In finding an answer for "Is there a list of failed/closed betas?", I decided to look through the results to see if there appeared to be a common cause of failure in a beta site.
Based on my analysis, it appears the common cause is a lack of high-quality daily visitors. Is a lack of high-quality daily visitors to a beta site the number one cause of their failure, if not, what is the most common cause?
Here's a link to a list of closed SE beta sites.
Your assertion leads to a bit of a circular argument. You almost have to define "high-quality daily visitors" as those who are needed to make a site work. So, yes, if you don't have the people needed to make the site work, the site wont work.
Looking at the individual proposals, there's been plenty of discussions and study about why the site failed to attract that needed audience. The following had a clear and concise reason for failure…
Artificial Intelligence - Community was mostly people curious about the topic but not enough actual practitioners/experts to ask or answer concrete, interesting questions. I wrote about it extensively here: No Artificial Intelligence in Area 51.
Electronic Gadgets - We pushed "Gadgets" hard (too hard) in our early enthusiasm to launch a site which seemed like such a perfect fit for the Stack Overflow community. The result was a broad base of users about a nebulous subject. The resulting site was a pop culture-y, light treatment of the subject where no one felt like an "expert." No core community formed around the subject and the site floundered irrecoverably.
Game of Go - Still too niche for our largely-technical, US-centric audience. A proposal before its time.
Atheism - The subject had quickly run it's course and simply had nowhere else to go leading to an uncharacteristic steady decline early in it's life cycle. But even beyond issues of traffic and the size of the subject and audience, most of the early users stated sincerely that they lost interest as the questions became uninspired and "painfully uninteresting."
Sexuality - Similar to AI, mostly an audience interested in the subject but not enough studied in the field of expertise. They made a good run of it, but the site became mostly cut-and-paste content from Wikipedia.
How Things Work - The site became little more than a mirror of many other popular sites on the subject.
Freelance Workers - Too many "students" (those looking for work) and not enough "teachers (those expert on the subject).
Guitars - Too niche. The interest was merged into the larger base of expertise of Music Practice & Performance.
