Feb 22012

Follow-up: How Much Do "Top" Advertisers Matter to Google?

In a recent post, George Michie broke down the impact of an individual advertiser to Google's bottom line.  He concluded that in tightly packed, competitive auctions, an individual advertiser leaving the auction doesn't impact Google's revenue much at all. In more loosely packed auctions, even the top advertisers leaving would only cost Google an estimated 5-10% of the revenue they receive from those auctions.  In other words, an individual advertiser doesn't necessarily mean as much to Google as the advertiser might be inclined to think! After reflecting on this topic, I thought I'd continue the discussion with a few follow-up considerations.

There are two other hypothetical scenarios/elements that I considered, both focused on the "Top" of the ad auction. Top ads are those that Google "promotes" to appear above the organic listings.

Since Google provides data on Top vs. Other ad position performance, we're able to see the impact of ad position to CTR and ad spend in Google Auctions. For many clients, it is generally the case that over 85% of clicks and spend are incurred by ads in Top ad positions.  One might gather then that the answer to the question, "How much do you matter to Google?" largely resides in how many Top auctions each advertiser qualifies for. The severity of the impact a Top individual advertiser's exit from the auction is judged by their Level of Bid, Level of Quality or Both.

Mark Ballard previously discussed the elements of Top auctions in two contexts. First, he called out the Top of Page Bid Minimum Threshold. Later, he identified where automated rules to push to such a minimum would be a risky endeavor. In this last piece, he gave some hypothetical auction scenarios. I'll springboard off of the auctions used in that discussion for the purposes of this post. There, Mark constructed a model of QS and CTR correlation which is helpful in these hypothetical auctions. If you'd like details about those assumptions, read them in that post, and see the application in my tables below.

Scenario 1: High Max CPC Bids at the Top:

Where there are one or more advertisers with significantly above average Max CPCs, the impact of those advertisers to Google's revenue is significant.  First, the initial auction:

By eliminating Advertiser C, we see a sizable hit to Google's revenue (a 15% decrease). The reason for this is there are no new advertisers eligible to appear in the Top ad spots. Since Advertiser D does not have a high enough Ad Rank to be eligible to be promoted to the Top ad positions, Advertiser D will remain the first listed, right-rail (side) ad:

Note, we've even made the assumption that Advertiser D moving from position 4 to position 3 will experience a higher CTR, and therefore ultimately pay more to Google, without changing physical positioning on the page. Overall, however, Google still loses significant revenue due to the exit of a key advertiser.

Scenario 2: High Quality Advertisers at the Top:

Where there are one or more advertisers with significantly above average QS, the impact of those advertisers to Google's revenue is also significant, by similar reasoning.

Google needs high quality advertisers to capitalize on Top ad slots. Take an example where 3 advertisers have a 10 QS, while the remaining set have 5's and bid levels are tightly packed.

Here, we see a similar divide in Ad Ranks, and with the removal of advertiser C, no new advertiser is able to claim the now-vacated third Top ad spot. This costs Google greatly, to the tune of 14% in this example.

What this proves is that absent high-quality advertisers (those with high QS) Google loses it's ability to "promote" ads to Top ad slots on the page. This reduces overall Ad CTR and ultimately costs Google revenue. Furthermore, if the 'Top of Page Bid Minimum Threshold' (in this example set at 50) were higher, the impact would be more severe in all cases.

Remember, that only Google controls this dial, and can set it to whatever they wish. Google can lower the Top of Page Bid Minimum Threshold in this auction, and regain some of the loss, but definitely not all of it. Would they do this though? Their Top of Page Bid Minimum Threshold is meant to allow only the most qualified advertisers to be promoted, so shifting it around willy-nilly seems contrary to that assertion.

For the sake of argument, I wanted to show the impact of a different advertiser leaving. If Advertiser B exits, as opposed to Advertiser C, the impact to Google is greater (16% decrease), but relatively similar to the previous example.

Conclusion - A special focus must be given to Top (promoted) ad positions when measuring the impact of a single advertiser on Google's revenue. Highly aggressive bidders running away from their competitor's bids, and high quality advertisers amongst a set of average or lower ones are very important to Google. More important still are high-quality advertisers with aggressive bid strategies!

RKG would also like to recognize Martin Roettgerding for his comments and development of the discussion on his blog - PPC Ephiphany. Please consider reading his analysis, and Martin, thank you for adding to the discussion!


5 Responses to "Follow-up: How Much Do "Top" Advertisers Matter to Google?"
Matthew, Thanks a lot for the kind words, I feel honored. It's great to see that you still do research in this area. The thresholds for both top of page bid and first page bid are something of a mystery. In your simulations you used a static threshold, but I think in reality it must be somehow connected to the advertisers in the auction. This makes me think... I might have an idea on how these thresholds work. I have to think this through... guess I'm hooked again ;)
Hey Martin! You're completely right about the Top of Page Threshold not being Static. Google addresses that here: http://support.google.com/adwords/bin/answer.py?hl=en-AU&answer=73810 "This minimum price varies based on the quality of each ad per search query. For this reason, our system doesn't display the minimum price." While it's still a mystery as you point out, this help page only makes my argument for high-quality players stronger. Excited to see any additional angles from here... ball is now officially back in your court :)
Nigel says:
Surely though if Advertiser C leaves then the Advertisers A and B will also benefit with a slightly higher CTR and since they're paying more per click, then Google's revenue isn't decreased that drastically? In Scenario 1, the total number of clicks dropped from 490 to 476 and in Scenario 2 the total number of clicks drops from 539 to 502 but this is not taken into consideration. Where did the missing clicks go? The number of searches would not change if Advertiser C leaves, so all the clicks that would've gone to Advertiser C must be re-assigned and you cannot simply assume the CTR remains unchanged for the top 2. You noted in Scenario 1 that the CTR changes for Advertiser D, and in fact altered the CTR for all advertisers in the right block, but why not also for Advertisers A & B? Would this not alter your conclusion?
Martin says:
Hey, I don't think that the total number of clicks would have to remain unchanged because organic listings could get a big part. But I agree with your point about the changed CTR for A and B. That was my idea too and I think this could be the key to how these thresholds work. I actually have a half written blog post about this... Think about it the other way around: Not from the perspective of an ad dropping out but a new ad being added. Say there are two ads on top positions already and now a third ad is being added. Basing on Matthew's model and the point Nigel made, the existing ads would lose some clicks. This would mean that Google's revenue from these ads would decrease. So in order for Google to come out ahead, the new ad's revenue would have to be higher than the revenue lost. This would translate to a threshold the new ad has to exceed. And that threshold would be a dynamic one, completely based on the other ads... What do you think?
Really good questions/points Nigel and Martin! Martin, can't wait to read your new post, let us know when it goes live. You're right that I didn't take a potential CTR gain by ads A and B into account. If either CTR increases by lower competition, then it does change the numbers. However, I still believe there is a strong drop-off in the ad spend for advertiser B that will not be fully offset by a CTR increase. Martin, with the introduction of a new ad, I think Google will usually still 'win'. First, they are ensuring maximum competition above the promoted minimum threshold. Additionally, they are still in control of the minimum threshold amount, so they can technically keep the minimum in their favor to keep new advertisers out of the top auctions (should they choose to do so).

Leave A Comment