Notion and Confluence: separating confidential HR and finance pages before a data review

Why start with permissions?#

Permissions are the best first signal because people restrict the pages they already consider sensitive. In Confluence, space permissions and page restrictions show where access was deliberately narrowed, and a view restriction on a parent page also hides its child pages. In Notion, private pages, closed or private teamspaces, and pages shared only with named people play the same role.

Permissions are a starting point, not a verdict. Some HR pages sit in open spaces because nobody thought to restrict them, and some restricted pages are restricted for reasons unrelated to sensitivity, such as an unfinished draft. Owners and content checks close those gaps.

Sharing outside the company is a separate signal worth capturing. Pages shared with guests, contractors or customers, or published through public links, often hold material that belongs to someone else, such as a customer's onboarding plan or a contractor's statement of work. List them alongside restricted pages and review them as a group.

Space-by-space triage checklist#

The triage runs on a list of spaces or teamspaces with their metadata, not on page content. A workspace admin can usually produce that list without opening sensitive pages, which keeps the first pass limited to people who already have access.

Plan the export route at the same time, because each route behaves differently. Confluence Cloud space admins can export a whole space or selected pages to PDF, CSV, HTML or XML, and a Confluence site backup includes attachments only if they are selected. A Notion workspace export leaves out pages the exporting user cannot access, such as other members' private pages, so an export run by a regular member is not a complete inventory. Settle the route before the triage is finished, so the excluded spaces can be left out of the export itself rather than removed afterwards.

• 1List every Confluence space and Notion teamspace with its owner, permission summary, approximate page count and last update date.
• 2Classify each space as exclude, include after review, or mixed.
• 3Inside mixed spaces, list pages with view restrictions or limited sharing and mark them excluded by default.
• 4Search labels and tags such as confidential, hr, comp, finance, legal and board.
• 5Search titles and text for sensitive terms, then read the hits.
• 6Ask each space owner to confirm the classification in writing.
• 7Record every decision, with the reason and the owner, in one triage sheet.

Which spaces should be excluded by default?#

Spaces built around people, money, legal matters or company strategy are excluded whole, because reviewing them page by page costs more than the little licensable content they hold. Spaces built around product, engineering and support work are where licensable knowledge lives, and they get page-level review.

Where sensitive pages hide in open spaces, and how to flag them#

HR and finance content often hides in spaces that look safe. Engineering managers keep hiring notes beside architecture docs, and product teams paste cost figures into planning pages. These are the places to check inside every mixed space.

Inside a mixed space, a few signals catch most sensitive pages, and each one is cheap to check. Combine them rather than relying on any single one, then send flagged pages to the space owner for a decision.

Notion databases deserve a specific look. A database in an engineering teamspace can carry a property such as salary band, rating or start date that turns an innocent-looking table into an HR record. Check database properties as well as page text.

• 1Interview scorecards, hiring loop notes and candidate databases kept by engineering teams.
• 2One-to-one templates and meeting notes that include performance feedback.
• 3Team retrospectives and incident postmortems that name individuals.
• 4Headcount, level and promotion planning inside engineering or product planning pages.
• 5On-call pay, bonus or overtime arrangements in operations runbooks.
• 6Vendor pricing, invoices and contract terms on procurement or tooling pages.
• 7Embedded spreadsheets and dashboards that pull from finance systems.

Illustrative: a data integration SaaS triages two wikis#

Illustrative: a fictional data integration software company runs Confluence for engineering and support and Notion for company operations. The COO leads a triage before a data review.

In Confluence, the engineering and support spaces are kept for review, while an engineering leadership space holding hiring plans and level guides is excluded whole. A text search finds restricted performance documents filed under the engineering space, which are excluded. In Notion, the people, finance and leadership teamspaces are excluded without review, and an interview scorecard database inside the engineering teamspace is excluded on its own. Each owner signs off in the triage sheet, and the review set ends up narrower and far easier to prepare than the full workspace.

The triage also surfaces a handful of customer onboarding plans shared with external guests in the support space. They are moved to the review queue with the sales pages, because they name customers and their internal contacts.

How SourceX approaches internal documentation#

SourceX starts with metadata only: which wikis you use, which spaces exist and roughly how many years they cover. No pages are shared during the fit check, and HR, finance and legal spaces are expected to stay out of scope.

In the Rights and Preparation steps of the SourceX five-step transaction, the triage sheet becomes the record of what was excluded and why, and it feeds the privacy record and release authorization in the SourceX Evidence Packet.