How convective clouds in the trades self-organize into mesoscale patterns is not well understood. Using airborne observations from the EUREC⁴A campaign, this study shows that the size distributions of boundary-layer thermals and cloud bases follow a double-exponential law. Analytical calculations and statistical simulations show a complex interplay between thermals and clouds that constrains the type of mesoscale cloud pattern that emerges as well as the total cloud fraction. The resulting framework suggests an intriguing path forward for representing these thermals and clouds in weather and climate models.