the problems we're facing

9
The problems weʼre facing Peter Evans-Greenwood [email protected] Companies are engaged in an arms race. For years they have been rushing to beat competitors to market with applications designed to automate a previously manual area of the business, making them more efficient and thereby creating a competitive advantage. Wal*Mart is the poster child for this approach. As the first retailer with an end-to-end logistics solution Wal*Mart slashed operating costs, leveraging this ad- vantage to deliver “Every day low prices” and drasti- cally undercut their competitors 1 . Wal*Mart used this competitive position to become the largest retailer in the world. Dell used a similar strategy in the PC market, pioneering an IT enabled direct-to-customer model al- lowing them to hold a minimum of inventory, drastically reducing operating costs. In some cases Dell was able to create an environment where customers paid for their orders before Dell paid suppliers, creating, in effect, negative working capital. Dell used this advantage to become the largest PC vendor in the world 2 . The end of applications Today, enterprise applications are so successful that it is impossible to do business without them. The efficiencies they deliver have irrevocably changed the business envi- ronment, with an industry developing around them a range of vendors providing products to meet most needs. It is even possible to argue that many applications have become a commodity (as Nicholas Carr did in his HBR article “IT Doesn’t Matter”), and in the last couple of years we have seen consolidation in the market as larger vendors snap up smaller niche players to round out their product portfolio. This has levelled the playing field, and it's no longer possible to use an application in the same way to create competitive advan- tage. Now that appli- cations are ubiquitous, they’re simply part of the fabric of business. Today, how we man- age the operation of a business process is becoming more im- portant that the busi- ness process itself. Marco Iansiti brought this into sharp relief through his work at Harvard Business Review when he measured the effi- ciency of deployment of IT, and not cost, and correlated upper quartile efficiency with upper quartile sales reve- nue growth 3 . Efficiently dealing with business excep- tions, optimizing key decisions and ensuring end-to-end consistency and efficiency will have a greater impact than replacing an existing application. Wal*Mart's supply chain solution cost billions in today’s money, a massive investment impacting their entire op- eration. The application created a competitive advantage, and the investment it required prevented competitors from responding quickly with a competing solution. The next generation of systems of a similar scope are taking 10% of the staff and 10% of the time compared to Wal*Mart’s original investment. Wal*Mart’s competitive advantage now rests on its buying power, not the level of IT support for its business processes. We are finished the big effort: applications are available from multiple vendors to support the majority of a busi- ness’s supporting functionality. The law of diminishing returns has taken effect, and owning or creating new IT asset today will not simply confer a competitive advan- tage. Competitive advantage now lives in the gaps between our applications. Exception handling is becoming in- creasingly important as good exception handling can have a dramatic impact on both the bottom- and top-line. If we can deal with stock-outs more efficiently then we can keep less stock on hand and operate a leaner supply chain. Improving how we determine financial adequacy allows us to hold lower capital reserves, freeing up cash that we can put to other more productive uses. Extending our value-chain beyond the confines of our organisation to include partners, suppliers and channels, allows us to optimize end-to-end processes. Providing joined-up support for our mortgage product model allows us to put the model directly in the hands of our clients, letting them configure their own, personal, home loan. Page 1 1 Charles Fishman , The Wal*Mart Effect, Penguin Press, ISBN: 1594200769 2 Steven Holzner, How Dell Does It, McGraw-Hill, ISBN: 0072262540 3 Reference U.S. software sales in billions of dollars Source: INPUT

Upload: peter-evans-greenwood

Post on 10-Apr-2015

411 views

Category:

Documents


0 download

DESCRIPTION

Companies are engaged in an arms race. For years they have been rushing to beat competitors to market with applications designed to automate a previously manual area of the business, making them more efficient and thereby creating a competitive advantage.Wal*Mart is the poster child for this approach. As the first retailer with an end-to-end logistics solution Wal*Mart slashed operating costs, leveraging this advantage to deliver “Every day low prices” and drastically undercut their competitors1. Wal*Mart used this competitive position to become the largest retailer in the world. Dell used a similar strategy in the PC market, pioneering an IT enabled direct-to-customer model allowing them to hold a minimum of inventory, drastically reducing operating costs. In some cases Dell was able to create an environment where customers paid for their orders before Dell paid suppliers, creating, in effect, negative working capital. Dell used this advantage to become the largest PC vendor in the world2.Today, enterprise applications are so successful that it is impossible to do business without them. The efficiencies they deliver have irrevocably changed the business environment, with an industry developing around them a range of vendors providing products to meet most needs. It is even possible to argue that many applications have become a commodity (as Nicholas Carr did in his HBR article “IT Doesn’t Matter”), and in the last couple of years we have seen consolidation in the market as larger vendors snap up smaller niche players to round out their product portfolio.This has levelled the playing field, and it's no longer possible to use an application in the same way to create competitive advantage. Now that applications are ubiquitous, they’re simply part of the fabric of business.Today, how we manage the operation of a business process is becoming more important that the business process itself. Marco Iansiti brought this into sharp relief through his work at Harvard Business Review when he measured the efficiency of deployment of IT, and not cost, and correlated upper quartile efficiency with upper quartile sales revenue growth3. Efficiently dealing with business exceptions, optimizing key decisions and ensuring end-to-end consistency and efficiency will have a greater impact than replacing an existing application.Wal*Mart's supply chain solution cost billions in today’s money, a massive investment impacting their entire operation. The application created a competitive advantage, and the investment it required prevented competitors from responding quickly with a competing solution. The next generation of systems of a similar scope are taking 10% of the staff and 10% of the time compared to Wal*Mart’s original investment. Wal*Mart’s competitive advantage now rests on its buying power, not the level of IT support for its business processes.We are finished the big effort: applications are available from multiple vendors to support the majority of a business’s supporting functionality. The law of diminishing returns has taken effect, and owning or creating new IT asset today will not simply confer a competitive advantage.Competitive advantage now lives in the gaps between our applications. Exception handling is becoming increasingly important as good exception handling can have a dramatic impact on both the bottom- and top-line. If we can deal with stock-outs more efficiently then we can keep less stock on hand and operate a leaner supply chain. Improving how we determine financial adequacy allows us to hold lower capital reserves, freeing up cash that we can put to other more productive uses. Extending our value-chain beyond the confines of our organisation to include partners, suppliers and channels, allows us to optimize end-to-end processes. Providing joined-up support for our mortgage product model allows us to put the model directly in the hands of our clients, letting them configure their own, personal, hom

TRANSCRIPT

Page 1: The problems we're facing

The problems weʼre facingPeter Evans-Greenwood

[email protected]

Companies are engaged in an arms race. For years they have been rushing to beat competitors to market with applications designed to automate a previously manual area of the business, making them more efficient and thereby creating a competitive advantage.

Wal*Mart is the poster child for this approach. As the first retailer with an end-to-end logistics solution Wal*Mart slashed operating costs, leveraging this ad-vantage to deliver “Every day low prices” and drasti-cally undercut their competitors1. Wal*Mart used this competitive position to become the largest retailer in the world. Dell used a similar strategy in the PC market, pioneering an IT enabled direct-to-customer model al-lowing them to hold a minimum of inventory, drastically reducing operating costs. In some cases Dell was able to create an environment where customers paid for their orders before Dell paid suppliers, creating, in effect, negative working capital. Dell used this advantage to become the largest PC vendor in the world2.

The end of applications

Today, enterprise applications are so successful that it is impossible to do business without them. The efficiencies they deliver have irrevocably changed the business envi-ronment, with an industry developing around them a range of vendors providing products to meet most needs. It is even possible to argue that many applications have become a commodity (as Nicholas Carr did in his HBR article “IT Doesn’t Matter”), and in the last couple of years we have seen consolidation in the market as larger vendors snap up smaller niche players to round out their product portfolio.

This has levelled the playing field, and it's no longer possible to use an application in the same way to create competitive advan-tage. Now that appli-cations are ubiquitous, they’re simply part of the fabric of business.

Today, how we man-age the operation of a business process is becoming more im-portant that the busi-ness process itself. Marco Iansiti brought this into sharp relief through his work at

Harvard Business Review when he measured the effi-ciency of deployment of IT, and not cost, and correlated upper quartile efficiency with upper quartile sales reve-nue growth3. Efficiently dealing with business excep-tions, optimizing key decisions and ensuring end-to-end consistency and efficiency will have a greater impact than replacing an existing application.

Wal*Mart's supply chain solution cost billions in today’s money, a massive investment impacting their entire op-eration. The application created a competitive advantage, and the investment it required prevented competitors from responding quickly with a competing solution. The next generation of systems of a similar scope are taking 10% of the staff and 10% of the time compared to Wal*Mart’s original investment. Wal*Mart’s competitive advantage now rests on its buying power, not the level of IT support for its business processes.

We are finished the big effort: applications are available from multiple vendors to support the majority of a busi-ness’s supporting functionality. The law of diminishing returns has taken effect, and owning or creating new IT asset today will not simply confer a competitive advan-tage.

Competitive advantage now lives in the gaps between our applications. Exception handling is becoming in-creasingly important as good exception handling can have a dramatic impact on both the bottom- and top-line. If we can deal with stock-outs more efficiently then we can keep less stock on hand and operate a leaner supply chain. Improving how we determine financial adequacy allows us to hold lower capital reserves, freeing up cash that we can put to other more productive uses. Extending our value-chain beyond the confines of our organisation to include partners, suppliers and channels, allows us to

optimize end-to-end processes. Providing joined-up support for our mortgage product model allows us to put the model directly in the hands of our clients, letting them configure their own, personal, home loan.

Page 1

1 Charles Fishman , The Wal*Mart Effect, Penguin Press, ISBN: 1594200769

2 Steven Holzner, How Dell Does It, McGraw-Hill, ISBN: 0072262540

3 Reference

U.S. software sales in billions of dollars

Source: INPUT

Page 2: The problems we're facing

Filling in the gaps & extending our reach

Service-oriented architecture (SOA) is a new approach to enterprise software supported by a broad range of tools and techniques across the industry, promising to dramatically lower the cost of integration and enable a more granular approach to IT. From humble beginnings as a light weight approach to integration, driven by the needs to quickly connect systems in complex and de-manding IT environments, SOA has grown into a revolu-tion similar that of client-server and the impact it had on how we design and build applications. After the Internet became established the web brought about the new ways of working, and now SOA is redefining how we plan and deliver IT.

SOA allows us to align new IT assets (services) with the business activities they are intended to support, letting us focus our IT investment on the high value differentiating business activities (those that a disproportionate impact on our competitive advantage) that live between applica-tions while leveraging low cost commodity functionality provided by applications where appropriate. It also pro-vides a holistic view of IT enabled business functional-ity, one that spans applications and extends beyond the bricks and mortar of traditional back-office operations of enterprise software to include field operations, client, partners and even competitors. Rather than being forced to treat each business function and requirement as equal and delivering solutions in application-sized chunks, SOA provides us with a holistic view of the business and allows us to focus our attention and energy where it will have the most impact.

In the back office we’re using composite applications to fill gaps between existing IT assets to provide users with a single consistent view into enterprise data and func-tionality. This allows them to focus on the task at hand rather than mediating between siloed applications, acting

as a manual integration point. The more holistic view of business functionality SOA provides enables us to string together the individual activities and create true end-to-end processes. Combined, these approaches allow us to deliver direct cost, efficiency and quality benefits to the business by providing a consistent level of automation across the enterprise.

The light weight and low cost integration provided by SOA lets us to expand IT operations out beyond the tra-ditional bricks and mortar of the back-office, bringing field operations, partners and even (in some cases) com-petitors into our IT environment. Mobility solutions are integrating team collaboration and back-office function-ality to deliver innovative solutions directly into the field. We’re all familiar with signing for a package on the device most delivery people carry, sending the deliv-ery receipt directly to the back-office. Airlines are using mobility solutions to provide maintenance teams with seamless IT support for turning aircraft around at the gate, connecting team actions with back-end scheduling and planning systems. Police departments are providing their people in the field with devices that that allow them to access back-end application functionality from the patrol car.

In many industries the sensor networks and the commu-nications network to support them are finally affordable, allowing us to instrument our operations to degree uni-maginable only a few years ago. For example, Wiscon-sin Public Service (WPS) (a little discussed mid-western utility power) deployed 1.2 million meters to collect intelligence directly from the field, pulling over 60 mil-lion data points a day. WPS has leveraged this data to optimize operations and is now moving to become the lowest cost operator of a grid in North America.

And finally, the (technical) ease of partner integration is causing many companies to rethink how they manage their operations. Wal*Mart was the supply chain leader in an application-centric world, causing Target to re-

Page 2

!"#$%$&'(!! !"#$%&'%()#*$)+$,-*)./).0+)

1,-2).3-$)

!! 4+.5'&'%()6$7&)8.+2$+9)8'#*):.'%$&;0<)=>)90<<.+#)

)*+,"-(.*&/"-0,(

!! ?$5$+,($)#$-*%.7.('$9)90-*),9)@A=B),%&)9C,+#),<<7',%-$9)#.)'%9#+0C$%#).<$+,D.%9)

)"1/2-*(2,(2()*-3$4*(5(!2,6789,(

!! !%,17$)$%&;#.;$%&)<+.-$99$9)E! ?.8$+)$++.+)+,#$9)

E! =%-+$,9$&)#*+.0(*<0#)

!! ?.8$+)09$+)8.+27.,&)E! F.)C.+$)98'5$7;-*,'+)

'%#$(+,D.%)

E! 4+.5'&$9)#*$)+'(*#)&,#,G),#)#*$)+'(*#)DC$G)<+$9$%#$&)#.)$%,17$)09$+9)#.)C,2$)1$H$+)&$-'9'.%9)

:2-&+*-(;+&*<-2="+(!! I,7$9),%&).<$+,D.%,7)&,#,)

!! @$9<.%9'1'7'#J)/.+)$%&;#.;$%&)109'%$99)<+.-$99$9)

!! A,-'7'D$9),%&).#*$+),99$#9)

Process & Information Integration

Platform

Application Services

UI UI UI

Field Operations Back Office Partners

Platform

Application Services

UI UI UI

Platforms

Application Services

UI UI UI

Page 3: The problems we're facing

spond by developing a new approach. They did this by handing responsibility for much of their supply chain over to their vendors, incentivizing the vendors to be-come very good at keeping their products on the shelves. Home Depot took the idea even further and stepped around the barrier of the supply chain to hand not only the supply chain, but all the stocking, store layout and pricing of the product category directly to many of their the vendors, providing the vendors with a direct feed on sales. Home Depot has significantly reduced the cost of what they do while increasing the number of sales per item in their stores above the level from when they did it themselves.

The boundaryless environment

The Internet has connected systems in number and scale unimaginable and outside anything known before. This is leading us into a new and different direction, away from the monolithic systems of the past.

We saw this effect first in the web, and we can see mi-grating to the enterprise today under the guise of SOA. The destruction of boundaries between departments is having a dramatic effect on our businesses as the old concepts of inside and outside no longer make sense.

Previously we were forced to restrict ourselves to what was happening within the walls of our organisation. Si-loed approaches to enterprise IT created an environment where we could only see the data we generated, we could only influence the business decisions and proc-esses we own, and we were responsible for only those business activities within our own organization.

The boundaryless environment is fundamentally chang-ing these assumptions and our zones of visibility, influ-ence and responsibility are changing. We have visibility across the entire value chain, as we share what was once private data with a customers, partners and suppliers. We can now negotiate and collaborate, in real time, with our partners and suppliers (and even their partners, suppliers, and so on further across the value chain) to influence their actions. At the same time our sphere of responsibil-ity is shrinking, as we pass more and more of our busi-ness directly into the hands of our partners and custom-ers.

SOA enables us to create a boundaryless environment, eroding walls between departments and organisations and bringing the exception rich, differentiating activities into sharp focus. This provides a wealth of opportunities to create new value propositions for companies by automating and then optimising of these differentiating activities, with the potential to create a step change in the market—a sustained competitive advantage.

The advantage that IT can provide has moved from indi-vidual applications, and now depends on successfully combining a suite of applications to deliver more effi-cient end-to-end processes. A company’s ability to inno-vate and build a competitive advantage rests on its abil-ity to support and optimize these differentiating excep-tion handling activities; with the significant intellectual investment this requires creating a barrier to competi-tion.

The challenge: scaleable computing

The shift in focus away from applications, and to the exception rich business activities that live between them, is impossible to support with an application centric ap-proach. Applications force us to treat all business proc-esses and activities as equal, spreading our investment evenly.

The next step-change in IT requires us to find ways to support the activities that live between existing applica-tions. Conventional approaches have reached their limits and we need to do something different if we want to continue to move forward.

Consider a financial adequacy solution for a major bank. Today, teams of analysts slave over spreadsheets to cre-ate what are effectively intelligent guesses as to the banks’ overnight financial position and obligations. SOA makes it possible to look across the boundaryless envi-ronment, integrating company, partner and market data in support of a more accurate and timely financial ade-quacy calculation.

A second example is the automation and optimization of the end-to-end supply chain. Existing supply chain solu-tions are strongly siloed into planning and event man-agement solutions, with little support for business excep-tion handling. SOA allows us to pull data from planning and event management and from departments and part-ners across the boundaryless environment to provide workers with the right data, at the most appropriate time, to facilitate the best decision and resolve the problem in an efficient manner.

Despite recent advances, key differentiating exception handling activities remain firmly in the domain of em-ployees and the extended teams they function within. The current technology stack has succeeded in removing a great deal of make-work from our employees’ day-to-day lives, but automating the complex, trial-and-error processes that we rely on them for is beyond the capa-bilities of our existing technology stack.

We, people, are very flexible decision makers, capable of finding creative solutions in uncertain environments, but we do have limitations. Our relatively slow “thinking speed” (compared to computers) and the low number of options we can consider at once (typically somewhere between four and six) restricts us to creating solutions that are acceptable in the local context. We find a solu-tion that works for us, rather than the best possible solu-tion.

Human involvement—the requirement to put employees inside transaction streams—is the new limiting factor. Our limited capacity to processes data and consider op-tions is prevents us from capitalising on the opportuni-ties unlocked by the boundaryless environment.

SOA has enabled us to deliver a consistent level of automation across the business and drive down costs. However, it hasn’t provided us with the tools to attack more complex problems, to automate the business ex-ception resolution and complex decisioning that we cur-rently rely on people for.

Page 3

Page 4: The problems we're facing

If we want to deliver the step change promised then we need to remove employees from opera-tional roles within the transaction stream, where simple human limitations are holding us back. We need to automate these key differenti-ating activities. Automation allows us to apply the much greater scale of information technol-ogy—it's ability to consider vastly more op-tions, rules, and interactions, and look across the boundaryless environment—to create glob-ally optional solutions. Employees move from an operational role, to a supervisory and opti-misation role where they can leverage their creativity to enrich these differentiating activi-ties and create a sustained competitive advan-tage.

This is beyond the capability of conventional technologies. The programming languages, development approaches, and rules and busi-ness process engines we currently use in the enterprise are mature products with well under-stood limitations. Aside from theoretical argu-ments, the simple fact that they haven't been used to support these differentiating activities can be seen as evidence that they is beyond their capabilities. The potential benefits of, for example, automating financial adequacy analy-sis and calculations would have resulted in the development a financial adequacy solution it were pos-sible.

While the current technology stack places us at the lead-ing edge of the step change, delivering the change will require something new.

TechnologyWe must answer a difficult question: what technologies, tools and techniques can we use to capture and automate these differentiating, exception handling business activi-ties?

Human flexibility is part of our ability to tolerate con-flicting goals and rules while working toward a solution. For example, work to capture and automate government legislation4 with rules engines quickly found contradic-tions (otherwise known as loopholes) in the legislation’s business rules. While the government was sure of the legislations intended effect, conflicts and inconsistencies within the act meant that no one, not even the people responsible for drafting the act, could be sure of the ac-tual effect. However, government found the act workable as the people responsible for interpreting it were per-fectly capable of working through the contradictions and inconsistencies as they sought solutions that met the legislation’s original intent.

Conventional enterprise technologies—such as Java or C#—represent bottom up technical languages. These languages are designed to help us craft a set of instruc-tions for computer hardware. This makes capturing

complex business logic difficult, if not impossible, as we’re forced to map the trial-and-error processes hu-mans use to navigate complex and inconsistent environ-ments into computer instructions. Capturing the incon-sistencies and conflicts, and the tribal knowledge, is a combinational problem beyond the capabilities of a tra-ditional programming language. There are simply too many options, alternatives and exceptions to consider.

More recent technologies—such as rules and business process engines, and domain specific languages—repre-sent an attempt to create business friendly languages which simplify this knowledge capture process. These approaches have seen some success, and they do have a much more business-like appearance, but both are still firmly founded in the technical traditions. The conflicts and inconsistencies inherent in complex business logic, the business logic at the heart of a company’s ability to differentiate, are still treated as exceptions rather than as part of the course.

Technologies based on a bottom-up approach cannot successfully deal with these inconsistencies, making them unsuitable as tools for capturing and automating the differentiating activities that are the key to unlocking the potential of the boundaryless environment.

An alternative approach

A more fruitful approach is to work top down: working from the business to technology. Businesses rely on net-works of people to plan and manage their operation. We should try to understand how people deliberate and col-laborate, how they work through the conflicts and incon-sistencies, and use that as the basis for our technology.

Page 4

4 Serot and Kowalski's work on the British Nationality Act in the early eighties. http://doi.acm.org/10.1145/5689.5920

Page 5: The problems we're facing

We also need to integrate failure—the error in trial-and-error central to our flexibility as decision makers—into the core of our approaches.

A case in point is the development of NASAs develop-ment to support fault-finding in the space shuttle. NASA started a project in the 80s to automate a manual fault finding process5. The process relied on an engineer and shelf full of binders at mission control, but NASA thought that it would be more efficient to reify the knowledge in a rules engine, enabling astronauts to carry it with them on the shuttle. The project encountered problems when the rules engine failed, incapable of sup-porting the non-linear, exception rich, fault finding proc-esses. Searching for inspiration they found Michael Bratman's Belief, Desire, Intention (BDI) theory of hu-man action6. Using this basic understanding of how a human would solve the fault finding problem, pulled from outside computer science, the team created a “BDI agent tool” called PRS7 (the Procedural Reasoning Sys-tem). PRS provides a scenario-based approach to captur-ing and automating knowledge that mirrors human de-liberation, breaking the problem up to create a decision making process that mimics the way a person would navigate their way through the conflicts. Using this tool the team was able to successfully deliver a solution that was beyond the reach of conventional programming lan-guages or rules engines.

Conventional enterprise technologies have an ideal view of the world, where all business logic can be represented with clean business processes or idealized predicate rules. This breaks down in the real world where rules conflict and the exceptions in a business processes are often more important than the body of the business proc-esses itself. Automating the differentiating activities that are holding us back requires us to find new technologies and tools that allow us to mimic human deliberations which are the soul of a business.

The tipping point

We’re rapidly arriving at a tipping point as our existing technology platform matures, bring key differentiating activities into focus where we can apply new ideas and concepts to capture and automate their complex deci-sioning processes while, at the same time, we’re faced with the reality that between 30% to 50% of the people that have been employed by industry in the last 30 years are retiring in the next three to five.

A range of suitable technologies are already available, from BDI agents mentioned above, through machine learning, neural networks to multi-agent systems which allow us to mimic how people communicate. These technologies allow us to leverage ideas from economics and complex systems theory to create applications that

exhibit the richness and capability of markets and social groups.

Many of these technologies have long individual track records, where they have successfully solved problems beyond the reach of more conventional approaches. However, successful proof-of-concepts and prototypes have faced difficulties when moved to production. The problem has been that each technology lives apart from the conventional enterprise technology stack. Past at-tempts at commercialization have tried to push these technologies into the enterprise, solutions looking for a problem and failing to meet enterprise requirements for performance, reliability and maintainability.

A more pragmatic approach is to identify the problem that need to be solved and then pull, rather than push, in the technologies required to solve the problem. Initially working with small, isolated problems, we will work our way toward more ambitious and pervasive challenges. This adoption process can be broken into three general phases8: role-based automation, the creation of teams and developing markets.

Role-based automation

First we capture the business process within a single differentiating activity, moving from the locally accept-able solutions generate by employees to locally optimal solutions that automation enables. Automation results in increased speed and accuracy, delivering an immediate benefit to the business by lowering operational costs, increasing throughput and lowering the number of er-rors.

Employees move from an operational role, where they are integrated into the transaction stream, to supervisory role where they are responsible for oversight and opti-mization of key exception management business logic.

Role-based automation is widely applicable, with com-mon examples across a diverse range of industry sectors for financial service, supply chain and logistics, manu-facturing through to the public sector. Activities such as real-time fraud detection, supply chain exception man-agement, automated loan approvals and policy manage-

Page 5

5 http://www.ai.sri.com/~prs/rcs.html

6 Michael Bratman, Intentions, Plans and Practical Reason, CSLI Publications, ISBN: 1-57586-192-5

7 http://www.ai.sri.com/~prs/

8 First laid out in the article Moving Beyond Composite Applications to the Next Generation of Application Development: Automating Exception-Rich Business Processes by Peter

Evans-Greenwood and Mark Stason, published in the May/June 2006 issue of Business Integration Journal.

Page 6: The problems we're facing

ment within the public sector are similar to the scenario-based approach to problem solving that NASA saw when developing the fault-finding solution for the space shuttle. Fraud detection is generally a process of testing transactions to see if they are part of any known fraud scenarios. Supply chain management and loan approvals both use a scenario based approach to problem solving, starting from the optimal solution (“Can I replace the product from excess inventory?” “Does the customer qualify for the home loan without any supporting docu-mentation?”). Policy management requires us to use a scenario based approach resolving inconsistencies in legislation, using the context of the problem to deter-mine which scenario to use in navigating the legislation

While SOA enables us to break down barriers between applications, departments and organisations to better support users and create a more efficient work environ-ment, role-based automation allows us to tackle more complex problems and augment users, allowing them to support more sophisticated business logic.

The creation of teams

Next we focus on capturing the collaboration and nego-tiation between distinct roles to create virtual teams within our organisation, allowing our automated roles to look across its social network to negotiate with peers in the creation globally optimal solutions within an enter-prise. This builds on the role-based automation from the previous phase, expanding our view beyond that of a single role to include the role's social content—its team.

Where role-based automation allowed us to capture and then optimize the business logic within a role (treating other roles as simple services on the network), a team allows us to consider a group of roles as a whole and capture the negotiations between teams members as part of the solution. This allows a role to influence other roles as they reach a decision, simply delegating work or even agreeing for both roles to change their current plans in order to resolve an issue.

Developing markets

Finally we can network the teams and roles together, leveraging market-based approaches to form automated networks of roles, departments or even separate compa-nies that dynamically share and manage resources. This allows us to expand our view of a role's social network

beyond the team to support boundaryless business proc-esses. Rather than requiring a single, central, planning authority, we can leverage ideas from economics to cre-ate flexible markets where roles gather to manage re-sources.

The boundaryless enter-priseHow will breaking down barriers between departments and organizations affect an enterprise? What will the impact be on the role of applications?

We can expect the emergence of the boundaryless envi-ronment to trigger a dramatic change in the organiza-tions we work for. The change in how we engage IT—a shift from managing assets to capabilities—must have a direct effect on ITs role in business.

IT as a facilitator

Many IT departments were created in a time when en-gaging IT was a major engineering undertaking. Deliver-ing large applications required large departments led by teams of skilled technologists. Companies created cen-tralized technology groups responsible for capturing business requirements, defining applications, setting standards and ensuring quality of delivered projects, and technology was seen as something apart from the rest of the business. While this was appropriate for an application-centric enterprise, we need a new model for the boundaryless enterprise.

The economics of IT has changed as it has become more diffuse. IT has spread out from the centralised depart-ment of old to find a place in every department. It’s common for the CFO to be responsible for desktop sup-port, with departments deploying their own portals, web sites and collaboration tools.

More recently Web 2.0 technologies such as wikis and blogs have been sprouting across the enterprise. Tech-nology is no longer seen as something special, some-thing to be maintained by a separate department, but as a simple tool to be purchased where and as needed. The current generation of enterprise applications are the last technologies to live in the world of centralized IT, as the increasing use of partnerships and service-based ap-

Page 6

Page 7: The problems we're facing

proaches is gradually reducing the need for a separate technology focused department.

We can see a point in five to ten years time when cen-tralised IT departments will have ceased to exist9. Soft-ware development will have moved to the fringes of the organisation, where development teams are focused on delivering and maintaining key differentiating services that are two valuable to move outside the business; how-ever the teams will be using a mixture of business fo-cused tools to deliver business mashups or support the development of complex business logic, rather than Java or C# which are common today. Software architecture to continue as a function, but it will have moved from a standardization and engineering development role. Its new role will be as a facilitator working with the busi-ness to understand the problem at hand, determine the required technology footprint, and then co-ordinate the planning and delivery of the heterogeneous capabilities that comprise the boundaryless enterprise.

Master planning the boundaryless en-terprise

We need a new organizing principle to manage the boundaryless enterprise. Or existing practices are based on managing major assets, and are no longer suitable.

The Paris Guide to IT Architecture (reference) points the way by establishing three guiding principles:

• Work with the business landscape, moving beyond our own boundaries to consider the assets, capabilities and constraints of the organizations we interact with.

• Establish capable and flexible infrastructure

• Actively manage the process—a commitment to ongo-ing monitoring and improvement

Together theses principles provide us with a framework to define a new model for planning and delivering the next generation of IT support. However, as with cities, different planning models suit different circumstances. Paris exemplifies one town planning model, as do Mil-ton Keynes, Levittown, Chandigarh, and Canberra. The tension in all master-planning is that which exists be-tween planners imposing structure (zoning) while allow-ing inhabitants (us) the freedom to change and imprint our surroundings. Too much master-planning stifles ex-pression and diversity, too little results in shanty-towns. This aspect of the analogy seems to hold for SOA and the boundaryless environment. Enterprise architecture provides infrastructure and standard connectivity, the business uses this platform to craft individual services and to string services together.

Working from the three principles we can create a sim-ple hub and spoke planning model that aligns well with business desire to innovate at the customer coal-face while, at the same time, ensuring compliance and effi-

ciency in common operations10. The hub represents our core enterprise data. Spokes are the major services radi-ating from the hub—the core applications that form the backbone of an enterprise. The rim is the coal face where we need to deliver business functionality to cli-ents, employees and partners.

The wheel provides us with a simple model for our ef-forts in city planning, allowing us to strike an appropri-ate balance between innovation and chaos. Core enter-prise processes, the spokes, are relatively static and con-sistent across the enterprise. We rely on them for com-pliance and to ensure the smooth operation of our or-ganization. Innovation is relegated to the rim, where we interact with clients, partners and even competitors. The framework provided by the hub and spokes is allows us to manage innovation by providing us with a clear un-derstanding of how much flexibility we can allow before cores processes begin to bend and break.

Making the business case

The boundaryless environment is making asset based return on investment (ROI) calculations that we’ve re-lied on to date irrelevant, as the next generation of solu-tions will leverage a range of technologies, combining them with non-technological elements to deliver focused capabilities.

From asset to capability

Solutions will have more similarities with the business mash-ups11 that are currently emerging, rather than the applications of old. Our financial adequacy solution from before is a good example, as it pulls in a range of existing services, new and old as well as from in and outside an organisation, combining them with new tech-nologies to deliver a tightly focused capability. The cost of the final solution cannot be determined by simply tallying up the capital investment required to deliver the mash-up since its functionality relies on a range of tech-nology as well as functionality provided by other serv-ices, some of which are sourced externally.

Page 7

9 The Economist, Great expectations: The changing role of IT in business, The Economist Intelligent Unit

10 The use of a spoked-wheel metaphore is explored more extensivly by Steve G. Jones in his book Enterprise SOA Adoption Strategies, InfoQ, ISBN: 978-1-84728-398-6.

11 Andy Mulholland, Chris S. Thomas, Paul Kurchina, Mashup Corporations: The End of Business As Usual, Evolved Technologist Press, ISBN: 0978921801

Page 8: The problems we're facing

We can see this as part of a broader trend to move from asset to activity based accounting, part of the broader trend to activity-based models of business we see in the business community. Rather than focusing on the assets owned by a business, we’re starting to focus on the product or services it provides and the capabilities re-quired too support them. The groundwork for this ap-proach was laid out in the Strategy Maps by Robert S. Kaplan and David P. Norton, which expanded their work on measuring the efficiency of business from The Bal-anced Scorecard.

In this new model cost, capability, risk and solution footprint form the basis of establishing a business case. A solution is measured on its ability to deliver a clearly defined capability that adds value to one (or more) busi-ness activities. The capabilities delivered will address real and pressing needs. Can I dramatically reduce the amount of stock in my supply chain? Can I reduce my capital reserves while maintaining financial adequacy? How do I create an effective 4PL market to deal with the last mile?

The tight connection between this approach and how we measure and understand the business provides a basis for prioritizing initiatives, something many companies struggle with. Is it more important for customer interac-tion processes to be standardized across all business units; or for vendor processes to be? Is it more important to provide flexibility and room for innovation in the product development, or to drive the creation of products that are consistent, uniform and low cost to create and distribute? None of these are easy questions to answer, but by explicitly connection the investment we are con-sidering we manage to greatly simplify the challenge.

Applications, but not as we know them

But if we are beginning to focus on capabilities, how will we justify the sort of major investment required to develop, deliver and maintain a complex enterprise ap-plication? How can we continue to support the core en-terprise functions that are the bread and butter of today’s current enterprise application?

Applications are capable and efficient at supporting the business functionality within their scope. It’s not un-common for older businesses to have twenty or thirty year old applications that are still successfully support-ing the business, processing huge volumes of transac-tions beyond their original design specifications. Their problem is that they support yesterday’s business re-quirements, not today’s, and updating the 10% of busi-ness logic that has changed can often necessitate com-pletely redeveloping the application.

SOA has step into the gap, as the more granular ap-proach to IT it enable allows us to break applications into their logical components and consider each compo-nent separately. The smaller size of these components

naturally lends itself to a capability based approach, as each component is focused on supporting a single busi-ness activity such as managing customer data or pricing transactions.

The undifferentiated nature of many of these compo-nents sees them on the path to becoming services a com-pany uses, rather than an asset is owns. If there is little strategic value in the functionality, then most business would prefer to account for the functionality per use rather than as a capital investment.

For example, existing billing systems are heavily com-moditized, and a billing solution that allows clients to retain control of core pricing functionality (the only dif-ferentiating activity in this part of the value chain) could be used as the basis of a "billing and collections serv-ice". Client companies submit transactions (with pre-computed prices) to the service, which then generates the invoice, sends it (email or paper), collects the cash, and taking a percentage of the top before sending the cash to the client.

We’re already starting to see the development of some of these services. Amazon is leading the charge by opening up their capabilities and offering them on the open mar-ket under Amazon Services12. Starting with simple shopping functionality, they are now branching out to offer more traditional enterprise business functions such as storage13 and, now, fulfilment14.

Applications as major IT assets will cease to exist, with their functionality provided by a suite of services. Capa-bilities central to a companies ability to differenti-ate—the 10% of business functionality that will make a difference—will be supported in-house where a com-pany have tightly control over its creation and delivery. Less important capabilities, many of which are currently support by in house enterprise applications, will be moved outside the enterprise.

ConclusionsInformation technology has come a long way since its birth over forty years ago and technology has matured to the point that IT is no longer something “different” in the business. The application centric approach to IT that we’ve successfully used to date is reaching the end of its useful life, and new need to look for new approaches that can take us forward into the future. We've finished the big effort: applications are available from multiple vendors to support the majority of a business’ context functionality. The law of diminishing returns is taking effect, and owning or creating new IT asset today is not going to deliver a sustained competitive advantage.

The recent emergence of service-oriented architecture is allowing us to reinvent enterprise IT, destroying the bar-

Page 8

12 http://www.amazonservices.com/

13 http://www.amazon.com/s3/

14 http://www.amazonservices.com/fulfillment/

Page 9: The problems we're facing

riers between departments and organisations to create a boundaryless environment. This is having an immediate impact on our business’s areas of visibility, influence and responsibility. Some forward thinking companies are leveraging this to reorganise their businesses and deliver dramatic increases in efficiency.

This has created a boundaryless environment, providing a wealth of data and interactions whose latent capability is capable of delivering a step change to the business. However, these capabilities are beyond the reach of the mature applications and integration technologies we currently use, and continue to rely on people to support the non-linear, trial-and-error deliberation processes that are becoming core to a company’s ability to differenti-ate. This has created a new limitation: the people we continue to rely on for the complex deliberation and ex-ception management that is at the heart of a company’s ability to differentiate. To move forward we need to move up the value chain and attack more complex prob-lems. If we can capture and automate key, high value, deliberations and negotiations then a company can use this to build a sustained competitive advantage. Deliver-ing solutions that capture, automate and optimise these deliberations and negotiations will enable a company to differentiate. The investment required to develop this complex business logic provides the barrier to competi-tion.

Solving these complex problems is beyond the capabili-ties of the current generation of technology. Remove this limitation and we unlock the potential to deliver a step-change to the business. Delivering this step change re-quires us to find new tools and techniques capable of capturing and automating the complex, non-linear delib-erations that we, as people, find easy, and to change the way we engage with IT.

Page 9