+3.7%
App conversion
+7K
Paid installs
20+
Enterprise interviews
Context
Atlassian Marketplace is a $6B+ ecosystem with 6K+ apps from 2K+ partners serving 300K+ customers. The Marketplace connects enterprise teams with third-party apps for Jira, Confluence, and other Atlassian products. Growth had plateaued as the existing discovery experience relied on basic keyword search and editorial curation.
Problem
App discovery was underperforming. Customers struggled to find relevant apps among 6K+ listings. Conversion from browse to install was declining. Key issues:
- Search results didn't account for user context, team size, or existing product stack
- Recommendations were static and editorially curated, not personalized
- No intelligent discovery path for AI agents and new app categories
- Funnel drop-off between app page views and installs was high
My Role
Led growth loop strategy for Marketplace discovery. Conducted 20+ enterprise customer interviews and applied AI-driven insight synthesis to shape strategy and experiments. Coordinated with engineering, data science, design, and partner ecosystem teams.
Decisions & Trade-offs
Parallel experiments vs. sequential
Ran 5 ML/AI experiments simultaneously (recommendations, personalization, sponsored ads, funnel optimization, AI agent-assisted discovery) rather than sequentially. Higher coordination cost but 3x faster learning velocity.
AI agent discovery strategy
Shaped strategic vision for onboarding AI agents, services, and standalone third-party apps on Marketplace. Aligned with company AI productivity goals. Projected +$25M ARR uplift (+20%).
Monetization model for AI apps
Defined consumption-based and value-based pricing models for AI agents and apps, delivering ~$20M ARR (16%) across 6K+ apps. Chose hybrid pricing over pure seat-based to accommodate AI usage patterns.
Metrics & Results
- App conversion improvement: +3.7% within 6 months
- Paid installs driven by ML experiments: +7K
- AI agent/app monetization strategy: ~$20M ARR
- Projected ARR uplift from AI agent platform roadmap: +$25M
- Enterprise customer interviews: 20+ conducted to validate strategy
Links
Learnings
- Parallel experiments compound. Running 5 experiments simultaneously taught us more in 6 months than sequential testing would have in 18. The interaction effects between personalization and funnel optimization were unexpected and valuable.
- Enterprise interviews shape AI strategy. The 20+ customer interviews revealed that enterprise buyers wanted AI agents integrated into existing workflows, not as standalone products. This reshaped our onboarding strategy.
- Pricing follows usage patterns. Consumption-based pricing for AI agents aligned incentives better than seat-based — customers paid for value delivered, which drove adoption faster.
- Platform plays compound. Marketplace growth doesn't come from optimizing one funnel. It comes from creating a flywheel: better discovery → more installs → happier partners → more apps → better discovery.