I'd say release the beta for free to a super small group of trusted people who have already played the game, and have shown themselves to be good feedback givers, or people local to you that you know personally. The idea here is that they're people who will actually take the time to feed back, and that towerclimb doesn't get released before you intend it to.
also, I agree with newt. make it uber cheap, and try to sell more copies
for instance, I think you will get more than twice the sales at $5 than you would at $10
for a little perspective
plants vs zombies is $10 on steam ( http://www.the-nextlevel.com/media/pc/p ... bies_2.jpg )
braid is $10 on steam ( http://gamernode.com/upload/manager///E ... 877407.jpg )
ragdoll kung fu: fists of plastic on psn is $10 ( http://www.psu.com/media/rag-doll-kung- ... c-ss-2.jpg )
castle crashers on psn is $15 ( http://gaming.icrontic.com/uploads-gami ... encap9.jpg)
[upon it's announcement, many forum posts can be found saying (I would have gotten it if it were $9.99, which once again goes to show the difference $5 makes psychologically]
graphics don't make the game, but books are often sold by their covers.
the three I mentioned also happen to be great games.
the last screenshot I saw of towerclimb was in an uberretro, almost c64 style of gfx.
it might have changed, and I don't doubt you could get the game a decent audience even with those gfx, but at $7 or more, I think you're going to greatly reduce the size of your audience
I think the figures you mentioned would be good to be honest, for the final product, not the beta.
it makes it an easy impulse buy, and also easy to recommend if it turns out to be fun.