Book Review - Enterprise Architecture as Strategy

Enterprise Architecture as Strategy by Jeanne W. Ross, Peter Weill, David C. Robertson.

This book attempts to describe how and why enterprise architecture matters to organisations. There are several questions that this book attempts to address:

1. What is enterprise architecture?
2. How can organisations justify the investment in developing an enterprise architecture?
3. What are the business outcomes from doing enterprise architecture well?
4. In what terms should an overall enterprise architecture be defined?
5. How can an organisation realise its architectural vision - and in doing so support the realisation of its business vision?
6. How can the conflict which often occurs between short-term business imperatives and long-term business strategy be managed in terms of its impact on architecture?

The first thing which the book recommends doing when developing an enterprise architecture is to define the organisation's operating model. This represents the organisation's commitment to a way of doing business. The operating model is defined in terms of two dimensions; business process standardisation and systems integration. The book argues that these two dimensions drive how enterprise architecture should be undertaken at an organisation. It then goes on to classify the various permutations possible with these dimensions into four operating models:

Diversification - Low Standardisation, Low Integration

Coordination - Low Standardisation, High Integration

Replication - High Standardisation, Low Integration

Unification - High Standardisation, High Integration

An organisation's position on the quadrant should be driven by a) the extent to which data sharing between business units is necessary and b) the extent to which commonality of business processes across business units is required. The book goes on to give examples of organisations which have applied each of the four operating models. It also describes the impact of each approach on IT architecture strategy and the implications of each with regards to how the organisations can grow. For example an organisation employing a replication or unification scenario which is acquiring another organisation will need to rip and replace the newly acquired entities' existing systems in order to fit the acquiring organisation's operating model. Conversely an organisation which employs a diversification approach can more easily digest newly acquired organisations - but it clearly will struggle with any data sharing requirements between business units. It is important to note that an organisation can simultaneously employ different operating models at different levels in its organisation. 

The identification of an organisation's operating model is not purely an IT responsibility. It is as least as much (if not more) a senior business leadership decision. 

The book's authors see much of EA as being focused on business objectives and outcomes rather than being a purely technical discipline. This makes a lot of sense especially, as the book asserts, since senior management understanding and sponsorship of EA is critical to its success in an organisation. To that end the book recommends the creation of a a Core Enterprise Architecture Diagram. This is one-page diagram intended to illustrate the key elements of an organisation's IT architecture to all interested parties in an organisation. The four most commonly shown elements in the diagram are:

1. Core Business Processes - company-wide processes

2. Shared Data Driving Core Processes - data shared between processes 

3. Key Linking and Automation Technologies - e.g. middleware and ERPs

4. Key Customers

The book then helpfully provides real-world example Core Enterprise Architecture diagrams for organisations with differing operating models. The book argues that organisations which in its words 'digitise the core to make it predictable, reusable and reliable' will be successful. Unfortunately the book does not define what digitising the core actually means in practice - beyond the obvious of building or (especially) buying systems to do so. Nevertheless the core diagram approach could well be useful as a mechanism for helping organisations to decide where to focus their IT investment. 

The book then goes on to define and explore various stages of enterprise architecture maturity. It defines these as:

1. Business Silos Architecture

2. Standardised Technology Architecture

3. Optimised Core Architecture

4. Business Modularity Architecture

As an organisation moves through these stages it will spend less and less on local, business unit-specific IT and spend more on developing a common IT architecture which still preserves the ability (where necessary) for diversification. The book discusses the maturity journeys of various organisations. It also argues that it is difficult for any organisation to jump multiple maturity levels in one leap as (when attempting to do so) the rate of change is usually too much for an organisation to bear. This section of the book is very successful in clearly defining the maturity stages, their individual characteristics and also the activities and IT funding priorities which should be associated with each stage. It also describes how an organisation's architecture practice can and should mature along with the architecture itself. 

Time and again the book asserts that in order to increase agility an organisation needs to "digitise its core processes". Unfortunately the book does not define what this means in any meaningful way. The inference which most readers are likely to draw from the phrase is that successful companies implement Enterprise Resource Planning (ERP) software. 

One of the key outcomes for organisations which undertake this journey is that as the flexibility of local business units starts to fall the overall global flexibility begins to rise. The outcome for the most architecturally-mature organisations is that the benefits of their overall gains in global flexibility outweighs the loss of flexibility at a local level.

One of the key challenges faced by any enterprise architecture team is how to define KPIs which demonstrate that the investment made in made in order to create and maintain an architecture practice is worthwhile. The book does provide some helpful material for these efforts by describing the various benefits which should accrue to organisations with a well functioning enterprise architecture practice. These include reduced IT costs, increased IT responsiveness, improved risk management, increased management satisfaction and enhanced strategic business outcomes. However, actual examples of enterprise architecture-relevant KPIs are not present. 

The book does an excellent job of describing how architecture should align and interlock with both senior IT folks and also project teams in order to ensure that both long-term architecture goals and short-term project needs are met. It does this by describing how the architecture function should be involved in the governance and guidance of projects throughout the entire project life-cycle.

Another key challenge for enterprise architecture efforts is how to turn what is often seen as an ivory tower-type of endeavour into one which has practical, useful effects throughout an organisation. Part of this is typically the question - who pays for the shiny new architecture when there is typically no specific budget to do so?

The book really does not adequately address this critical question. The only relevant example provided is that of Toyota Motor Marketing Europe. In this example where projects could not be persuaded to fund the additional costs sometimes required in order to meet architectural strategy the architecture function would pick the cost of the additional work. Given the way that architecture teams tend to get funded this does not seem to be an approach which could be widely applied. Additionally, the incentive for project teams to "do the right thing" is much reduced when they realise that someone else will pick up the tab if they don't choose to adhere to architectural standards. The book does suggest that business units which pay for the upfront infrastructure costs associated with the implementation of strategic items could subsequently recoup the wider costs involved via some sort of chargeback model. However this idea is not explored in sufficient depth. 

The book then goes on to outline three outsourcing models together with the implications of each. The models described are:

1. Strategic Partnership - responsibility for operational activities outsourced

2. Co-sourcing - project management and implementation outsourced

3. Transaction - narrowly-defined repeatable processes

Unfortunately these models are quite poorly defined by the book. Additionally the book is now rather out-of-date with regards to the commercial models now being offered by outsource vendors. The book does attempt to advise how and when to outsource with regards to the previously described architectural maturity model. However since the outsourcing models are themselves poorly defined this is of limited value. It would have been useful if the book had discussed the impact of the commercial models used during outsourcing on the achievement of architectural goals. For example, how can a move from capex to opex-based expenditure enable strategic projects to become more easily funded?

The book then moves on to describe the different growth opportunities enabled by each of the different operating models (diversification, coordination, replication and unification). It also describes how each copes with a merger or acquisition scenario. This is genuinely useful material and will enables architects to help business managers understand the likely impact of a given merger or acquisition in both cost and systems terms. 

Having previously defined four stages of architectural maturity (business silos, standardised technology, optimised core and business modularity) the book also proposes a fifth; dynamic venturing. This seems to involve businesses reaching a level of agility at which they can can plug-and-play entire businesses or business units. This notion is predicated (at least in part) on the assumption that software agents and standards for information interchange reach a level of maturity which allows such seamless integration (both business and IT) to happen. Given the rate at which both meaningful AI and standardised business software component models are improving this seems to be quite an optimistic position to adopt. 

The book closes with a chapter which first describes how the previous chapters have addressed the problems associated with organisations with low architectural maturity and how this can alleviated via a more mature approach. It then closes by offering some useful guidance on how to approach an initiative to plan and implement an enterprise architecture strategy. 

In summary the book is very successful in describing the practical types of operating models which can be applied when defining the high-level architectures and business approaches for global organisations. It is also strong on the maturity journey which organisations need to undertake in order to realise a vision involving maximised global organisational agility. Where it is weaker is in defining the practicalities involved - what does "digitising the core" actually mean, how does one do it?

The other major failing of the book is that it does not adequately address the tension between short-term project imperatives and long-term architecture vision - i.e. balancing the tactical with the strategic. 

The book contains much sound advice for enterprise architects and provides a good framework and starting point for them to engage with senior business stakeholders in order to begin or deepen the implementation of an enterprise architecture. 

On balance, the book is a highly worthwhile read for any enterprise architect who wishes to put in place a high-level strategy for enterprise architecture which will exist above the day-to-day imperatives of the role.