Permission Model Conflict Across Brands
A holding company runs three brands in one Widen instance. Brand A's team must not see Brand B's unreleased campaign assets, but a shared "Corporate" folder must be visible to all three. A junior admin previously granted the "All Users" group upload and edit rights at the top-level folder to speed things up. Now there are 180,000 assets, 40 users across three brand groups, and reports of Brand A users finding Brand B launch imagery by searching. Legal has flagged this as a confidentiality issue two weeks before a joint announcement. There is no documented permission matrix.
1. Explain why the current setup leaks assets across brands and outline the correct permission structure.
Read the worked answer
Why it leaks: permissions in Widen are inherited down the folder tree, and search surfaces anything a user can access anywhere they have rights. By granting "All Users" upload/edit at the top level, the junior admin gave every user access to the entire tree, including Brand B's folders - inheritance flows downward, so a top-level grant overrides the intent of brand-level separation. Search then dutifully returns Brand B assets to Brand A users because they technically have access. Upload/edit at top level also lets anyone drop assets anywhere, compounding the mess.
Correct structure:
- Remove the top-level "All Users" grant. That is the root cause; nothing else matters until it is gone.
- Build brand-scoped groups (Brand A, B, C) and grant each group rights only on its own brand folder subtree.
- Grant "All Users" access only to the Corporate folder (and only the access level needed - likely view/download, not edit).
- Keep inheritance intentional: set explicit permissions at each brand root and the Corporate root; avoid broad grants higher than necessary.
- Document the matrix: group โ folder โ access level, so this is auditable next time.
Verify by logging in as a test user in each brand group and confirming search no longer returns other brands' assets.
2. With two weeks to the announcement, what is your remediation sequence and how do you verify the leak is closed?
Read the worked answer
Sequence (fast, reversible, verifiable):
- Contain now. Remove the top-level "All Users" edit/upload grant immediately - this is the single change that stops the bleed. Do it first, before building the perfect matrix.
- Re-grant Corporate access to All Users at view level so shared assets keep working.
- Stand up brand groups and assign folder-scoped rights. Audit the 40 users' group membership against brand.
- Audit recent activity for Brand A users who accessed Brand B assets - Legal will want to know scope of exposure.
- Document the matrix as you go so it is not tribal knowledge.
Verification: create (or use) a test user in each brand group, log in, and run searches for known Brand B launch asset names/keywords as a Brand A user. Confirm zero results. Repeat across brands and confirm Corporate is still visible to all. Do not declare it closed on configuration alone - prove it with a search test from each persona, and record the results for Legal.