Is it enough to improve answered KPI (currently 90%) to get it out of beta phase? (and keep other KPIs at "Excellent")
I don't think that we should try to improve that stat at the expense of other things. First, I don't think that 89.7% or whatever is too low. Second, I think that the unanswered questions that we have is a result of more of a lack of international experts. We have lots of people answering US questions but very few who can answer obscure questions about sub-Saharan Africa, Southeast Asia, or Latin America.
I actually think that having some unanswered questions is good. It means that if we do pick up an expert in a relevant area, there is a lot of room for that person to answer questions.
Note that the GameDev site had almost the opposite statistics. Most of its questions are answered and it has many answers per question but it would have fallen short on views, number of questions, and avid users. Yet it graduated. I suspect that "critical mass" is as much art as science.
What does a critical mass mean? Having excellent for all KPIs except for one (e.g. CodeReview site stats at the end of beta)?
In general, it means that they feel that the site has matured into a sustainable state. I don't think that either of the answered stats is the real problem for this site. The larger problem is that for several years, the site wasn't meeting the views and questions standards. Because this site is heavily related to the news, one worry is that when the news dies down, so does the site. The site was much slower before the presidential election, even years after starting.
Just looking at those stats, there is a solid argument for allowing the site to graduate. The greater question is if the stats are maintainable. A year from now, Trump's every action won't be new anymore. The basic thrusts of his legislation will have been established. If he pauses travel from a country, there will be a well worn judicial precedent to determine what happens.
The current changes are newsworthy and politics worthy. When the administration moves into a more steady state, will they still be? Will the issues of the midterm elections drive questions the same way?
Will other countries step in and provide question fodder at the same rate as the United States does? A site that answers questions regarding a variety of countries would be more stable than one that concentrates purely on the United States.
Will we be able to get back to the more theoretical parts of the site? E.g. talking about the Condorcet criterion and other voting measures. That is less news dependent, but it also seems to have been lower volume.
I think that there is more risk about graduating the site early than leaving it in beta. The public beta allows for more participation than a graduated site has. Many who currently have privileges would lose them on graduation. Graduation does give a site redesign, but I don't know that the generic beta design is hurting the site.
Graduation is certainly a site milestone, but I don't know how much it matters. Particularly with an older site like this. I don't know that there are many who might participate if only the site were graduated.
If they graduate it, great! If not, that can always change.