Enterprise architecture is not TOGAF

The Open Group Architecture Framework (TOGAF) is currently promoted as an industry consensus framework for enterprise architecture (EA) representing the best practice of numerous EA practitioners.

However, my observations of successful companies using TOGAF as a basis for their EA efforts suggest that most TOGAF recommendations are usually found inapplicable, while the most critical parts of EA practice are usually established from scratch in company-specific ways.

Therefore, my findings question the value of TOGAF as a standard for EA practice since successful TOGAF-based EA initiatives hardly overlap with the actual TOGAF prescriptions. More realistic and evidence-based sources on EA are needed.

The value of TOGAF

In the recent years The Open Group Architecture Framework (TOGAF 2011) has gained prominence as the best-known framework for EA practice. It is positioned as an ‘open, industry consensus framework for enterprise architecture’ (TOGAF 2011, p. xxiii) representing the best practice of multiple companies practicing EA.

Its core feature, Architecture Development Method (ADM), is ‘the result of continuous contributions from a large number of architecture practitioners’ (TOGAF 2011, p. 45). Unsurprisingly, TOGAF is heavily promoted by various consultancies, experts and trainers. Is it really so valuable for organisations? Does it really reflect the best practice of successful organisations using EA?

My observations of leading Australian companies (who wished to remain anonymous) suggest negative answers to these questions. Although these companies use TOGAF as the key EA framework (one of them is even included in the ‘official’ list of TOGAF users on The Open Group website, (see TOGAF (2015)), their actual activities hardly resemble TOGAF prescriptions.

For instance, ADM - the core of TOGAF - is found inapplicable in practice and its steps are not followed by any of these companies. Similarly, other essential TOGAF features, including the Architecture Content Framework and Enterprise Continuum, are also not used in companies ‘using’ TOGAF.

Instead of following the step-wise and iterative ADM they established idiosyncratic sets of continuous EA-related processes. Instead of developing heaps of EA artifacts recommended by the Architecture Content Framework they established small pragmatic sets of value-adding documents with clearly defined aims, stakeholders and purpose. Consequently, the most critical parts of TOGAF-based EA initiatives were essentially ‘invented’ from scratch in unique company-specific ways.

On the other hand, useful TOGAF features adopted in these companies, such as four EA domains or architecture principles, are not really TOGAF-specific but rather conventional and have been used for decades (Davenport et al. 1989; PRISM 1986; Richardson et al. 1990). Therefore, the core parts of TOGAF were found largely useless, useful parts of TOGAF were not TOGAF-specific, while the most critical parts of EA practice were established from scratch.

In light of these findings the growing popularity of TOGAF can hardly be attributed to the real usefulness of its advice, but rather to a lack of any better alternative sources on EA. This conclusion supports the opinion of Bloomberg (2014a, p. 1) that ‘TOGAF’s accelerating success is simply because it’s the only game in town’ and that ‘TOGAF has gained traction simply because it’s better than doing nothing’.

For you

Be part of something bigger, join BCS, The Chartered Institute for IT.

Despite the fact that many TOGAF recommendations are considered as overly complex and largely unrealistic, it still remains the most comprehensive available source on EA where some useful ideas can be found, while other EA frameworks are even less helpful. Consequently, the choice between TOGAF and other EA frameworks is essentially the choice between something and nothing, bad and very bad, largely useless and totally useless.

Although TOGAF is briskly discussed, heavily promoted and its certifications are required for enterprise architects at many companies, my findings suggest that successful EA efforts, even TOGAF-based, are built on pragmatic common sense ideas and rarely resemble the actual TOGAF recommendations.

Therefore, TOGAF is used, arguably, more as a symbol than as a realistic actionable guidance. Following TOGAF does not guarantee a successful EA practice to an organisation and getting TOGAF-certified does not guarantee a successful career to an EA practitioner. In short, TOGAF does not define EA practice, and successful EA is not TOGAF.

New sources on EA are needed

Despite the promise that ‘TOGAF provides a best practice framework for adding value’ (TOGAF 2011, p. 7), the analysis of its usage shows that this best practice can be hardly applicable and TOGAF can be hardly used in any real sense: only as a ‘dictionary’ where some useful information can occasionally be found or as a ‘horoscope’ from which vague recommendations can be adapted and interpreted to fit all organisations, but, not providing any real practicable guidance to any of them. Though positioned as a comprehensive end-to-end EA framework covering most parts of EA practice, TOGAF rarely provides useful advice on its most critical aspects.

Surprisingly, my conclusions are not innovative and far from new, but rather old and widely known. They are supported by numerous previous empirical studies (Ahlemann et al. 2012; Buckl et al. 2009; Haki et al. 2012; Holst and Steensen 2011; Ross et al. 2006; Smith et al. 2012) that consistently demonstrate that EA frameworks, including TOGAF, have only a minor value at best or are even detrimental at worst.

Consequently, this article is not a breakthrough, but yet another reminder that successful EA initiatives have not much to do with the popular EA frameworks that are, arguably, no more than a typical management fad since their value is more than questionable. ‘Frameworks are cocaine for executives - they give them a huge rush and then they move to the next framework’ (Bloomberg 2014b, p. 1).

Therefore, I argue that the EA community should take a sober realistic look at existing EA literature, critically revise and rethink its idealistic ideas that are presently very distant from the real-world EA practice in organisations. New evidence-based comprehensive sources on EA are desperately needed as a substitute for largely useless, but aggressively promoted, EA frameworks.

The views in this article are the views of the author and are his own personal views that should not be associated with any other individuals or organisations.