The number of registered accounts means not too much and it's totally useless to make a measure of their reality.
It's not difficult to imagine guys like these:
http://www.gamespot.com/articles/guy-spends-22-500-on-star-citizen-has-no-regrets/1100-6426321/
spending all days doing whatever they can to support the game, including registering new accounts with or without pledge associated.
That, and a few years doing that in a daily basis and you can transform easily your population from a very low number to 850k.
There are some evidences that works as a proof of that... like the own aggressive and risky marketing attitude of CIG, the # of watchers of their live streams and weekly shows barely growing, the # of forum users commenting also do not growing, their pools receiving very few votes (and you can even vote with an account without any pledge associated), and so on... but one of the major evidences is their leaderboards which, for the sake of reporting a leaderboard to motivate people buying more expensive ships to win and be a famous community member due that, ends revealing an important data, showing everyone who played in their maps, how many in total, for how long, etc... since AC was released (not intentionally probably, since the objective seems to be just to motivate people to pledge more dollars, not to report sales status):
https://robertsspaceindustries.com/community/arena-commander/leaderboard?mode=BR
The link above shown the most popular map, and it only has 11k users. If you consider that a lot of users are the same individual with different users (and different ships),
this number is even minor. Less than half of that played for more than 10 minutes, summing all tries.
But since AC released, they probably earned about 650k new registered users. Had already about 150k registered users before its release (with Roberts confirming in the past that about 40k never pledged, but ignoring to tell that the rest still included many individuals with multiple accounts).
Meaning that you have to believe that neither the first pledgers, probably the most excited and thirsty fans nor the 650k who came later had interest to play the game released so far, which is extremely difficult to believe from people who pays and "subscribe" to be part of an alpha/beta.
It's pretty much like to believe in fairy tales or to live in a deep level of denial.
The number of people playing do not grow in any acceptable/reasonable proportion when compared with the counter of registered users of their website.
For those who consider this project as a "massive" success, or that the team is doing the right thing that everyone loves because the ultimate proof is that their counters grow in popularity, no matter what, with new 1k/2k individuals "registering" every single day, Spock would say:
"His pattern indicates two-dimensional thinking."
Based on that, we can understand that the surprise effect (lack of that) is not a problem, considering that the huge majority does not listen to them or play their game, whatever quantity of people they really have (I imagine something around 100-150k).