Best practice -- and I'm shamelessly stealing this lesson from Joel Spolsky -- is to grandfather SaaS accounts in indefinitely if you raise prices. This a) lets you get a nice bump in sales if you announce the incoming price increase (no reason you can't do the A/B test quietly but announce the conclusion that prices are going up "loudly"), b) rewards people for taking the risk on you "early", and c) incentivizes them to keep their account active because if it lapses then they lose their super-sweet not-available-anywhere deal.
If I were to redo a pricing scheme in such a way that some customers would benefit from switching, I'd switch them automatically if the new pricing dominated their existing plan ("More quota for less price!") and tell them to make the call if it weren't obvious ("More quota for less price but this new plan doesn't have a particular feature that you may want to have in the future.")
What do you do when your old pricing scheme is no longer profitable?
There's some costs which are difficult to anticipate prior to launch. One is ongoing customer support, and it is entirely possible to price below profitability if you get a lot of pathological customers in on the ground floor. If you've just got a few and the business is fundamentally sound, consider it a marketing expense and let them get weeded out by attrition. If you're like Spreedly or Chargify and discover that a lot of the early adopters in your space are toxic, then it's time to have the "Look, we're not in the business of subsidizing you" discussion. I'd have the messaging for that discussion stress that you'd like to continue offer the old terms but circumstances are tying your hands, and try to be generous on e.g. giving 6 months or so of lock-in to the old prices or assistance in moving them to a provider which is more suited to their needs. But honestly, at the end of the day, if the choices are a) not making payroll for my employees and b) ticking off a bunch of pathological freeloaders, I know what I'd pick every single time.
I think so long as you don't try and cover it up and offer those who discover the choice between the new/old pricing, it shouldn't cause any problems.
that's just because you aren't big enough. Amazon has been caught testing prices previously, and there were a lot of upset people. Noticing the test isn't what caused the outrage.. wait until you have thousands of people who find out they paid more than another person for the same item (ie in their minds: were ripped off). Then you'll see how consumers feel about your test.