Decision Guides
Use these guides when you’re unsure which pattern to apply. Each guide addresses a common decision point.
OU vs Group
Section titled “OU vs Group”Question: Should this be an OU or a Group?
- OU = Policy boundary (device posture, session length, DLP rules). Answers “what security controls apply to this person?”
- Group = Access list (drive permissions, system roles, mail distribution). Answers “what can this person do?”
Rule: If it controls security posture, it’s an OU. If it controls what they can access, it’s a Group.
Department vs Team
Section titled “Department vs Team”Question: Should I create a new Department or a new Team?
Create a Department when:
- The function owns budget, policy, or platform scope
- It will contain multiple Teams for 12+ months
- It has a distinct compliance or audit surface
Create a Team when:
- It’s a working group within an existing Department
- It might not exist in 6 months
- It doesn’t own budget or policy scope
Rule of thumb: If it doesn’t own budget, it’s a Team.
Role Ownership
Section titled “Role Ownership”Question: Which team should own this role?
| If the role affects… | Owner is… |
|---|---|
| Org-wide security posture | plt or sec |
| Public-facing content | mktg (with plt gating dangerous options) |
| Engineering systems | eng |
| Customer delivery | ops |
| Financial systems | fin |
| HR systems | hr |
| Legal systems | legal |
| Cross-functional analytics | biz |
Mail Pipe Chooser
Section titled “Mail Pipe Chooser”Question: Which mail pipe should I use?
Is the sender a human? YES → Is the audience internal? YES → Is it a broadcast (few senders, many readers)? YES → Mail (Announce) NO → Mail (Discussion) NO → Mail (Public or C-Inbox) NO → Is it urgent / on-call actionable? YES → Alerts NO → Is it a report for later review? YES → Intake NO → Infra (router)Public Contact: C-Inbox vs Standard Mail
Section titled “Public Contact: C-Inbox vs Standard Mail”Question: Should this public address be a Collaborative Inbox?
- C-Inbox when you need assignment, queuing, and SLAs (support, sales inquiries).
- Standard Public when you just need moderated delivery to a small responder set (press, general contact).
Drive Channel Selection
Section titled “Drive Channel Selection”Question: Which drive channel should this content go in?
| Content Type | Channel | Why |
|---|---|---|
| Final deliverables, agendas, status updates | COLLAB | Safe for customer/external eyes |
| Working drafts, internal notes, AI summaries | TEAM | Internal only |
| Raw recordings, evidence, PII/PHI | STRICT | Need-to-know, chain-of-custody |
Shared Drive vs Folder vs Repo
Section titled “Shared Drive vs Folder vs Repo”Question: Should I create a new drive, a folder, or a git repo?
- New drive when: new customer, new major function, or sensitivity boundary needs isolation.
- New folder when: new project within existing customer/function.
- Git repo when: code, IaC, or version-controlled config.
Vendor vs Partner vs Customer
Section titled “Vendor vs Partner vs Customer”Question: Which taxonomy should I use?
| They serve us | → Vendor |
|---|---|
| We serve them | → Customer |
| We collaborate as peers | → Partner |
Security Label: ON vs OFF
Section titled “Security Label: ON vs OFF”Question: Should this group have the Security label?
- ON for groups that appear on ACLs or grant system access (Dept, Team, Role, some Customer groups).
- OFF for communication/routing groups (Mail, Intake, Alerts, Infra, People, Ident).
Warning: Security label is irreversible. When in doubt, leave it OFF — you can always recreate with it ON, but you can’t turn it OFF.
Infra Router vs Direct-to-Alerts
Section titled “Infra Router vs Direct-to-Alerts”Question: Should this system post directly to Alerts, or go through an Infra router?
Use Infra router when:
- Multiple noisy sources feed one on-call lane
- You need classification (one source, many topics)
- You need cross-tenant bridging
- Volume is high and you want centralized allowlisting
Use Direct-to-Alerts when:
- Single, high-signal source
- Low volume
- Simple routing (one source → one alerts group)
Breakglass / Emergency Paths
Section titled “Breakglass / Emergency Paths”When normal processes are blocked (SSO outage, IdP failure):
- Unsuspend breakglass account (two-person rule).
- Login with FIDO2 escrowed keys.
- Perform scoped admin tasks.
- Re-suspend breakglass account.
- Record evidence (ticket, approver, timestamps, actions taken).
- Post-incident review within 48 hours.