Organisational Antipatterns
Overview
Section titled “Overview”The previous two topics assumed an organisation that merely lacks an architecture practice. Some organisations have shapes that actively defeat the role.
Four of them, each with a diagnosis you can run in your first weeks and a response that is genuinely available rather than aspirational. The common thread is stated at the end, and it is the most useful sentence in the topic.
The cursed role
Section titled “The cursed role”What it is. A position nobody has held successfully for more than a year. There have been two or three predecessors. Each was competent. Each left.
Diagnosis. Ask how many people have held this role and what happened to them. Ask in the interview if you can, and in week one if you could not. The answers are usually freely given and rarely volunteered.
What it means. The failure belongs to the organisation rather than to its occupants. Something structural defeated three capable people, and arriving with more energy will not change it. Common causes: no authority paired with full accountability, a mandate that two executives define differently and neither will reconcile, or a role created to avoid a reorganisation that the organisation actually needed.
Response. Find out what defeated them before committing to a plan that assumes it is gone. If the cause is structural and unfixable from your position, the honest response may be to leave — which the next topic treats properly rather than as a throwaway.
Quiet sabotage
Section titled “Quiet sabotage”What it is. Not refusal. Slow reviews, dependencies that are forgotten, a team that deprioritises your work, enthusiastic agreement in a meeting followed by nothing.
Diagnosis. The gap between stated agreement and observable progress. Nobody says no; the work does not move. Measure the gap rather than sensing it — a list of agreed items with dates and status makes the pattern visible and is also the only form of this diagnosis that survives disagreement.
Response. Three things, and the third is a prohibition.
Make progress visible, so that absence of progress is visible too. A published status is uncomfortable for exactly the right reasons.
Find the interest underneath, per Module 1. Quiet resistance is almost always a rational response to an unaddressed interest — a team that fears losing autonomy, an engineer who thinks the design is wrong and lost the argument, a lead protecting a commitment you are threatening.
And never name it as sabotage. You would be accusing someone of something unprovable, and you would lose — both the argument and the relationship. The response to resistance is to address its cause, not to characterise it.
Inherited expectations
Section titled “Inherited expectations”What it is. Somebody promised what you would deliver before you arrived, and they promised different things to different people.
Diagnosis. Ask each stakeholder, individually, what they expect from you in the first quarter. Write the answers down and compare them. The divergence is the measure, and it is usually larger than anyone would predict.
Response. Surface and reset early, individually, in writing. This is Module 1’s charter conversation with a diagnostic behind it — you are not asserting your remit in the abstract, you are resolving a specific contradiction between two people’s expectations, which is a much easier conversation to have and a much harder one to defer.
Incompetence you cannot name
Section titled “Incompetence you cannot name”What it is. A senior person whose technical judgement is poor and whose position is secure.
Diagnosis. Decisions that do not survive scrutiny but survive anyway. The tell is a pattern rather than an instance: good arguments lose to the same person repeatedly, and the reasons given change each time.
Response. This is the genuinely hard one, and the honest answer is that you cannot fix it and you cannot say it.
What works is mechanism rather than confrontation: make the decision criteria explicit and public before the decision is made. A criteria table agreed in advance means a poor choice has to argue against a written standard rather than against you personally — and arguing against a standard the group agreed to is much harder than overruling an individual. Section 4’s comparison table is this instrument, and this is the situation it was built for.
What does not work: being right in a meeting. Being right harder. Escalating, in an organisation where their position is secure.
The common thread
Section titled “The common thread”graph TD A["Cursed role"] --> E["<b>A mechanism,<br/>not a conversation</b>"] B["Quiet sabotage"] --> E C["Inherited expectations"] --> E D["Unnameable incompetence"] --> E E --> F["Published status · explicit criteria ·<br/>written expectations · recorded decisions"] classDef prob fill:#D9695B,stroke:#B4483A,stroke-width:2px,color:#ffffff classDef key fill:#E8A33D,stroke:#C4832A,stroke-width:2px,color:#ffffff classDef out fill:#7D9A4E,stroke:#5F7A38,stroke-width:2px,color:#ffffff class A,B,C,D prob class E key class F out
Four different problems, one shape of response. You cannot out-argue a structural problem — which is why each response is an artifact or a process rather than a better conversation.
Published status makes absence visible. Explicit criteria make a poor decision argue against a standard. Written expectations make a contradiction somebody else’s to resolve. Recorded decisions stop the same argument recurring.
None of those requires authority, which is why they are available to you.
Feeds the workbook
Section titled “Feeds the workbook”This is your antipattern assessment. Assess all four, say which are present with the evidence, and for each present one name the mechanism rather than the conversation. An assessment that concludes “we should communicate better” has not used this topic.
→ One organisational force deserves its own topic, because it is both the strongest objection you will face and your best lever. — Scar tissue