Some years ago I inherited an architecture practice whose previous lead had printed the Zachman Framework at A0 and mounted it on the wall of the main meeting room. Six columns, six rows, thirty-six cells, and — as far as I could establish from the repository — four of those cells had ever been populated. The poster was beautiful. It had also become a standing joke, which is a worse outcome than never putting it up.
That poster is the whole framework question in miniature. People do not choose badly between TOGAF, Zachman and FEAF because they misjudge the intellectual merits. They choose badly because they treat three fundamentally different kinds of artefact — a method, an ontology and a government mandate — as if they were competing products in the same category. They are not. Once you see what each one actually is, the question of which to learn mostly answers itself.
The short version
TOGAF is a method for running an architecture function, and it won the market. Zachman is a classification scheme for architectural descriptions, and you should understand it in an afternoon and then stop. FEAF is a US federal compliance artefact, and it matters to you only if your paycheque depends, directly or indirectly, on a US government agency.
If you have limited study hours and you want them to convert into work, the allocation is roughly: TOGAF eighty per cent, Zachman ten per cent, FEAF ten per cent if and only if you are in that sector, otherwise zero.
The rest of this piece is why.
TOGAF: it won, and that matters more than whether it deserved to
I am not going to argue that TOGAF is the most elegant thing on the shelf. It isn’t. The ADM is a method with the confidence of a standard, the Architecture Content Framework has more moving parts than most organisations will ever populate, and the pre-10 editions had a genuine sprawl problem that gave a generation of engineers a permanent allergy.
What TOGAF has is distribution. It is the framework named in job descriptions, requested on procurement forms, and assumed in the vocabulary of the steering committees you will sit in. When a programme director says “we’re in Phase B”, you need to know without asking that they mean business architecture, and that Phase C is data and applications and Phase D is technology. That is not deep knowledge. It is table stakes, and not having it costs you credibility you would rather spend elsewhere.
The 10th edition materially improved the situation by splitting the Fundamental Content from the Series Guides and being explicit that the Guides are optional. That restructure is the single best argument for revisiting TOGAF if you wrote it off in the 9.x era — the core is now small enough to read honestly in a weekend. I have written a full review of
The TOGAF Standard, 10th Editionat /togaf-standard-10-review/.
Where TOGAF earns its keep in practice: giving a distributed architecture function a shared sequence and shared artefact names, so that four architects in four business units produce work that can be compared. Where it does not: as a substitute for judgement about what to build. The ADM will happily walk you through eight phases of governance for an architecture that should never have been approved.
Learn TOGAF because it is the lingua franca. Do not learn it expecting it to make your decisions for you.
Zachman: an afternoon, not a programme
John Zachman published the original framework in the IBM Systems Journal in 1987, and it remains the most intellectually honest thing in this comparison — partly because Zachman has spent decades insisting on what it is not. It is not a methodology. It does not tell you what to do, in what order, or when you are finished. It is an ontology: a two-dimensional classification of architectural descriptions, with the six interrogatives (what, how, where, who, when, why) on one axis and the stakeholder perspectives on the other.
Used as intended, it is a superb diagnostic. Draw the grid, put your existing artefacts in the cells, and look at what is empty. In my experience the where and why columns are almost always bare, and the “why” gap in particular predicts a specific organisational failure: an architecture nobody can defend because nobody wrote down the motivation. Half an hour with the grid tells you more about the state of an architecture practice than a month of maturity assessment.
Used as a programme — as a thing you resolve to fill in — it is a project-shaped hole in your budget. Thirty-six cells times any real enterprise is a documentation effort with no natural stopping point and no consumer. That A0 poster on my wall was somebody’s attempt to do exactly this, and the attempt is what killed the idea in that organisation for years afterwards.
So: read about it, understand the two axes, use the grid as a gap-finder in workshops, and do not spend money on a certification for it. Nobody has ever filtered a CV on Zachman. The framework’s value is entirely in the thinking, and the thinking is available in an afternoon.
FEAF: only if you sell to the US government
The Federal Enterprise Architecture Framework is what happens when architecture becomes a compliance obligation. Its current form, FEAF v2, was published by the US Office of Management and Budget in 2013, built around a Consolidated Reference Model — performance, business, data, application, infrastructure and security — plus the Collaborative Planning Methodology. It exists because US federal agencies are required to justify IT investment in a comparable way across the whole government, a lineage that runs back through the Clinger-Cohen Act.
That origin explains both its strengths and why it is largely irrelevant outside its jurisdiction. The reference models are genuinely good at what they are for: making spend and capability comparable across separate organisations that do not want to be compared. That is a real and difficult problem, and almost no private-sector enterprise has it.
If you work for or bid to a US federal agency, FEAF is not optional and TOGAF is not a substitute — you will be asked for artefacts in the reference-model shape regardless of what method produced them. The documentation is free, it is readable, and a couple of days with it is enough to hold a competent conversation. If you do not work in that sector, reading FEAF is a hobby. The same is true of its defence-sector cousins, DoDAF and NAF: essential inside the fence, noise outside it.
What hiring managers actually filter on
I have sat on both sides of enough EA hiring panels to be blunt about this.
The recruiter screen filters on the string “TOGAF”, usually because someone pasted it into the job description from the last one. That is the only framework filter that exists in volume. Nobody screens for Zachman. FEAF appears in US public-sector and systems-integrator postings and essentially nowhere else.
The interview filters on something entirely different, and this is the part candidates get wrong. What I am listening for is whether you can describe a decision you made, the options you rejected, and what it cost when you were wrong. Framework fluency shows up as a lubricant in that conversation — it lets you compress the setup — but reciting ADM phases in an interview reads as studying rather than practice, and it reads that way immediately. I have never declined a candidate for not knowing Phase E. I have declined several who could recite all of it and could not explain why their last target architecture went unbuilt.
The practical implication: get whatever certification your market genuinely filters on, then stop optimising for the filter. I have written the longer version of that calculation in is TOGAF certification worth it.
The reading that builds judgement rather than trivia
Sequence matters here more than volume. Three steps, in this order.
First, the strategy layer. Read
Enterprise Architecture As Strategybefore you read any framework. The reason is specific: TOGAF tells you how to run the process, and it presumes you already know what kind of company you are building for. Ross, Weill and Robertson give you the operating-model diagnostic that answers that question, and it is the thing that makes framework choices fall out naturally instead of being argued about. Read a framework first and you will produce very tidy documentation of the wrong target state. Full review →, and the operating-model quadrant on its own is covered in the four operating models.
Second, the framework itself, read as a source. The Fundamental Content of
The TOGAF Standard, 10th Editionis the primary text. Read it once, properly, rather than absorbing it through consultancy slides — the slide version is where most of the folklore about what TOGAF “requires” comes from, and most of that folklore is wrong. Skip the Series Guides on a first pass; they are explicitly optional and reading them front-to-back is how people end up believing TOGAF is enormous.
Third — and only if you are sitting the exam — the study companion.
Mastering TOGAF 10exists to solve a different problem from the standard: the OGEA-102 scenario questions test pattern-matching against the framework in a way the standard itself does not prepare you for. Full review →, and the sequenced plan is at the TOGAF 10 self-study path.
For Zachman and FEAF, buy nothing. The 1987 paper and the OMB documentation are both free and both short enough to read in a sitting. Anything sold to you as a Zachman or FEAF course is selling you a certificate for a filter that does not exist.
TOGAF vs Zachman vs FEAF — quick answers
- Is Zachman better than TOGAF?
- They are not comparable in that way. Zachman is an ontology — a classification scheme for architectural descriptions — and it deliberately says nothing about process. TOGAF is a method for running an architecture function. You can use the Zachman grid to check the completeness of artefacts that a TOGAF ADM cycle produced; they are complementary, not rivals. If you have to pick one to learn properly, learn TOGAF, because that is the one the market filters on.
- Do I need to learn FEAF if I am not in the United States?
- No. FEAF is a US federal government framework built to make IT investment comparable across agencies, and its reference models reflect that mandate. Outside US public sector work and its supply chain, it has almost no currency. Read it if you are curious; do not spend study budget on it unless you are bidding for federal work.
- Which enterprise architecture framework do employers actually ask for in 2026?
- TOGAF, overwhelmingly, in the sense that it is the framework name that appears in job descriptions and procurement forms. Zachman essentially never appears as a hiring filter. FEAF appears in US public-sector and systems-integrator roles. Check the last ten job descriptions in the segment you are targeting rather than trusting anyone's general claim, including mine — the picture varies sharply by sector.
- Can you use TOGAF and Zachman together?
- Yes, and this is the most sensible way to use Zachman. Run your architecture work with the ADM or whatever method your organisation actually uses, then use the Zachman grid periodically as a gap analysis: put your existing artefacts into the cells and look at what is empty. The 'why' and 'where' columns are usually the bare ones, and that gap tends to predict real problems.
- How long does it take to learn each framework?
- Zachman: an afternoon to understand the two axes and how to use the grid as a diagnostic. FEAF: a couple of days with the OMB documentation to hold a competent conversation. TOGAF: a weekend to read the Fundamental Content, and considerably longer — most people report tens of hours — if you are preparing for the OGEA-101 and OGEA-102 exams rather than just reading for comprehension.
Bottom line
Learn TOGAF, because the market runs on it and fluency in the vocabulary is cheap relative to what not having it costs you in rooms where decisions get made. Understand Zachman well enough to draw the grid from memory and use it as a gap-finder, then leave it alone — it is a thinking tool, not a programme, and the organisations that treated it as a programme are the ones where the poster became a joke. Read FEAF only if someone is paying you to comply with it.
And then spend the hours you saved on the thing that none of the three frameworks will give you: the judgement to know which architecture is worth building in the first place. That comes from the strategy layer, from being wrong in public a few times, and from reading outside the framework literature entirely. The frameworks are how you write it down. They are not how you decide.
For the wider reading list this sits inside, see the best EA books in 2026.