I'm generally of the opinion, though it is only a hunch, that any scheme requiring a user to actively tip for each site they like is doomed to fail.
I also think that the other problem with tip jar and/or Xanadu-style micropayment schemes is that people are uneasy about building up a surprising total. This is the advantage of Readability/Kachingle-style subscription services.
We’ve made everything about the service extremely flexible and low-key, and shouldn't cause any tense feelings. For example you can cancel Tips (up until they are paid out) if you made a mistake.
Tipjoy basically did this, but they still had the fundamental problem that the user has to remember to click in order for the tip to take place. I just don't think people are wired that way.
I admire the work you've done (I am, in part, working on a for-profit "competitor"), but I just don't think it's the right model. In terms of human factors, Readability is closest to perfect IMO.
If they don't have one already, I expect an opt-out feature before too long. It's the first objection raised to my system whenever I talk to people about it.
In terms of the business model, Contenture was almost exactly what I thought up on my own.