I agree with your general analysis OP. However your math is bad. 600 paying vs. 200 not paying implies 25% will stop paying according to that (obviously biased) poll, not 33%.