At least for us, it saved us a lot of discussion when we could say "is this particular feature helping us grow?" and if the answer as no, we would move on.
But I do think, even at the earliest stages, it isn't hard to gauge interest by showing demand. Sign up pages are simple enough to do.
Firstly, getting traction by number of users is much harder. But every customer you sign on implies huge growth for your company in terms of revenue. Take Palantir for example; afaik, their initial massive growth came from a single, but huge, government contract.
Secondly, growth will never be a smooth 10%/week type of curve -- more like an irregular staircase; but taken over the long-term (years), it might get you higher. So taking a granular snapshot of the early beginnings might show no growth at all.
Thirdly, there won't be many early signups. Simply because the volume is much lower.
Having said that, zooming out on the macro scale (years), your advice still holds (growth > everything) -- but it doesn't apply to the YC application process :)