Digital enterprise architecture must start with the data. First published on 7wData, October 2019.

A digital enterprises is an enterprise whose business is based on new digital technology. To succeed, it must have the right technology architecture. Data is its lifeblood. That is where the architecture design must start.

Image by Gerd Altmann from PixabayThe book Designed for Digital by Ross, Beath and Mocker, published in September 2019, was eagerly awaited by the Enterprise Architecture community. Since the appearance of Enterprise Architecture As Strategy (2006), Jeanne Ross has built a big reputation through articles and presentations at EA events. Now she has combined with business school professors Cynthia M. Beath and Martin Mocker to publish the fruits of their research into adoption of digital technology by big companies. The result gives us a new way of looking at the business architecture of digital enterprises, but falls short when it comes to the supporting technology.

Enterprise Architecture in the Digital Revolution

Enterprise Architecture developed from Information Technology Architecture, the discipline used by teams of professionals to plan the IT estates of enterprises. The IT is there to support the business practices and strategy. Enterprise architects start from business considerations, to plan for the whole enterprise. They do not set the business strategy, but they work closely with the business leaders to ensure that the technology meets business needs.

New digital technology has disrupted and still is disrupting business practices and strategy. Cloud computing, social networks, mobile connectivity, big data, the Internet of Things, and artificial intelligence have enabled new business models.

Books such as The Four Steps to the Epiphany by Steve Blank and The Lean Startup by Eric Ries describe how business leaders have successfully created enterprises using the new technology. Their approach does not follow the traditional methods of enterprise architecture dating from the 1990s. It takes a new path.

Enterprise architects understood that this was happening, but at first it did not much affect their work. They held that the fundamental architecture principles and practices must still apply, regardless of changes in technology.

Designed for Digital may change that. It is the result of five years of research into adoption of digital technology by big companies. In the main, its subjects are not the prime movers that seized technology opportunities to grow suddenly from nothing; they are companies that were already household names, saw what was happening, and reacted. It looks at what they did, at what worked, and what did not work. It draws conclusions, and describes patterns for success.

Digital Transformation Reference Model

The book is an impressive piece of work. It presents a reference model for use in digital transformation, and discusses how to create a roadmap to change an enterprise to fit that model. It is illustrated with copious real-world examples, drawn from the authors' research. The reference model is simple and clear. Creating a roadmap is not so straightforward, but the book presents ideas on how to do it, again with real-world examples.

The reference model for digital transformation has five building blocks. They are organizational capabilities that help companies succeed digitally. Each contains people and processes, as well as technology. They are not a complete kit; you will need additional blocks to build any enterprise; they are the ones to include if you are building a digital enterprise, or digitalizing an existing enterprise. They are:

• Shared Customer Insights - information about customers, and a set of processes for developing it and disseminating it through the company, using the methods of The Lean Startup, and supported by technology.
• Operational Backbone - a coherent set of people, processes and technology systems that automates the core business processes and makes them visible to users and other system components. It supports end-to-end transaction processing, and provides a single source of truth for core data.
• Digital Platform - a repository of business, data, and infrastructure components used to configure digital offerings rapidly. The distinction between it and the operational backbone is rather like that of Gartner's bimodal IT, where there are two separate but coherent styles of work, one focused on predictability, the other on exploration.
• Accountability Framework - a management structure and governance regimen that empowers people while coordinating their individual efforts. It supports agile software development and devops, but these are not its essential characteristics. It is based on lifetime responsibility for system components, with clear visions and metrics.
• External Developer Platform - a repository of digital components open to external partners. Not every company needs an external developer platform. Establishing one enables a company to participate in wider business ecosystems, but it must be justified by business payback.

Technology-Based Companies Need Technology Architecture

The model is an excellent way of viewing the business architecture of a digital enterprise, but the book is less helpful when it comes to creating a technology architecture to support the business.

The five digital enterprise building blocks depend on each other, and other building blocks depend on them. For example, Shared Customer Insights and the Operational Backbone both need to know who the customers are. The development and operation of digital business offerings uses Shared Customer Insights and the Digital Platform in combination with the Accountability Framework. The Technology systems in different building blocks must work together, and must be able to exchange data.

Successful prime-mover digital enterprises generally have well-defined technology architectures. Netflix, for example, has an architecture that includes appliances that distribute video, streamed data logs of customer events, and microservices that support digital offerings. It can be mapped superficially to the building blocks of the model, but its real value to the business comes from a deeper level of technical detail.

APIs Are Not Enough

The book accepts the need for components to interwork, and says that they do so through Application Program Interfaces, but this is only a superficial solution. A software component exposing an API is like a hotel publishing a picture of the entrance lobby, without describing the the bedrooms, parking space, availability of wi-fi, and so on. The user might not find out how good or bad it really is until too late.

Back in the 1990s, APIs were programming language function calls that applications made to operating systems such as UNIX®. They can now take many different shapes and forms, most usually HTTP requests to web services.

They do provide a basic means of inter-operation for software components, but it is not a very efficient one. There has been an explosion of web service APIs, but many are of poor quality, and with little or no documentation, making user programs hard to debug. In a large-scale system, undisciplined use of APIs soon results in a tangled mess of interdependent modules, a spaghetti architecture that is impossible to modify or maintain; exactly what an agile enterprise that needs to change its software frequently does not want.

Yes, the components must expose and use APIs, but in itself this is not enough. They must follow clear principles that will guide additions and changes. Often, they should use a software platform to reduce complexity and duplication.

Data Is The Lifeblood

Availability of information is the basis of many of the new digital businesses. Trudging down the High Street looking for bargains on a wet Saturday afternoon is what you do if you don't know who sells what for how much. Driving your taxi round in circles waiting for someone to hail you is what you do when you don't know who wants to go from where to where. These are things that, with the right digital technology, shoppers and taxi drivers can know easily. Companies such as Amazon and Uber took advantage of digital technology to make information readily availability to people that needed it.

The case of Netflix is more subtle. It is tempting to say that going to the cinema is what you do when you can't get video to your home, and this is one explanation. Another is that playing a movie to an empty theatre is what you do when you don't know what your customers like. These are two separate business reasons for success. The video distribution appliances support the first, the streamed event logs support the second.

Information is the oxygen of an enterprise. Data is its lifeblood, carrying the information between the parts that produce it and the parts that consume it.

Understanding the data flow is the key to configuring components to create digital business offerings. The components must be arranged to obtain and deliver the information that provides the business value. If they are to fit together, they must have common data models, so that they interpret the data in the same way. Often, new components are required to transform data so that it can be exchanged by components whose data models do not match.

Creating a Technology Architecture

Success starts with a vision. Eric Ries and traditional enterprise architects at least agree on that. The Lean Startup explains how to turn a vision based on new technology into business success. Designed for Digital explains how a company can use new technology to become more successful, fulfilling its existing vision.

Designed for Digital presents a reference model of organizational capabilities for digital transformation, but does not describe a detailed technology architecture. Perhaps that is too much to expect. The reference model and roadmap discussion are major achievements on their own.

Describing how to create a technology architecture for a digital enterprise will be a task for the enterprise architecture community. They will have to review their principles and practices. Many of the fundamental principles may still apply. Many of the practices must surely change.

Architects specifying technology platforms have always had a requirement to support enterprise data flows. New technology has led to new business possibilities, based on availability of information. If enterprises are to take advantage of those possibilities, that requirement must be brought to the forefront. Data must be the starting point for technology architecture development.

Design for Data!

Designed for Digital is an excellent book. It will help enterprise architects understand the business architecture of digital enterprises. It will help them less to understand how to create the technology systems to support the business architecture. To do this, they must design for data!