Success Story: SaaSJet's Experience Optimizing App Documentation with Scroll Sites for Confluence
SaaSJet has been around for over seven years, and as an Atlassian Marketplace vendor, it's clear that Confluence and Jira have been our long-time allies and helpers.
Our single source of truth
Confluence is where we keep what matters: decisions, docs, insights, and research. It keeps everyone on the same page—literally.
How it works for us
- One living workspace. Capture decisions, docs, and context side-by-side.
- Real-time collaboration. Comment, co-edit, and move work forward together.
- Jira, connected. Link work so nothing gets lost between brainstorming and delivery.
- Search & structure. Powerful search, page trees, and permissions keep knowledge findable and trusted.
- Templates & smart links. Start fast, stay consistent.
- Quiet automation. Create pages from triggers, nudge owners, and keep spaces tidy—no manual housekeeping.
- Built-in AI. Draft, summarize, and sharpen writing to get from blank page to clarity, fast.
What we get in return
- Fewer status meetings.
- Faster decisions.
- Onboarding in days, not weeks.
- A durable knowledge base that scales with the team.
Our link to customers
Confluence also connects us to the outside world. It’s where we publish product documentation—so users get clear, current guidance, and we keep a single source of truth from internal note to public page.
How SaaSJet used Confluence until recently
We’ve shipped ~20 apps to the Atlassian Marketplace. Each one has its own little “kingdom” in Confluence. Over time, those kingdoms grew… and grew. The more features we added, the more our docs puffed up like a balloon. Helpful? Yes. Easy to navigate? Not always.
This is what actually happened, what we broke, and what finally worked.
1) Structure: stop building mazes
When a user can’t find the answer—even if it exists—you either lose a lead or your support team gets a 10 p.m. “where is this?” ping. Been there. More than once.
What we changed:
- We set a hard rule: no more than 1–2 levels of nesting. If a page has one lonely paragraph, it probably shouldn’t be a page.
- We organized by features and user tasks, not by our internal teams.
- For our largest space (Time in Status is 100+ pages), we created the first page as a navigation hub, featuring a short human description under each page name and labels by feature at the bottom. Click a label, see everything that mentions it. Simple and boring in the best way.
Tiny checklist we actually use:
- 1–2 levels deep?
- Home page = navigation, not poetry.
- Pages grouped by what users try to do.
- Labels applied the same way every time.
And because it’s you and me talking: yes, we kept the “Minotaur of despair” joke. Documentation mazes deserve a monster.
2) Titles & content: less “commit message,” more “talk to a person”
We had titles like “JSON Data Feed Settings.” Cool. But what does that do for a user?
How we rewired our writing:
- Pretend you’re explaining it to a smart friend who’s new to the product. Feynman Technique, but without sounding like a textbook.
- Make titles task-y:
- ❌ JSON Data Feed Settings → ✅ Send sprint reports via smart link
- ❌ JQL Filter Configuration → ✅ Filter issues with JQL (copy-paste examples)
- Begin with a single, honest sentence: what you aim to achieve and the estimated time it will take.
- Show the thing. Screenshots. Numbered steps. Short, not epic.
- We do a “stranger test”: hand the draft to someone outside the team. If they stumble, we fix the doc—not the reader.
Micro-template we reuse:
- What you’ll learn (2 bullets max).
- When to use this (a couple of clear cases).
- Steps (5–8, numbered).
- FAQ / Troubleshooting.
- Related (smart links).
3) Releases: no more page whack-a-mole
A feature changes, and suddenly, nine pages go stale. Alerts help, but we still missed stuff.
What actually helped:
- We label by feature (average-report, grid, etc.). When something ships, we search by label and update all mentions in one sweep.
- We saved a couple of boring but life-saving searches: “Pages to update this sprint,” “Mentions of deprecated X.”
- On large pages, we include a minor change log block with a link to the release note. Users trust it; editors know what moved.
Release checklist:
- Which features changed? Map to labels.
- Search by those labels, update every hit.
- Re-run saved searches.
- If the structure changes, update the hub page accordingly.
4) Page quality: design = comprehension
We’ve all seen the wall-of-text page. Guilty, your honor.
What we use now (and it helps):
- Headings and sections so the eye can breathe.
- Callouts for “Important/Info/Tip” so key bits don’t hide.
- Tables for options, expanders for edge cases.
- Screenshots or tiny clips when words aren’t enough.
- A consistent template ensures pages feel familiar across all apps.
Before → After we aim for:
- Paragraph soup → short sections with H2/H3
- Inline aside → callout macro
- Bulleted steps → numbered with visuals
- Hidden gem → Tip/Important block
A quick word on analytics (the unglamorous hero)
- Searches with zero results = content we need to write.
- Top exit pages = where we need better “what’s next.”
- Support tags ↔ doc labels = find missing or unclear topics fast.
Why bother?
Because this cut down our “where is…?” tickets, sped up onboarding to tricky features, and made our docs feel like a companion—not a puzzle.
Next up, the part I’m oddly excited about: how automation and AI joined the story. Not as buzzwords, but as the reason I don’t spend Fridays chasing stale pages.
Unlocking More Value from Confluence: How Automation and AI Enhanced Our Documentation Process
As our documentation library grew, so did the operational challenges around maintaining content freshness, ensuring consistency, and making updates discoverable across teams. Confluence offered much more than just a space to store pages — we began exploring its automation features and Atlassian Intelligence (AI) capabilities to address these bottlenecks and scale our documentation efforts more effectively.
Automation That Saves Time and Reduces Manual Work
We implemented several Confluence automation rules that quickly became essential to our internal workflow:
- Automated content creation: Weekly changelog drafts and recurring meeting notes are now auto-generated from templates, consistently titled, and placed in the correct location within the space hierarchy.
- Content maintenance alerts: Rules automatically remind page owners to review documentation that hasn’t been updated in 90 days, helping us avoid outdated content.
- Consistent labeling: Newly created pages are tagged with standardized labels based on their context or location, supporting both internal navigation and external presentation strategies.
These automation rules eliminated repetitive tasks, helping our team stay focused on writing and refining content — not managing it.
AI-Powered Documentation Enhancements
Confluence’s AI tools have also played a meaningful role in enhancing the quality and clarity of our documentation:
- We utilize AI-generated summaries to provide concise overviews of lengthy how-to or configuration guides, which are beneficial for internal collaboration and creating condensed content for end users.
- Tone and clarity suggestions help ensure our product documentation reads professionally and accessibly, especially before external publishing.
- The innovative labeling feature, powered by AI, suggests relevant topics and applies thematic labels — a feature we now pair with our SEO keyword clustering strategy.
Together, automation and AI have turned Confluence into more than just a workspace — it’s become our documentation engine. However, we still faced one critical gap: how to deliver this well-maintained knowledge to users in a visually polished, SEO-friendly, and fully branded format. That’s when we turned to Scroll Sites by K15t.
Transform your Confluence documentation into a powerful traffic generation tool with Scroll Sites by K15t
We kept bumping into the same questions as our docs grew:
- We’ve built a solid knowledge base—couldn’t it also bring in organic traffic if search engines could index it?
- Where exactly are readers coming from? Which articles actually pull people in?
- We want a Help Center that looks and feels like a standalone site, but… we’re not spinning up and maintaining a whole new web project.
So we went hunting on the Atlassian Marketplace and, honestly, found the light at the end of the tunnel: Scroll Sites by K15t.
Why Scroll Sites clicked for us
We didn’t want a science project. We wanted “plug in, publish, see results.” Scroll Sites was exactly what we needed for our workflow.
- Publish Confluence → a clean static site. Point it at your docs, hit publish, and you’ve got a Help Center that lives on its own domain/subdomain.
- One site or many sites. We started with one, then realized we could spin up more for different audiences or products without duplicating effort.
- Flexible content sources. Bring in entire Confluence spaces or Scroll Documents—whatever matches how you write today.
- Brand it your way. Tweak design, navigation, and layout. Need more control? Add custom CSS/JS.
- AI-powered search. This was our pleasant surprise. It helps users get to answers faster (and reduces those “where is…?” tickets).
- Analytics & support hooks. Tie in your analytics and support tools to see how people arrive, what they read, and where they need a hand.
- OpenAPI docs & in-app help. Publish developer references alongside product docs, and embed contextual help so users don’t have to leave the product to learn.
We went looking for search indexing and nicer presentation—and ended up with a much bigger win (AI search was the “wow, didn’t expect that” moment).
Next, SEO — and the weird keyword puzzle
We did our homework on K15t’s tips. Great. But one question wouldn’t leave us alone:
how do we use keywords across everything without tripping over ourselves?
Because we weren’t just writing in one place. We had:
- our Atlassian Marketplace listing,
- articles on other sites,
- our website,
- and now a Help Center that also wants a slice of Google.
At some point, we looked at the spreadsheet and thought: okay, so we’re… competing with ourselves? Not ideal.
Here’s how we un-tangled it — not theoretically, but how we actually did it one Tuesday with coffee and a slightly chaotic Google Spreadsheet.
We consolidated all relevant queries into a single list and clustered them by frequency and intent. Nothing fancy: big, medium, tiny.
Then we gave each cluster a “home.”
- Big, high-intent terms (the money makers) → Marketplace page. That page does the heavy lifting, and we use those phrases as anchors when we link to the listing.
(Example: “time in status for jira” lives there.) - Informational/“compare” terms → website/blog.
- One-feature, very specific keywords → Help Center. If a user is typing “export time in status to csv,” they want a straight answer, not marketing.
We made one small rule that saved us from cannibalizing: one primary keyword per page. Synonyms can live in H2s and body copy, but only one page is “the grown-up” for that topic.
We linked everything together like normal humans would click: docs ↔ website ↔ listing. Descriptive anchors, no mystery meat.
And because guessing is expensive, we do a quarterly SEO check-in. It’s not a 40-page audit. We examine what’s climbing, what’s overlapping, merge thin pages, and relocate a keyword’s “home” based on the data.
Then we… waited. And this part honestly surprised us: the Help Center started catching the long-tail searches we used to ignore, while the Marketplace page kept its spot for the big terms. Support noticed fewer “where is…?” emails. We noticed clearer paths from search → doc → product.
It wasn’t elegant at first, but it worked. And yes, the spreadsheet is still messy. We just understand the mess now.
What SaaSJet gained by improving the documentation
We hooked everything up, hit publish, and then opened Google Analytics.
Year view (since switching to Scroll Sites):
- Traffic to our docs is up 7×. We double-checked because we didn’t believe it either.
- 5,545 organic sessions from search alone — which is wild considering our docs weren’t even indexed before. They used to be just “answers for existing users.” Now they pull new people in.
Audience snapshot:
- 8,000+ users in the last year.
- About 1 in 4 come back, which tells us the Help Center is actually useful, not a one-time detour.
Then we looked at Google Search Console — and that’s when we high-fived and went to celebrate.
GSC (last 6 months vs. previous 6 months):
- Total clicks: 924 → 2.37K (more than double).
- Total impressions: 104K → 367K.
- Average position: 28.6 → 21.7 (about 7 spots up).
That’s the kind of curve you print out and stick on a wall. Not perfect, not finished — but undeniable.
Scroll Sites turned our docs from “answers for existing users” into a genuine discovery channel. Our Help Center now looks and feels like a real site, shows up where people search, and actually guides them to what they need. We’re not chasing vanity metrics—we’re building clear paths from question to solution, and the feedback loop keeps getting tighter. It’s still a work in progress, but the direction is unmistakable: documentation that earns its keep.
Leverage Smart Forms by SaaSJet — easily collect partner contacts and customer feedback.
Scroll Sites does its job beautifully. Still, we’re product people—of course, we wanted one more thing. Part of our docs is a Partner Portal, and we’re always meeting new folks who might join us. What we didn’t want was to send them through a maze of Google Forms and manual copy-paste.
So we used something we know well: Smart Forms for Jira. It lets us spin up clean forms, collect responses from outside the org, and land everything straight in Jira—no retyping, no missed details, no “who owns this?” confusion.
The best bit (and the reason it clicked with Scroll Sites): you can embed a Smart Form on any web page. We dropped a “Become a Partner” form right into the Help Center. When someone submits, a Jira issue is created in the right project with all fields mapped. From there it’s pure flow: triage, assign, follow up, measure.
How it feels in practice:
- Visitors stay in the same branded Help Center—no jarring handoffs.
- We capture structured data, not free-form chaos.
- The team works from Jira, where the process already lives.
It’s a small add, but it made the Partner Portal feel like a living part of our docs—not a side quest.
When Documentation Earns Its Keep — Thanks to Scroll Sites
If this story has a theme, it’s simple: documentation isn’t an attic—it’s your front door. We did the tidy-up inside, but Scroll Sites was the moment we switched on the porch light and put a proper sign over the bell. It transformed a collection of smart Confluence pages into a place where people can find, trust, and act on our knowledge —shaping it into a real experience.
Scroll Sites didn’t replace the craft; it amplified it. It gave us a clean stage for the structure we built, a clearer path from “I have a question” to “I know what to do,” and enough flexibility to plug in the rest of our flow—search, journeys, even forms—without breaking stride.
If you’re sitting on a solid base in Confluence, give it a job. Let Scroll Sites be the wrapper that meets users where they are, with a face that feels like you. Open the door, turn on the light, and invite people in. The rest is listen, refine, repeat.
With ❤️ from SaaSJet
Was this helpful?
Thanks!
Iryna Komarnitska_SaaSJet_
2 comments