Lapizuli - 9:04 pm on Jan 23, 2011 (gmt 0)
I'm wishy-washing about opting in even to try it out...doing fine now, afraid to rock the boat. I'm pretty superstitious about testing this kind of thing - not only is it hard to control for variables, but also because I think about the human beings behind the campaigns and figure it could do long-lasting damage.
Here is my thinking:
My ads are pretty much always relevant to the context. My earnings are pretty steady, overall. This suggests to me that I have some consistent advertisers.
So if my advertisers have to compete with higher-budget external-to-niche advertisers to display their ads, that would initially cause a surge in CPC. But then the small-but-relevant advertisers will wake up to their quickly depleted budgets and re-evaluate whether it's worth it and, realizing a pet grooming tool simply can't compete financially with debt consolidation loans, or whatever, simply drop out of the pool for my pages, sending the CPC on a downward slope.
Of course even if that doesn't happen, then my CTR, nothing special but basically healthy right now, might decline. Which might be a big deal, or it might not, depending on how the bidding algorithm adjusts.
I wish Google would share data about what kinds of sites would do better to enable interest based ads. They're focused on averages, but we're not. My instincts say there's a big difference between how, say, fiction content would perform with IBAs and latest-cool-gadget content. Does the data support this?