This is perhaps the most valuable observation of this entire thread. So few people have let non-performing ads run longer than 2-3 weeks that a lot of what we think is "data" is incomplete. I am very guilty of this myself and am beginning to rectify it.
That said, two of my book ads seem to convert well enough that I am more than happy to pay .20 cents per click (bidding in the 30's). So I am still torn about whether I should bid really low on keyword ads. Given that I now intend to let them run longer, I could go quite a while without results before I figure out whether my action thrillers really do require me to bid in the .30's. I know the romance folks almost all have to bid high to get results.
I really don't know the answer.
All I know is that I can't control what Amazon know and I don't, so I don't try to second guess. I only have my own data that I've posted and garnered over quite a number of months. I start by knowing I only have 70% royalty ($2) to play with for the ACOS and a small % of of actual page reads income. I did a previous post how I calculated this after working it out as a percentage of AMS sales to non-AMS sales rather than considering them all attributed to AMS, which I would consider it like fools gold to apply all page reads.
From the data I've posted, I don't need to know what Amazon are upto, or where I sit on the pages. I can quickly work out from overall results or individual campaigns how increased bids would bite into my ACOS if I didn't increase sales and increasing bids only guarantees early pages not clicks and sales, but it ensures higher costs. Regardless, my click to buy ratio would likely stay the same (Think on that a moment). It wouldn't take much of an increase across the board to soon get to a 70% + ACOS, so I always keep that in mind rather than starting from high bid pricing to get impressions that I know will end up producing a loss at my sale price of $2.99 as many have shown by posting their data on here, albeit from short term campaigns.
As an example, If my average bid cost is 8c (Which means bidding more in some instances), then I have a 25-1 click rate to buy, to eat up all my royalty.
If you apply that to my data on the previous page, in relation to % clicks to sales, then it's obvious I will make a profit @ an actual cost of 6 - 8c, so 15c bids are not out of the question, so long as the don't cost me that.
If however, I bid 50c and I'm charged say 25c then it translates to a 8-1 click rate to sale to eat up my royalty. So the way I look at it is that if I don't want to make a profit, I should bid 50c because I am not getting that ratio of clicks to sales for it to work. God forbid if I bid 52 cents and I was charged 50c, which would mean it would take only a ratio of 4-1 to eat up my royalty.
I'm tired so I hope I have the math right and that I am making sense. Bear in mind that I am shy of bidding higher as I didn't get the sort of results in increased sales when I increased bids as the previous post to this enjoyed to increase(or is that decrease) my click to buy ratio for the better. Also note I have no series as they are all standalones.
You can only determine your click to buy ratio with a long term campaign. For me, that ratio determines my bid price and not the pack chasing a front page. I still get sales from ad placement down the pecking order, just not as many as if I'd bid higher.
There is a contra argument to this as a strategy, but this post is too long already.