Free-tier math: when the funnel eats the margin
Three companies published their conversion curves. Two of them should not have shipped a free tier.
October 2, 2026
The shift didn't arrive as a single announcement. It showed up in headcount plans, in the quiet retirement of a dashboard, in a comp plan that stopped rewarding the thing everyone had spent three years optimizing.
We spoke with eleven operators running revenue at companies between $8M and $140M ARR. Every one of them described the same pattern: the motion that got them to Series A stopped compounding somewhere around the second sales hire, and the fix was structural rather than tactical.
The uncomfortable part is what it implies about budget. Teams that made the transition cleanly moved between 20% and 35% of spend out of top-of-funnel and into product, lifecycle, and partner surfaces — before the pipeline data justified it.
What follows is the org chart, the numbers behind it, and the three failure modes that showed up in the teams who tried the same move and reversed it within two quarters.
The first failure mode is treating this as a marketing reorg. It isn't. The buying committee changed, which means the qualification criteria, the demo, and the first ninety days of onboarding all have to change with it.
The second is measurement lag. The new motion reports later and messier than outbound did. Teams that kept weekly pipeline reviews as the primary forum reverted almost immediately, because the forum could only see the metric that was disappearing.
The third is the comp plan. If the account executive is still paid on net-new logos sourced from a list, no amount of strategy deck will move behavior. The teams that succeeded rewrote quotas first and org charts second.
The conversion curve told the real story: 4% of free users would ever convert, and the CAC on those conversions was 3x what we were charging.
The conversion curves from the three companies, including the math on when free tier works and when it doesn't, are in this week's pricing analysis.
