An enterprise architecture team has worked through a formal framework's method for eight months and produced catalogues of applications, data entities and technology standards. Delivery teams do not open any of it. Which part of the method was skipped?

What is being tested

Whether you understand what a framework is for. TOGAF's method is a sequence of questions, and its content framework exists so that each artefact answers a named concern held by a named stakeholder. That lineage comes from ISO/IEC/IEEE 42010 (2011), where a view is defined as addressing a viewpoint which addresses concerns.

Skip the concern and you still produce the artefact, because the method tells you what artefacts exist. The artefact then has no reader, and "nobody opens it" is the correct outcome rather than a cultural failure.

The reasoning

A catalogue of 400 applications is not useful or useless in itself. It becomes useful when someone has a question it answers: which applications process personal data in the EU, which run on a runtime losing support in fourteen months, which two systems do the same thing for different business units. Each of those questions determines different columns, and a catalogue built without the question has the wrong columns, which is why it is both expensive and unread.

Eight months of a five-person team is roughly 7,000 hours, and the cost of that is not the hours but the credibility: the next catalogue the team proposes starts from a worse position. The test is one sentence per deliverable: who will act differently because of this, and what decision are they making. If it cannot be written, the deliverable is not architecture work, it is inventory work. Running the test before the eight months is the entire difference between a framework used as a checklist and a framework used as a production plan.

Why the other options fail

• Gap analysis is valuable and arrives too late to save this. A gap between a baseline nobody asked for and a target nobody agreed to is a document about two documents. Gap analysis is how you sequence work once the concerns are known; it does not generate the concerns.
• The complete content metamodel is the opposite of the fix. Tailoring the metamodel down to what the organisation will actually maintain is explicit guidance in the framework, and adopting it whole is the most common way to produce unmaintainable catalogues. More coverage increases the decay rate without increasing the number of answered questions.
• The architecture repository is a storage decision. A repository with version control holding unread documents is better organised and no more used. Tooling is the last problem here and usually the first one procured.

When this is the wrong answer

If the organisation genuinely has no inventory and cannot answer basic questions about its estate, a deliberately minimal catalogue built ahead of demand is defensible — but then say so explicitly, keep it to the handful of fields you can keep current, and attach a named decision it will support within the quarter. The failure in the question is not that a catalogue existed; it is that eight months of artefacts were produced without a single named reader.

Common weak answers

• "They needed better communication." The deliverables were not under-communicated, they were unmotivated. Presenting an unread catalogue more loudly does not create a question it answers.
• "The framework is the problem." The framework explicitly says to tailor it and to drive from stakeholder concerns. Following half a method and blaming the method is the most common diagnosis and the least useful one.