mate i'm doing well on the free trials at the moment. may i suggest if you are using search:
1) Keep it small at first and build your way out from the center. think of your buying market as a target with the buyers at the bullseye, shoppers in the ring around and browsers on the outside. that in mind start just bidding on the brand name and very close derivatives for a bit until you've got the ad->lp->conversion bit down. its way easier to convert the brand names as peeps are in buying mode already and you're removing a lot of uncontrollable factors while keeping control of the important bits. get some lp split tests going too. i added 3 words to the header of one of mine and ctr went from 43%->78% overnight and stayed there. conversions upped proportionately too. while it may not be as cool to just work with 1 keyword as it is to say bomb 1000 out there it is infinitely more easier to control and once your base is steady and converting nicely (my base hits 300%+ daily now) you have a load of steady funds to expand with and the confidence to make small changes to your ads and lps to hook the buyers and browsers in.
2) Start with Yahoo or Msn not Google first. once you are stable move to G and reap the rewards. G definately doesn't convert as well as the other two but a small tweak here and there should be enough to bring each adgroup nicely into $.
3) throw up a large 100+ adgroup into yahoo or msn with low bids to guage keyword traffic levels without spunking money, just for a night. when you see a keyword getting lots of impressions i optimize it to its own adgroup with new ads and up the bids to place 2-4 for a night or two to see if its converting and work from there. cut the rest.
4) daily spend levels are your friend!
5) once you're all set and you've got a nice list of converting keywords bid high in G - set positioning to 1-3, burn a load of money for one night, then enjoy half price clicks for the next month curtesy of the 20-50% ctr you gained that night. don't try this until your adcopy/lp is tested first. otherwise the spend will be a waste.
i'm new to this but so far after a couple of blowouts on the content network the above strategy is really working out for me on search.
good luck
oh and just thought if you're not already doing so get professional lps made up - it makes a big difference and if you have the time learn how to make them yourself - then editing them for split tests/new copy is a 5 minute job not a 3 day headache