Hey everyone! I've been lurking for a bit and finally have something to share. I'm part of a data engineering team that just finished migrating all our project docs and runbooks from Notion to Confluence. We're about 15 engineers in total, split between data and platform. The migration itself took a solid month of evenings, but we've been fully on Confluence for about six months now.
I wanted to get some perspective from folks who've managed similar tools at this scale. Our leadership pushed for Confluence mainly for "enterprise integration" with Jira (which we also use). The per-seat cost for the Premium tier adds up, and I'm honestly not sure if the productivity gains are there for an engineering team like ours. The search feels slower than Notion, and the editing experience is... clunkier for technical docs. We do like the official page versioning, though!
On the data side, we tried to use it for some basic pipeline documentation—linking to our dbt docs and Airflow DAGs—but it feels like we're just pasting URLs into a table. I'm curious if anyone has built a more integrated setup? Maybe with some automated syncing?
And the big question for the room: for a 15-person technical team, does the tighter Jira integration and permission granularity justify the invoice over something like Notion's team plan? I sometimes wonder if we're just paying for the brand name. Would love your honest takes!
-- rookie
rookie
I'm a product manager at a SaaS startup with 40-odd folks, 12 engineers. We run both Confluence (for product specs and release docs tied to Jira) and Notion (for engineering wikis and runbooks) in prod.
Four things I'd check:
**Actual pricing** - Atlassian's pricing per user is straightforward but expensive at scale. At my last shop, Confluence Premium was $7.50/user/month, so your 15 eng team is ~$1125/year. Notion's Business plan is $8/user/month billed annually ($1440/year for 15). The hidden cost is the admin time to manage permissions and structure in Confluence, which we spend about an hour a week on.
**Integration vs. just linking** - Confluence wins if you're truly embedding Jira issues, sprint reports, or live data via Atlassian's own apps. But if you're just pasting URLs (like to dbt docs), that's not integration. I haven't found automated syncing for pipeline docs that feels native; you end up with manual updates.
**Technical editing friction** - For pure technical docs, Confluence's editor is slower. Our engineers say pasting code snippets and updating markdown-like content takes about 3-4x longer than in Notion. The versioning is a real win for change tracking, though.
**Search performance** - Confluence search felt slower for us too, especially on cold starts. We see results in 2-3 seconds on average, versus under a second in Notion for our internal wiki.
My pick: I'd keep Confluence only for product/Jira-linked docs, and move engineering runbooks back to Notion. But that depends on two things: how strict your leadership is about a single tool, and if you actually use live Jira widgets in pages. If you don't, tell us that.
You're right about pricing being deceptively simple, but I've found the true long-term cost is in rigidity. The per-user fee is clear, but it assumes a static team. In a 15-person engineering group with contractors, interns, or occasional external reviewers, you're forced into managing users as a system. That admin hour a week balloons during onboarding periods or re-orgs, which for engineering teams can be frequent.
Your point on technical editing friction is understated. It's not just 3-4x slower for markdown-like content. The real productivity sink is when you're documenting a data pipeline and need to paste a sample JSON payload, a snippet of SQL, and a CLI command in sequence. In Notion, that's a few keystrokes. In Confluence, you're switching between the 'code block' macro, manually selecting the language, pasting, then fighting the editor's tendency to reformat your CLI command into a numbered list. The versioning is indeed a win for audit trails, but it doesn't offset the daily drag on creating content. Engineers stop documenting nuances because the tool fights them.
Where I diverge slightly is on integration. For pipeline docs, true automated syncing is rarely native to either platform. The win with Confluence isn't live data; it's the bidirectional link with Jira that actually propagates status changes and enforces a workflow. A spec page linked to an epic automatically reflects when that epic is marked done. That's powerful for process-oriented teams, but it's a cultural shift, not just a tool feature. If your team wasn't already disciplined in Jira, you won't get the value, and the friction becomes a net negative.
—BJ