re-entered the new data. And, Page admits, they often did so late, and inaccurately.

It was, of course, a classic enterprise application integration (EAI) problem, which at least a dozen good software companies are expert at solving. But Eastman did not turn to an EAI vendor – such as Tibco, WebMethods, or SeeBeyond – for his solution. He turned to the company’s e-procurement software supplier, Commerce One.


BPM in a nutshell

Business process management software, and similarly business process integration software, is designed to exploit two basic premises. First, that application software largely consists of businesses’ processes that have been captured and automated in a formal way; and secondly, that these processes can be accessed and used by other, process-oriented high-level programs.

These high-level programs may be designed to carry out a number of jobs. For example, they may link together a few key processes in each application, and then provide a new, ‘composite application’ consisting of a number of ordered, workflowed steps; they may identify, analyse and extract many processes, and link them into new, real-time applications and processes; or they may be designed to centralise the management of many similar processes necessary to each application – such as an address change process.

The wide variety of tasks for which process-oriented software may be optimised partly explains why so many suppliers from different backgrounds believe they have a role to play.

A workflow software company such as Staffware, for example, will focus on linking processes together to form a structured flow of tasks, often involving human interaction; an integration vendor such as Tibco will focus on integrating critical, real-time processes from one core application, perhaps analysing the activity at the same time; an application company, such as Siebel, will focus on centralising and managing tasks that are core to their competency, such as managing processes that involve customers; and a development tools company, such as the Rational division of IBM, will focus on tools that enable a new application to be modelled, designed and developed to include new and existing processes.

The more radical of the process management advocates believe that many products carrying a process label are halfway solutions. They say processes should be clearly exposed using open standards (debatably, BPEL, backed by Microsoft and IBM), there should be a minimum (or no) amount of proprietary or low level programming, and systems should support the manipulation of many processes in different ways.

The so-called pure play BPM suppliers, such as Metastorm, Procession, Fuego, Intalio and Savvion, tend to take this more radical line, but not all will scale well and not all of them avoid the use of proprietary languages.


A few months earlier, Commerce One had approached Eastman with an outline of its revolutionary new product for building composite applications using web services standards – Conductor.

The idea was that, rather than use specific adaptors to convert data formats from one application to another, the processes involved in carrying out a task in two or more applications are simply linked together, using open web services, and used as the basis for a new ‘composite’ application. And it worked: “This tool was the fastest and the most capable of bringing off the project,” said Page. Eastman became an enthusiastic beta test customer.

Reconciling the prices held in two enterprise applications systems is valuable, if hardly revolutionary, but Page believes it is only the start. The goal now is to build more and more ‘touch points’ between the SAP and Commerce One e-procurement system, creating, along the way, more composite applications, each based on automating a single process.

Critically, these processes can be developed quickly, and, because they don’t interfere with the underlying applications, can be quickly changed. And, using the technology, it shouldn’t be difficult to integrate the processes with other applications – including those of customers communicating across the Internet.

What Page is talking about is process level integration – and that has become the new nirvana not just for Commerce One, but for just about every significant enterprise software supplier.

Suppliers head for P Spot

For Commerce One, Conductor is a bet-the-company product. If it fails, many analysts think it will be difficult for the downsized company to recover from the rapid disillusionment and contraction that has affected its original core markets.

But it is by no means alone in its headlong pursuit of process automation and integration software. All the major platform vendors (IBM, Microsoft, Oracle and BEA, for example), the big application vendors (SAP, Oracle, PeopleSoft), the integrated workflow vendors (Staffware, FileNet), the EAI companies (Tibco, SeeBeyond, WebMethods) and a host of start-ups and pure plays see ‘process’ – or composite applications – as one of the new sweet spots of enterprise software.

This optimism is partly, but not entirely, based on expected demand. Although the market is nascent, with users of BPM (business process management) or composite application systems numbered at best in their hundreds, businesses are expected to adopt this technology on a massive scale in the coming decades.

One good reason for this is that BPM is a technology that can give customers the ability to build, link and change their heavily computerised processes much more easily and, therefore, much more often, without having to invest heavily in expensive new IT systems or in enormous projects to link them together. (see box, BPM in a nutshell).

Some analysts’ firms have picked up on this and are advising clients to invest without delay: “Chief information officers should waste no time and should organise a process savvy team to kick off their first business process management (BPM) initiatives today,” said one Forrester Research report.

The potential cost savings are considerable. At present, enormous sums are invested in products, programmers and services to integrate processes. “Some 40% of [IT’s] cash and 80% of [its] manpower is today spent on integrating applications,” says Ian Howells, director of marketing for SeeBeyond, an EAI supplier that is re-engineering itself to become a composite applications builder.

But there is another reason why suppliers are rushing to embrace process-level integration. Technology strategists simply think that the software industry is irrevocably set on a course that differs radically from the orthodoxy of the mid-nineties, when big businesses were hard-wiring end-to-end processes into their complex ERP suites.

“Do not mistake BPM for some new ‘killer app’ or some fashionable new business theory. It is a foundation upon which companies can depend as surely as they depend on database management,” wrote Howard Smith and Peter Fingar in their book, Business Process Management: The Third Wave.

Now, in the software development community – whether at the leading universities or at IBM, Microsoft, or SAP – the fact that there needs to be a new architecture based on open, ‘plug in’ web services and applications, is almost universally accepted.

Making sense

But that universal acceptance isn’t leading to a clarity of vision. BPM is a classic early software market. The terminology is confused and customers are even more so. Some suppliers are evangelical and purist, some have edged in from adjacent fields, and some have the size and influence to become leaders in spite of the fact that they have been almost dismissive of BPM until now.

“The problem for anyone considering using this technology is working out who supplies what, where it fits in, and what it does,” says Anthony Howcroft, UK managing director of ‘pure-play’ BPM company, Intalio. He warns customers that all of the products “have a legacy in terms of the way they do things” and how they approach the problem.

Certainly companies from a development or integration background (such as IBM, BEA, SeeBeyond) tend to focus on tight, low-level integration, perhaps using J2EE. SAP (with its ‘X-Apps’ for linking different systems via a process engine) and Oracle (with its ‘business flows’) tend to focus on linking their own systems together using both standards and APIs (application programming interfaces), although it is possible to add-in external processes.

The approach of the pure-play start-ups, meanwhile, is to leave this lower level integration to other vendors. “With our product, there is no programming at all,” says David Chassels, CEO of UK-based company Procession.

Commerce One, with its procurement focus, is perhaps an unlikely contender in this area, but arguably no more so than Siebel, the CRM supplier, which introduced the universal application network (UAN), a platform for building composite applications that integrate front-office functions, in 2002.

Both suppliers would probably baulk at the comparison, but UAN has much in common with Conductor – not least of which is that both suppliers face a long struggle to win over customers and analysts (Commerce One claims about 12 customers, and Siebel about 25 customers).

Processes for sale

UAN, like Conductor, is a software platform for rapidly building composite applications that utilise the underlying functions and processes of the installed applications.

But both companies take the process a critical step forward. As application suppliers, their real business is in selling applications, not just platforms: and in this context, that means selling ‘processes’ – even, eventually, they say, to organisations that don’t use their main applications. These processes are co-developed with customers, and then packaged up as processes for resale.

Nimish Mehta, responsible for Siebel’s UAN business unit, for example, says the company has 121 business processes available, all of which will run on any software infrastructure that has been developed to support UAN. These range from customer lifecycle management to a simple change of address.

That is the ultimate goal for the process industry – commoditised, resaleable processes that run on any infrastructure. If – or when – it all works, it could dramatically alter the economics of IT.

There will be a few issues for the suppliers, too. As Dan Woods, a former chief technology officer turned author, writes in his book Packaged Composite Applications: “If vendors have been selling and caring for gorillas and chimpanzees, what will happen when they try to sell ants?”

Avatar photo

Ben Rossi

Ben was Vitesse Media's editorial director, leading content creation and editorial strategy across all Vitesse products, including its market-leading B2B and consumer magazines, websites, research and...

Related Topics