Why We Bet on Forge Before the Market Rewarded It (and What the Install Numbers Say Now)
I want to share something we've been watching quietly for the past several months, because I think it's useful for anyone in the Atlassian partner ecosystem who is making (or about to make) architectural decisions for their next Marketplace app.
This is partly a data story. But it's also a story about making a hard call under uncertainty, before the market gives you any signal that you're right.
The context: we launched into a shrinking category
Before we released Mria CRM for Jira, we started tracking install counts across the entire CRM category on the Atlassian Marketplace. We started this even before our launch, partly out of competitive curiosity, partly because we wanted to understand where the category was heading.
What we saw was striking - and, honestly, a little unsettling. Every established player in the category was losing installs. Not dramatically week-by-week, but steadily and consistently. One of the once leading apps dropped from roughly 2,400 installs to under 1,650 over the period we tracked — a decline of around 32%. Another one went from ~880 to ~680. The overall category was contracting.
We launched in September 2025 with 8 installs. Today, Mria CRM is at 375 (as of today), and has grown every month since launch. In a category where every competitor was going down, we were the only app going up!
I'm not sharing this to boast. I'm sharing it because I believe it points to something real happening in the Atlassian ecosystem right now - and because the architectural decision we made before we wrote a single line of production code is, I think, the main reason we ended up on the right side of that trend.
Below you can see just s short sketch of the recent days:
The decision that didn't feel obvious at the time
When we started designing Mria CRM, the natural path was Connect, or a hybrid Forge-Connect approach. Connect gives you more flexibility. You control your infrastructure, you integrate with anything, you build features faster when you hit Forge's constraints, and Forge has real constraints.
Anton Storozhuk (my co-founder) pushed hard for a different path: pure Forge, no external egress, fully eligible for the Runs on Atlassian badge. I'll be honest, I wasn't immediately convinced. We argued about this a lot.
Here's what the "rational" case against Forge looked like from where I was standing:
- Forge meant deliberate limitations on what we could build and how fast we could build it
- Some functionality had no established patterns yet — we'd be figuring things out as we went
- Our competitors had years of head start and thousands of installs; we didn't have the luxury of choosing the slower path
- The benefits on the other side of those constraints felt abstract and distant
There were real moments where optimizing for speed and short-term convenience would have felt like the smarter founder decision. My partner's conviction was that the Atlassian ecosystem's future would move toward trust, security, native experience, and platform alignment. That Forge wasn't just a technical requirement but a signal about where the platform was going.
So we kept building.
What "Runs on Atlassian" actually means in a sales conversation?
Six months of demos and trials teaches you what customers actually care about. Here is what we hear, over and over:
"Does this need its own security review from our IT team?" When you're Runs on Atlassian, the answer is: Atlassian has already done that review. Your app qualified by being built the right way.
"Where does our CRM data sit?" Same region as your Jira. If you've pinned to EU-West, your Mria CRM data is in EU-West. Automatically. No configuration required on the customer's end.
"What happens when we move our Jira data to a new region?" Your Mria data moves with it. Forge handles it.
"Can you send us your security documentation and vendor risk questionnaire?" The Runs on Atlassian badge significantly shortens this conversation.
For customers in finance, healthcare, government-adjacent industries, or simply in regions with strict data sovereignty requirements (let's be honest that this is a substantial portion of the Jira enterprise market!) these are not nice-to-haves. They are procurement blockers. An app that can't answer these questions clearly often doesn't get past IT review, regardless of how good the product is.
What Runs on Atlassian does in practice is make the answer to those questions visual, immediate, and credible. Customers don't need to read technical documentation or trust a vendor's self-reported security posture. They see the badge and they understand: this app lives inside Atlassian's infrastructure, with the same data residency and compliance guarantees as the rest of your Atlassian stack.
Why we think the broader category is contracting
The decline we're seeing across established CRM apps on the Marketplace isn't random. I believe it reflects a few forces converging:
Atlassian announced Connect deprecation for Q4 2026. Enterprise customers managing long-term risk are beginning to factor this in. An app with no clear Forge migration path is carrying technical debt that sophisticated procurement teams are starting to notice. Or, what I notice even more regularly, including in the CRM category - the migration-to-Forge path is not followed or delivered at all!
Enterprise Jira customers are becoming more sophisticated about data residency. The answer that worked a few years ago for us in Alpha Serve ("our data is on secure servers in AWS") is not enough anymore. Specific regions, specific compliance certifications, and clear ownership of where data lives are now table-stakes questions in many enterprise deals.
The Runs on Atlassian badge gives procurement teams an actionable signal. Before the badge existed, a customer would need a technical deep-dive to understand an app's hosting and security posture. Now they can see it at a glance on the Marketplace listing. This matters especially for larger organizations where the person evaluating apps is often not the person who would do the technical due diligence.
Teams living in Jira want CRM that lives in Jira, not CRM that integrates with Jira. There's a real difference in the experience, and customers increasingly feel it. The strongest point for Mria CRM is a deep and 2-way integration for the customer context and revenue into existing Atlassian workflows, and a recognizable UI that speeds up the adoption.
Our peers (I prefer not to say competitors as we deliver too different solutions) were built before these forces became dominant. They were making the right calls for the environment they were building in. Since then, the environment has shifted, and apps that got to Forge early are now benefiting from that timing.
The honest trade-offs, because this post would be dishonest without them
Forge's constraints are real. There are features our competitors have that we haven't shipped yet, not because we haven't tried, but because building them the Runs on Atlassian way is genuinely harder.
External integrations are limited compared to what a Connect app can do. Forge's storage has improved significantly but is still not a traditional database, and architectural creativity is required to work within its runtime limits. Some things that would have taken a week on a traditional stack took us substantially longer.
If your product's core value proposition is deep integration with many external SaaS tools - email clients, dialers, marketing automation platforms - Runs on Atlassian may genuinely conflict with what you're trying to build. That's a real consideration, not a small one.
Separate pain for me personally is also the "attachments" issue. And there are more others.
But if your product lives in Jira and serves Jira users, and your customers increasingly operate in regulated industries or have data residency requirements: build it the Forge way. Accept the short-term constraints. The market is moving in this direction, and based on what we're seeing in our own category, it's already rewarding apps that made this choice early.
What I'm taking away from this as a founder
The data is gratifying. But the part I keep thinking about is how the decision felt when we made it.
It didn't feel like an obvious strategic win. It felt like choosing the harder path without clear evidence that it was worth it. Anton's conviction wasn't backed by install-count trend data, as that data didn't exist yet. It was backed by a read of where Atlassian was going as a platform, and a belief that customers were raising their bar faster than most partners realized.
The thing I'm learning: data matters. Experience matters. Market analysis matters. But in moments of real uncertainty, when you're making a foundational architectural call before the market has validated it, sometimes the biggest advantage is the willingness to trust a conviction before it's confirmed.
Most important decisions are made without certainty. That's the part nobody tells you about building.
___________
For partners considering this path
The Atlassian Marketplace is not static. Enterprise customers are raising their expectations for what they'll install. Teamwork Graph is our today, not tomorrow. Atlassian's deprecation of Connect is a clear signal about where the platform is going, and the Runs on Atlassian badge is Atlassian's way of helping customers identify apps that are already there.
If you're starting a new Marketplace app: build on Forge, aim for Runs on Atlassian, and treat the constraints as design requirements rather than blockers.
If you're on Connect and evaluating migration: the friction is real, but so is the risk of staying put. Enterprise customers are starting to factor platform longevity into their vendor selection, and that trend will only accelerate as the Q4 2026 deprecation date gets closer.
We're a small team. We went live something more than six months ago. We're growing in a category where every established player is shrinking. The architecture was the strategy.
Happy to answer questions from anyone working through the Forge decision. It's a real tension and I don't think there's a universal answer. But we know what it looked like for us. Would be happy to chat! DM on LinkedIn!
Anna Odrynska, Co-founder, Mria Labs Inc.