shazam - 4:06 am on Oct 30, 2011 (gmt 0)

What you are saying is true. I spend a fair amount on adwords myself so I fully understand what you're saying.

This theory does make sense for huge sites with tons of traffic. They can suck an individuals daily budget down quickly similar to a straight media buy.

For lower traffic sites, it just doesn't make sense. A site that gets 1000 uniques/day then gets 2000/day isn't going to kill an advertisers budget.

You can simply watch the ads. What I see are advertisers that are always looking for as much traffic and impressions as they can possibly get. Big advertisers with huge budgets. My little adsense sites are a drop in the ocean in comparison to their daily spend.

I used to also think it was a traffic quality issue. It's not. An increase in high quality traffic has the exact same effect on earnings. It's the jump in traffic that triggers the algo, and on the flip side, the reduction in traffic that sends ecpm/cpc soaring upwards. Any sudden change in the baseline triggers the scrub algo even if the traffic is coming straight from g with a high conversion keyword.

The 'trick' if there is one, is to somehow get steady increases, to somehow raise the baseline. After a month or two of steady traffic inprovement, the scrub algo calculates a new baseline and the earnings go back to normal. The fraud and thievery is at it's highest when traffic fluctuates. Which of course is most of the time.

I still haven't come up with anyway of controlling this, but i have seriously considered coding ad rotators that can avoid these issues by removing adblocks after x amount of traffic. I'm scared of implementing such a thing though, as I'm sure g wouldn't like it, and probably ban my account. I'm sure it would work great as it seems the scrub algo is based on individual adblock impressions.