That’s interesting to read about, Thank you for sharing this informative stuff..

Furthermore, just from a ten-thousand foot perspective with the second-price RTB model…say bidding just begins and we start at oh, for the sake of simplicity, a penny for a set of impressions that can be appraised for say, $.30 CPM. If I know this is a second-price model, can’t I make a ridiculous bid such as $10.00 PM and then lock in some sort of low price? I guess I’m going under the assumption that A) the bidding happens in discrete time B) bid reception is instantaneous and C) all information is public (EG if I bid 30 cents, then all other bidders immediately know I bid 30 cents).

Now I’m probably missing something here. But I’m not sure what it is.

As a result, there is asymmetric information between the bidders (the person who has the current highest bid has revealed more information than others)

For ad serving you have sealed bids (so nobody knows what the other bidders are bidding). Optimum bidding strategies are different as a consequence.

their arguments are that floors reduce profitability for the publisher, and shouldn’t be used.

if you look at example tables, the first one makes $13 for the publisher, the 2nd $7 (which uses a smarter bidding strategy), and the 3rd (with floors) $6.50.

the smarter bidding strategy sounds the best, but it doesn’t include the probability that the user won’t get to the 3rd/4th impression (ie the risk associated with the browser not seeing the ad at all), and he might have to end up raising it to get his inventory filled.

the 3rd one sounds bad, but it assumes that the publisher’s house ads can’t make more than 50c over 6 positions (have a lower CPM than 9c), which may not be true.

what it does point to (to me) is that the publisher should be putting his house ads into the RTB as well, which they are doing by putting a floor on there. The floor being the house-ad or best offer the can get not using the RTB.