Last issue, I explained how allowing auto-bidding at Overture means that pricing has to be based off of max bids, rather than actual cost paid. As a result, there are some serious "cost gaps" that can occur. The situation still doesn't sit well with some people, and Steve Harrison is one who is proposing an answer. I'll explain what I understand him to be suggesting, as well as why I don't think it ultimately works.
Steve proposes adding a rule to Overture's logic where bids would only be raised if a gain could actually be made past the person above you. So let's say you have this situation:
Site - Max Bid A - .30 B - .20 C - .10
It works out to this cost:
Site - Actual Cost A - .21 (.01 above B's max bid rate of .20 ) B - .11 (.01 above C's max bid rate of .10) C - .05 (max bid of .10)
Steve would have the logic then work to look at all the listings and determine whether auto-bidding would allow a site below to move past a site immediately above it. If not, then auto-bidding would not take place. In other words:
Site - Actual Cost A - .07 (has max bid of .21, keeping it above B's max bid price. Since there's no way B can automatically rise past this, B's auto-bidding doesn't happen) B - .06 (has max bid of .11, keeping it above C's max bid price. Since there's no way C can rise, C's auto-bidding doesn't happen, so "cost gap" is closed) C - .05 (has max bid of .10, but auto-increase doesn't happen, because there's no way to exceed the .11 max bid of B)
To me, the problem is that people will still go in and then manually raise bids (or have third-party tools do this for them).
For instance, C is willing to pay up to .10 and can see that the people above them are below this amount. If they aren't moved automatically, then they may make the change manually. And if C moves up to .10, then B will move up to .11, since that's the max limit they've agreed to pay and one that even in Steve's model will gain them a position. And if B moves up, then A will automatically be able to move past them, as well.
The cost gaps under Overture's new system don't sit better with me more than anyone else. Unfortunately, I still don't see a way to get rid of them and allow auto-bidding. I've looked at a couple other suggestions from others besides Steve, and these also don't seem to work.
I suspect the ultimate solution will be when everyone (or many more people) moves to auto-bidding. I think that will naturally cause the cost gaps to disappear, though it is also likely that it will raise the overall price of bids.
Overture's Auto Bidding Feature Warrants Serious Consideration
Pay-Per Master, July 3, 2002
Harrison's proposal is explained here.
Up Close With Overture's Bid Management
The Search Engine Update, July 1, 2002
My last article on new auto-bidding at Overture.