Read article and learn about:
- The dangers of a decoupled IT strategy
- Why IT and business strategies must be aligned
- IT architecture management
- Advantages of a holistic IT strategy
About the authors:
Stephen Murer, Senior Advisor Credit Suisse.
Bruno Bonati, formerly Head of IT at Credit Suisse, now independent consultant based in Zug.
Frank J. Furrer, Specialist in application architecture and IT security, Independent consultant based in Zurich.
The phrase ‘silo effect’ refers to a lack of communication and common goals between business units, which prevents companies from efficiently responding to changing demands and opportunities. To enable organizations to utilise the synergies and flexibility created by pooling IT and business resources as a competitive advantage, the authors make a case for managed evolution as a management approach to transform a decoupled IT landscape to a business-empowering information system.
As in other large and complex institutions, the development of information systems in global investment management companies often depends strongly on the organizational set-up of the business. IT budgets are owned by the individual business units, and any change in the organization also leads to a corresponding change of direction in systems development, with IT investments sometimes driven purely by the local needs of the individual business unit.
As organizations are often more flexible than their underlying, tightly integrated information systems, the state outlined above can lead to a situation where multiple units implement changes simultaneously without strategic coordination. The changes are driven purely by their business requirements on the same information system and as nobody owns the system, nobody feels accountable for the overall architecture.
Unhealthy symptoms of the silo effect
As a consequence of the traditional silo approach just described, the efficiency in operating and evolving the information system decreases to a point where the system becomes unstable or requirements can only be implemented at disproportionate cost and risk. Over the years, the investment management organization will find that this situation leads to an IT landscape, which is either highly fragmented in silos or tightly coupled and highly complex.1 A situation arises where IT is hindering rather than empowering the business.
If a corrective attempt is made to bring these silos together in an integrated information system, however, a number of issues arise. First, the borders of these silos are cut in a way that makes integration with other silos very expensive, if not impossible. Second, whereas the silos are highly redundant, the functions and the data are not. And third, the silos build on independent architecture stacks, additionally complicating cross-silo integration. Often ‘horizontalization’ strategies, trying to break down the silos fail, because the integration challenge is too overwhelming.
Corrective IT strategy must balance three objectives
A good remedial IT strategy should ensure a systematic development of a future-proof information system by balancing three objectives:
- all business requirements should be met within reasonable time-to-market and cost;
- in order to deliver on the first objective, the IT strategy must maintain agility of the system with regard to implementing new requirements;
- to maintain agility and cost effectiveness, the strategy must contain complexity and continuously strive to remove redundant and outdated functionality and technology.2
A key prerequisite for achieving these strategic objectives is a good alignment of business and IT. In our view, the right strategy to fulfil these goals and to avoid the potentially fatal trend in the evolution of an information system towards unmanageable legacy is the concept of managed evolution.3
The key idea of managed evolution is to steer the evolution of an enterprise’s information system, such that the efficiency in developing and operating the system is preserved or even increased. This is accomplished by striking a balance between investments in efficiency improvement and investments in business functionality development.
Architecture management
The concept of managed evolution consists of different elements. One important element is architecture management. A target architecture needs to be defined and system evolution must be driven systematically towards the target architecture. Disparities with the target state need to be identified and used to steer architecture-driven investments. There must be a clear roadmap of how to develop the IT landscape from silos to a future-proof agile information system. The key to architecture management of large information systems is to break them down into a hierarchy of manageable pieces, or so-called domains.
Managed evolution is about developing and implementing a holistic IT strategy. This requires much deeper cooperation and more formal coordination in strategic business-IT alignment.
In a siloed environment, it makes sense to define the domains more horizontally in order to emphasise common functionality among the silos. Often, architects use two different domain models, one for infrastructure and one for the application landscape. The infrastructure domain model is mainly structured according to the market in IT infrastructure, with domains, such as databases, operating systems, and end-user platform.
The application domain model often groups domains according to common data and functionality. The domain structure also serves as a structure to federate the architecture organization with a small central group for coordination and decentralized domain architects allocated with the organizations in charge of engineering or development in the respective domain. Good architecture management is based on successfully dividing the very large information system into manageable parts.
Integration architecture as vital element
One particularly vital element in successfully pursuing managed evolution is integration architecture. This defines the appropriate coupling and decoupling of the system’s components. At the heart of integration architecture is the interface management process that manages evolving interface contracts between producers and consumers of a service. Accepting that a very large information system cannot be replaced as a whole requires a component replacement strategy. For this, it is paramount that the interfaces are stable or at least evolve in defined, predictable ways.
One area that is often neglected is the infrastructure technology portfolio underneath a very large information system. It is essential for a long-lived system to continuously adapt to modern technology. If not managed carefully, this leads to a very broad, expensive and often partially outdated technology portfolio.
This is a typical problem of silo-based organizations, where each silo follows an independent infrastructure strategy. One difficulty with technology portfolio management is that applications depend in many ways on the underlying technology. So if the abstraction layer between infrastructure and applications is not managed with care, changes in the underlying technology portfolio can become extremely expensive.
In the case of a silo structure, bringing the silos together may be harder than necessary due to incompatible infrastructure. In this context, the approach of defining platforms that combine a technology stack with well-defined processes and services for running and developing applications is a very powerful abstraction concept.
Taking a balanced approach
A strong relationship exists between business-IT alignment and managed evolution. Managed evolution requires a method of business-IT alignment that ensures balanced investments in new business functionality and in IT efficiency at one and the same time.
It is current business-IT alignment practice in most organizations to ensure that business requirements are defined by the business in a systematic and efficient way, that IT departments understand and accept these requirements, and that they provide adequate IT solutions fulfilling the requirements. Focusing entirely on business requirements results in a project portfolio that is driven purely by business requirements and little or no consideration for efficiency and agility issues.
Business requirements in such a portfolio are implemented in the system where it takes least effort, easily available resources and in the shortest time from the point of view of the individual project instead of where change arises from a conceptual integrity. As more and more functionality is added without regard to architectural principles, the system becomes more and more complex. This is common in a decoupled IT strategy.
This opportunistic approach drives the system into the complexity trap (see Figure 1). Once an information system is caught in the complexity trap, it is extremely costly to evolve and becomes in many cases unstable and unreliable.
Developing and implementing a holistic it strategy
Often information systems grow opportunistically over years, while the organisation keeps changing. Over time, the architectural system boundaries no longer fit the business organisation. This is to be expected, because organisational change is much easier than structural system change. This leads to a situation where business and IT are no longer aligned to the architectural scope of the system. In a very clear sense, the system becomes decoupled.
A key prerequisite for achieving these strategic objectives is a good alignment of business and IT.
As a consequence, this can lead to uncoordinated silos, with massive redundancy and divergence in functionality, interfaces and data. Such silos are overly complex, difficult to integrate and, therefore, slow down the strategic development of an information system. Consequently, the first step towards a strategic development of the IT platform is to shape business-IT alignment based on the scope of the system.
Another important aspect is the balance of power between business and IT. The common practice in many companies is to develop individual IT strategies in each business unit with little coordination at the corporate level. Strategic responsibility is decentralised. In this model, business unit heads decide on their IT budgets, within the constraints of the overall business unit budget.
Managed evolution is about developing and implementing a holistic IT strategy. This requires much deeper cooperation and more formal coordination in strategic business-IT alignment. Collaboration in joint committees fosters the understanding of IT managers for business requirements and of business managers for IT needs.
Of crucial importance is that both business and IT are represented in decision-making in a well-balanced way. This reflects the key idea of managed evolution – to balance investments into business value, as well as into IT agility. This is implemented by matching roles for business and IT participants at all levels and balanced representation of IT and business in all decision-making bodies.
IT organisation embedded in the business
In order to support business-IT alignment, the IT organisation needs to be properly embedded in the investment management business it serves. It is very hard to manage the complex processes and dependencies necessary for managed evolution, if the IT organisation working on a single system is distributed among many business lines.
Large organisations and large systems only move in a certain direction if progress is measured and, in the case of insufficient progress, corrective action is taken. The simplest way to measure architectural progress is to gauge the quality of the processes by measuring whether all standards are kept up to date, projects are properly reviewed and review results are followed through.
The next level is to measure the progress towards the target architecture, by looking at adoption rates of technology standards or reuse factors of interfaces. The highest level is to measure whether the system makes progress towards better agility or not. There is a clear correlation between improved agility and business value of the system.
The case is made for a more balanced approach that takes architectural aspects and the long-term viability of the underlying information system into account. Ideally, this governance is based on a domain model that reflects the architectural structure of the information system, on an IT organisation that spans the whole system scope under one lead, on a strong architectural leadership, on a balanced governance model bringing all stakeholders together around one table, and finally on a culture where stakeholders understand and appreciate how they strategically depend on one another.
One proven way to move from silos
A highly developed, efficient and cohesive information system is an indispensable enabler of many of the daily operations run by investment management organisations, and must evolve in order to support the business it serves in a competitive and changing environment. Information system evolution should be guided by a suitable IT strategy, which has to cope with the business strategy and a number of influencing factors.
A sound alignment of business and IT strategies is the primary ingredient for a successfully integrated IT approach applied by investment management organisations. It must be more than just collecting and prioritising business requirements within the available budget in each business line if the enterprise wants to avoid ending up with siloed systems that are hard – if not impossible – to change.
This article has attempted to establish one proven way to meet the challenges of the development over time of mission-critical, very large information systems that include significant legacy parts. The recommended strategy is managed evolution. Managed evolution leads to a continuous improvement and sustainable lifecycle evolution of information systems. It also improves agility in the form of reduced time-to-market and lower cost for adapting to new requirements.
1. Tom Graves: Bridging the Silos – Enterprise Architecture for IT Architects. Tetradian Books, Colchester, UK, 2008. ISBN 978-1-906681-02-9.
2. Jeanne W. Ross, Peter Weill, David C. Robertson: Enterprise Architecture as Strategy – Creating a Foundation for Business Execution. Harvard Business School Publishing, Boston, USA, 2006. ISBN 978-1-59139-839-4.
3. Stephan Murer, Bruno Bonati, Frank J. Furrer: Managed Evolution – A Strategy for Very Large Information Systems. Springer Verlag, Berlin & Heidelberg, Germany, 2011. ISBN 978-3-642-01632-5.