It's also unclear whether it's a causal relation or not - it's also plausible that the startups who happen to be "more mature" from the very beginning are both

It's also unclear whether it's a causal relation or not - it's also plausible that the startups who happen to be "more mature" from the very beginning are both growing faster, and also able to afford an office faster; perhaps that early office doesn't influence anything but is simply a consequence or side-effect of 'better startups'. This kind of observation simply can't distinguish between correlation and causation.