Because there are humans in the loop, and "Google" isn't a monolithic all-seeing entity. Sales guys, devs, etc. are rewarded on a quarterly (or yearly) basis, and short-term gains are prioritized over long-term effects ( I mean, how _do_ you measure the fact that, say, after getting 9 months of pointless clicks, the advertiser decided to cut their Adwords budget by 30%?)
So, someone runs an experiment (replace the obvious radio button with a hidden drop-down box, say), the clicks on the advertiser's ads go up, so the experiment increases revenue week-over-week by, say, 3%; and voila, everyone involved gets a pat on the back and a nice phat bonus come the next eval cycle, and maybe a promotion too.
Oh man, this is _the_ problem with metrics. We must find a way to measure and demonstrate aversion (instead of conversion).
They _might_ lose a customer — I suspect that they've determined that enough people either simply don't review their results closely or increase their spending to make it a net profit for them to do this. That would fit with the general pattern of hubris you see in many of their decisions where they've been coasting on their old reputation for at least a decade and don't seem to be fully aware of just how much loyalty they've sacrificed.
An end user might say 15% is pretty good and don’t touch it.
Google just earned 5% by giving you worst ROI.
If you start comparing with other ad providers you’l you might notice you’re being screwed. But low spenders rarely do proper checks.
I really don't get this part. What's Google's interest in pointless clicks? Short term they could get some extra charge for those, but if they don't convert they lose a customer. On the other hand, for quality clicks they could just charge 5x the price and nobody would complain.