This baby picture from the Planck satellite shows the fluctuations from perfect uniformity, with the red “hot spots” corresponding to the underdense regions and the blue “cold spots” corresponding to overdense ones: the ones that will grow into star-and-galaxy-rich regions of space. The Universe required these
But to go from those first stars to a dark energy dominated Universe — one where no new structure will form if you’re not already gravitationally bound — that’s not such a big leap. It takes only about 7.8 billion years from the Big Seed for the Universe to begin accelerating, meaning that if the initial fluctuations were much smaller, so that we wouldn’t have formed the first stars until, say, ten billion years after the Big Seed, the combination of small fluctuations with dark energy would ensure that we’d never get stars at all.
If it weren’t for "dark energy" — if all we had was matter and radiation — then in enough time, we could form structure in the Universe no matter how small those initial fluctuations were. But that inevitability of an accelerated expansion gives our Universe a sense of urgency that we wouldn’t have had otherwise, and makes it absolutely necessary that the magnitude of the mean fluctuations be at least about 0.00001% of the average density in order to have a Universe with any notable bound structures at all. Make your fluctuations smaller than that, and you’ll have a Universe with nothing at all. But elevate those fluctuations up to a “massive” 0.003% level, and you have no problem getting a Universe that looks just like ours.