All Google ad auctions act on the same principles outlined in Chapter 3, regarding Ad Rank, Max Bid, and Quality Score. However, the auction for display campaigns (and shopping campaigns) is more complicated, as individual advertisers have the opportunity to win multiple ad placements on the same webpage.
First, let’s cover the concept of incremental clicks. We first mentioned this when describing how 86 percent of dynamic search traffic is incremental, meaning 86 percent of traffic earned from dynamic search would not have been earned if you were using traditional keyword search campaigns.
Incremental clicks can also refer to clicks earned by having more favorable ad positions. A Google Search ad in Position 1 on page 1 of the search results will receive incrementally more clicks than the same ad on page 2 of the search results, for no other reason than its actual position on the SERP.
This concept also applies to display ad placements and is factored into the display ad auction. A banner ad at the top of a NYT article would receive more incremental clicks than the same banner ad if it was placed at the bottom of the page. The auction is designed to ensure that advertisers are paying a fair price for incremental clicks.
With that said, here’s how the display ad auction differs from the traditional search ad auction:
● You’ll pay what’s required to rank higher than the next best ad position only for incremental clicks you get from being in the current position.
● You’ll pay the price you would have for the next best ad position for the rest of the clicks.
● You may pay an additional service fee for ads that use audience targeting. In such cases, your maximum bid is reduced before the auction, and the fee is added to the closing auction price. 130
Regarding this last point: This is related to the generation of in-market, affinity, and custom intent audience. The description is relatively vague and the only follow-up information Google provides on this is:
The data used to generate audiences (for example, page visit history, past Google searches) may be used to improve the bidding and targeting of your audience campaigns.
I have not confirmed this directly, but I interpret this as Google saying that they partner with data providers for some of this information, which will take a cut of the ad revenues. This is how most traditional demand-side advertising platforms operate.
Regarding the auction, let’s say there is an ad placement with two positions. Your ad might get ten clicks by being in the topmost visible position, but only eight clicks if it showed in the less visible second position. Here, the two additional clicks are considered the incremental clicks for being in the top position, as opposed to the second position.
If you are the advertiser that wins the top position, you will typically pay about one penny more than what’s required to rank higher than the advertiser immediately below for the two incremental clicks. For the remaining eight clicks, you will pay a lower price (the amount you would have paid if it had ranked in the second position).
Google provides a helpful example in their support documentation:
In this example, there are three advertisers with the same Quality Score competing for an ad unit that can show only one ad. Alice wins the auction because she is bidding the highest, based on a combination of Max CPC bid and Quality Score.
The amount that’s required for Alice to rank above the next best ad—Bob’s—is $3.01. Because this ad unit only shows one ad, all of the clicks Alice receives are considered incremental to what she’d have received in a lower position. This is equivalent to not being shown because there is no lower position. Alice then pays an Actual CPC of $3.01 per click.
Now, say that the ad unit shows more than one ad. We use “Relative CTR of Position” to show that not all ad positions are equally visible, and that higher ad positions can get you more clicks than lower ad positions. (“CTR” refers to your ad’s clickthrough rate.)
In this particular example, the top position is much more visible than the second position: the ad will get three times more clicks in the top position than it would have gotten in the second position.
All of Bob’s clicks are considered incremental compared to what he’d have received in a lower position, not having his ad show. Therefore, Bob pays an Actual CPC of $1.01 per click.
Alice’s Actual CPC in this example depends on both the incremental clicks she gets from being in the top position and Bob’s Actual CPC. According to the Relative CTR of Position, Alice is getting three times as many clicks by being in the top position compared to being in the second position. So, two-thirds of her clicks are considered incremental, while the remaining one-third are clicks she’d have gotten if she had occupied the second position.
Let’s say Alice gets three clicks total, for the first click she’d pay the same price Bob pays. For her next two clicks, she’d pay $3.01, which is the price required to rank above Bob’s ad. Her costs would look like this:
Her actual CPC would be $2.34, which equals the total she spent divided by the clicks she got.
Lastly, ads entering the auction on the display network are always using CPM bidding, even if you choose Manual CPC as your bid strategy. When an auction takes place for a display network placement, all advertisers’ bids are converted to CPM bidding using a CPM multiplier, so that they can be compared apples to apples. The CPM multiplier takes expected CTR into account when your bids are converted over, and your Actual cost per click will never be greater than your Max. CPC Bid. This fact doesn’t change anything from a strategy perspective, but it’s useful to know.