The context
After launching new Free, Team, and Enterprise plans for SonarQube Cloud, the Growth team expected conversion to pick up. It didn't. Over 200,000 customers were on the free plan, but only a very small percentage were upgrading. The assumption was a pricing problem. It wasn't.
Results at a glance
20%
Increase in active organizations
From a single LaunchDarkly experiment
3
Free-plan user types mapped
Surfaced from mixed-method research
The friction
The Growth team was asking: "why aren't users converting from free to paid?". That framing assumed conversion was the right goal. My research surfaced a more specific and actionable question: "how do we enable users with some projects on a free plan to fully migrate to a paid plan?"
That reframe changed what the Product-Led Growth (PLG) team explored next.
What I found
- A large share of users said the free plan already met their needs. So there was no real motivation to switch.
- Many teams were deliberately using free-plan projects as sandboxes for work not directly tied to business i.e., R&D, experimentation, pre-revenue projects. These weren't failed conversions. They were intentional.
- A key pivot: free plan users often had some projects on a Team plan already, as a workaround for budget limitations. This meant the real opportunity wasn't converting free users, it was consolidating split accounts.
- Conversion wasn't primarily a pricing problem. It was a timing, relevance, and internal advocacy problem. Users needed to encounter the right feature at the right moment, and have a way to build a business case internally before being asked to upgrade.
The decisions this enabled
- Shifted onboarding sequencing and campaign timing to target users at moments of natural expansion, not arbitrary upgrade prompts.
- Directly informed the LaunchDarkly experiment roadmap for the PLG team. One experiment drove a 20% increase in active organizations.
Risks and tradeoffs
- Reframe risk: Shifting the question from "why aren't users converting?" to "how do we consolidate split accounts?" was the right call analytically. But it required convincing the team to focus on a specific segment of free plan users first, which isn’t what they were optimizing for.
- Experiment causality: The 20% lift was the strongest result across a series of experiments. But being more deliberate about experiment sequencing and impact attribution across the full program would have made it easier to know which specific lever drove the change.
- Sandbox users: A meaningful share of free-plan users are there intentionally and will never convert. Treating them as a conversion opportunity wastes resources. Identifying and excluding them from campaigns was as important as finding the right message for users who might convert.
What stayed human — and why
The decision about which experiment to prioritize first wasn't purely data-driven. It required judgment about what the team could build quickly, what would be most legible to users, and what would generate the most actionable signal. Research can narrow the space. It can't make the call.
Result: 20% increase in active organizations from a single experiment · Reframe adopted by Growth team · Experiment roadmap directly informed by research segmentation
If I ran this again…
- Instrument free-plan project behavior from day one, knowing which projects were sandboxes vs serious work would have sharpened the segmentation faster.
- Validate the account consolidation finding quantitatively. We surfaced it through interviews and it shifted the team’s framing, but we never ran an explicit experiment to measure its scale.
- Build the internal advocacy angle into the product itself. There needs to be affordance for users to share value signals with their manager before being asked to upgrade.