I am sorry, but I have to disagree with those against IBA.
The fact still remains that if the algo shows that such IBA has a higher chance of a higher return, it will be displayed over that of an interest based ad. Even if it takes time for a user to click such ad after seeing it on multiple sites, multiple times an algo can easily take that into consideration when determining which ad to place. For example:
Let’s say that the algo determines that said IBA has this % chance of being clicked based on times seeing on different sites:
1 time 1% chance
2 times 2% chance
3 – 5 times 3.5% chance
6 – 7 times 2.5% chance
Greater than 7 times .5% chance
They could use the above numbers to determine which ad to display based on expected returns between contextual and IBA ads.
3 - 5 times = ($5.00 IBA bid)(3.5%) = .175
Contextual = ($2.00 Contextual bid)(4%)= .08
IBA wins and is placed. The numbers would change as above each time, so there would be times when the contextual would win out. Say after greater than 7 times the IBA is shown.
7 times = ($5.00 IBA bid)(.5%)= .025
Contextual = ($2.00 Contextual bid)(4%)=.08
Contextual wins and is placed.
The analogy of a shop with an ice cream sign out front selling shoes inside is not the same as a website taking in visitors that may have just been looking for that pair of shoes that IBA is now displaying on your site.
If your physical store were digital, and could collect the massive amount of data that Google does, and it could place a product that your visitor was recently looking for in front of that store visitor (even if the item had nothing to do with your store), that visitor may indeed purchase it. You can't compare apples to oranges.
Don't get me wrong, I'm no Google-Fan-Boy, but...