Model Purpose,
not Processes
Chris Marshall
SES
Software Inc
4694
Monticello Circle, Marietta, GA 30066, USA
|
Abstract
This paper describes
how modeling of businesses may be simplified by separating the concepts of
business purpose (why) and business process (how). In particular, the
hierarchical decomposition of a complex business process may be simplified or
eliminated by an equivalent decomposition of its purpose. This distinction is
also useful in describing commerce between organizations in terms of
contracts. Examples from traditional business are used to illustrate these
concepts. Keywords Contract, organization, process,
purpose, value. |
Every business has a purpose
towards which its energy and enterprise are directed. As we move into the
twenty-first century, business is as much between as within organizations.
While the internal purpose of an organization has been modeled by hierarchical
decomposition of its vision into attainable objectives, the purpose shared by
organizations for them to collaborate effectively is often not formally
recognized, let alone described. This paper outlines various concepts and
artifacts which businesses use to manage their internal and external purpose.
Business purpose describes why an organization exists, and business
processes describe how such purpose
is to be achieved. By separating these concepts, and by delaying the choice and
instantiation of a particular process to achieve a purpose until the latest
possible time, a business increases its ability to adapt to changes in its
markets and availability of its resources. This becomes increasingly important
as technology enables new patterns of e-commerce, and as the average size of
organizations decreases. Outsourcing, manufacture-to-order, sub-contracting,
niche-marketing, direct sales, and other such trends illustrate this position.
In a such a world, where value
networks form and disperse rapidly and continously, there is no central point
of control, and it is not feasible to devise or enforce systems to manage the
disparate business processes of participating parties as if they were one
process. Indeed, the processes may be confidential to a party, or may be
selected from a range of alternatives according to prevailing local conditions,
or may be delegated legitimately to other parties, and so on.
To avoid such problems, in
particular premature decisions as to how work is to be done, business should
describe itself in terms of desired outcomes, not prescriptions as to how such
purpose is to be achieved. This is even recognized in contemporary military
organizations, where authority is absolute, in which "mission command and
control relies on the use of mission tactics in which seniors assign missions
and explain the underlying intent but leave subordinates as free as possible to
choose the manner of accomplishment" [Sander 1997].
Naturally, the means by which
outcomes are achieved must also be described, particularly if the processes are
to be automated. However, such process definitions are typically for small
grained objectives, so are simple and direct. Where alternatives exist, each
may be described separately, to be selected at execution time according to the
ruling purpose. For example, selection among alternative shipping modes (air,
road, rail) is often based on a delivery deadline, which depends on its
purpose, not on the process itself.
Small grained processes of this
nature are often reusable, have alternatives, and may be refined piecemeal,
thereby avoiding the need for expensive and risk prone re-engineering efforts.
By communicating purpose between organizations, each remains free to determine,
and improve, the processes by which the purpose is achieved. While this appears
to be contrary to the centralized planning of most ERP systems, it is not so.
Such systems merely need to manage objectives, not the means by which the
objectives are to be met. In most cases, resource scheduling such as materials
and capacity planning may be delegated to the peripheral organizations
responsible for achieving the objectives.
Finally, the paper elaborates
on the concepts of contract and business messages introduced in last year's
paper [Marshall 1998], in the light of developments since that time.
[Marshall 1998] Chris Marshall,
Organization in a Chaotic World, Business Object Design and Implementation
IV, OOPSLA ‘98, 1998
[Sander 1997] Emile Sander, Marine Corps Intelligence and Security
Doctrine: Command and Control, US Marine Corps, 1997.
Chris
Marshall is a (semi-retired) director of a number of software companies that
have been active and successful for many years in delivering mission critical,
enterprise wide information systems using object technology. He has been responsible for managing the
development of large scale manufacturing, engineering and accounting systems in
C for UNIX, and more recently in Java, for companies throughout Southern
Africa. He was previously a general
manager of a number of divisions of the largest manufacturing company in South
Africa, and before that developed power systems software and managed projects
on the Zambezi. He has an MBA from the
University of Cape Town, and MA Mechanical Sciences from Cambridge in England.