Atlassian Team ’25 – Discover and Indeed share hard-won lessons on migrating from data center to cloud
Discover and Indeed share their experiences of migrating thousands of users and millions of items to Atlassian Cloud instances of Jira and Confluence.
Migrating to the cloud sounds like it’s just a tech decision. In reality, it also forces a reset on how the organization operates. When systems like Jira and Confluence underpin thousands of daily workflows, the stakes are far higher than infrastructure.
That’s the message from US financial services company Discover and global job site Indeed at Atlassian Team ’25, where senior leaders from both companies took the stage to share the complexity behind their recent cloud transitions.
Together with ISOS Technology, an Atlassian solution partner that helped facilitate both efforts, they detail a pragmatic roadmap for cloud migration, grounded in thorough planning, meticulous rehearsal, and continuous communication.
The scale and the stakes
For Indeed, the migration involved over 15,000 users on Confluence, 11,000 users in Jira, 4 million Jira issues, and more than 75,000 pages of content. At Discover, it was 12,000 users and more than 2 million issues, alongside 1.5 million attachments. In both cases, the teams had to reconcile massive data volumes, years of underutilized servers taking up more space and resources than needed, and institutional dependencies on Atlassian tooling.
Eduardo Lessa, Senior Manager for Atlassian Products at Indeed sets the scene:
We decided to migrate because Indeed is adopting a SaaS mentality — not just to be on cloud, but to stop allocating our internal resources to managing infrastructure
Alex Jones-Duffey, Jira Product Owner at Discover, echoes that rationale, explaining:
We were trying to get a more modern infrastructure and take advantage of vendor-supported patches, security updates, compliance features like team-managed projects and analytics.
Long before the first line of data moved, both teams invested heavily in documenting their environments and scoping their projects.
Jones-Duffey elaborates on this:
We spent several weeks documenting literally everything — fields, configurations, issue types, processes.
We broke everything into pillars so we could organize the massive scope of work.
The investment paid off during a week-long onsite workshop with ISOS.
Jones-Duffey continues:
We walked through every single item on our list — what’s current state, what does it look like in cloud, can it migrate? Some were easy.
Others required recommendations from ISOS or follow-up from Atlassian.
The effort also extended to a deep inventory of business processes.
She notes:
Organizations evolve. Some processes change or disappear.
We tried to minimize what we brought over and took the opportunity to de-scope.
At Indeed, the decision to prioritize speed took the project down a different route.
Lessa recalls:
We had to choose: optimize before migration or after.
We decided to archive inactive projects and standardize later. We didn’t want to disrupt our users — we wanted the migration done in 48 hours.
It probably goes without saying that the audience collectively took a deep breath at this. The tight timeline meant rehearsals were critical.
Lessa continues:
We timed every single aspect of the migration in each rehearsal.
By the time we hit cutover, we knew exactly how long everything would take.
Governance and change management
Across both migrations, the importance of governance and change management was hard to overstate.
Lessa recounts at Indeed that there was a strong program management group, weekly steering committee meetings, and clear executive alignment.
He says,
When risks or decisions came up, we could clear them quickly,
At Discover, the team structured their work into four streams — business-as-usual, migration planning, change management, and incoming business needs. That helped to assign the right skills and avoid death by committee.
Jones-Duffy says:
Not everyone needs to be in every meeting.
We focused our time deliberately.
That structure extended to Isos Technology as well.
Nick Sommerfeld, Manager of Solution Engineering, says:
We’re not a revolving door,
“We bring the right resources for each phase — and everyone is held accountable.”
Practice and post-migration support
One clear lesson was underscored by both companies — successful migration is as much about post-move experience as it is about data movement.
Indeed made sure there was a team available 24/7 for five days straight to catch and resolve any issues quickly after the migration, anticipating global time zones and ongoing configuration work. Discover, meanwhile, held daily office hours for two weeks and conducted multiple dry runs — ultimately eight, up from an original plan of four.
Jones-Duffey says:
We had a runbook for the cutover.
One surprise we uncovered was that attachments alone were taking 16 hours to migrate. So we split that out and ran it the weekend prior. That saved us a third of our migration time.
Importantly, both companies prioritized user experience over technical wins.
Lessa sums it up:
Full parity was our goal.
On Monday morning, we needed users to log in and just keep working.
More than just infrastructure
Since migrating, both organizations have begun unlocking cloud-specific benefits — some expected, others pleasantly surprising. Discover has taken advantage of the data lake and Atlassian Analytics, noting that it took lots of time to configure extracts, tables, and BI tools. Jones-Duffey says: “Now, data access is much more efficient.”
Lessa adds that feedback on the new UI has been overwhelmingly positive:
It’s more intuitive.
It feels more like a social platform. There’s excitement around features like OKR [Objectives and Key Results] goal tracking and Loom AI, which were just announced.
Jones-Duffey remarks that team-managed projects have also gained traction due to the specialized ways of working by business partners. Both speakers emphasize that successful migration demands more than tooling — it requires organizational alignment, meticulous rehearsal, and empathetic communication.
Sommerfield agrees with this, noting:
If you’re working with a solution partner and don’t get a detailed project plan — not just for the cutover weekend but the entire timeline — your risk skyrockets.
And while some enterprise IT leaders may consider change management a soft skill, the panel disagrees.
Lessa said:
Change management isn’t just about systems — it’s about people. From communication to training to post-support, it was critical.
Despite different starting points and constraints, both Discover and Indeed describe their migrations not just as technical shifts, but as organizational milestones.
As Sommerfield comments:
It’s not just a migration — it’s a transformation.
Yes, you’re moving data. But you’re also unlocking new capabilities, shifting how your teams work, and setting the foundation for what comes next.
Jones-Duffey agrees:
When you get to the other side, you’re not just up to date — you’re ahead.
My take
During his keynote, CEO Mike Cannon-Brookes framed migration from data center to cloud as the first step in adopting Atlassian’s System of Work. His comments on the move to a modernized architecture are clear — it can be a game changer, but it’s not a lift-and-shift exercise. Instead, he says, “It’s like a giant leap from 2015 to 2025”.
What stands out in this session is not just the technical complexity of the projects, but the transparency of the participants. Rather than simply offering polished lessons learned, Atlassian and its customers approach this conversation with candor — sharing practical advice, hard-won missteps, and the courage it takes to talk honestly about what change really looks like at scale.
READ the latest news shaping the data centre market at Data Centre Central
Atlassian Team ’25 – Discover and Indeed share hard-won lessons on migrating from data center to cloud, source
