What determines 1) Ad position 2) Actual CPC?

We are debating this at the moment...

[Mark_A]

1) What determines Ad position?

Is it just Max CPC or are there other factors at work.

If I bid £1.00 and my next competitor bids £0.50 I assume I will be ranked above him, but how much more will I pay per click?

2) What determines actual (and average) CPC?

If you bid max £1.00 per click how come on average you can be charged £0.66 per click?

[exmoorbeast]

Hi

Try searching your query on Google, there are loads of videos by Google on You Tube - some really amazing ones about the ad aution.

I got this from a search:

Determining Click Cost:


Actual CPC = (AdRank to beat/Quality Score of Advertiser) + $.01


Each advertiser only pays the minimum amount required to maintain his position.

An advertiser's actual CPC is determined by dividing the AdRank to beat (the AdRank of the competitor below them) by their own Quality Score plus one penny. The lowest positioned ad will only pay the minimum price required by Google to show on the page.

Hope that helps. If you are at this point now, try reading blogs like Click Equations ROI Revolution etc.

[Mark_A]

Hi exmoorbeast, thanks for your response. (I will check out the blogs.)

It has introduced a couple of new concepts for me, Adrank and Quality Score!

So, what determines Adrank and Quality Score?

And what is actually happenning when we are regularly encouraged by Google to increase our maximum cost per click?

[faust1]

i think when you start an adwords campaign you have to pay google first in order to let google pay you back.. so initially you may have to set high bids to rank high .. .

[LucidSW]

Do a search on "lucid adwords faq" to find my own document which has the answers to your questions.

[Mark_A]

Do a search on "lucid adwords faq" to find my own document which has the answers to your questions.

Thanks LucidSW - it is a useful document.

[Eurydice]

Lucid, on p. 11-14, you write about Quality Score (QS). On p. 11, you write that QS has many factors. But later, you tend to write that QS is the value that one sees for the keyword QS in Keyword Editor. Your formulas tend to use 1-10 as values for QS.

However, I've seen formulas at Google where QS is much higher. They're not using KW QS. They're using a "Total QS", as in what you describe at the beginning (QS based on many factors). That QS value can be much higher than 10: I've seen formulas at Google where the QS was 100.

I asked a Google engineer if this was the upper limit and she said there is no upper limit. I asked "so it could be, say, 400?" and she said yes.

If you recalculate your formulas and use higher values, say QS 40 or QS 60, you'll see more interesting results. An advertiser can bid $0.50 and beat another advertiser who bids $10.00. (By the way, it's very confusing in your examples that you don't show cents, e.g., p. 19, where 30 and 23 should be shown as $0.30 and $0.23).

Google doesn't show the Total QS, only KW QS. How to tell if you have high (or low) Total QS: look at the difference between bid and CPC. If the difference is low (i.e., you bid $2 and you pay $1.93), then your QS is poor. I've seen cases where the bid was $2.00 and the CPC was $0.05. That's very good QS.

Lucid, thanks for writing about this and making it public. I wrote a complete description of the QS formula in my book that was published a bit more than a year ago. Most books on the market don't understand the QS formula or its implication.

In summary: Bids are a minor (very minor!) factor in your advertising. QS is the main factor. If you rely on ABM tools such as Kenshoo or Marin, it's extremely dangerous; your budget will be looted by Google. Spend 80% of your time on improving the QS, i.e., better ads, better AGs, better landing pages, better conversion process, etc.

(Download Lucid's doc before the moderators find it and remove the pointer :-)

[LucidSW]

Eurydice, I think you misunderstand.

The QS that is shown to you is a number between 1 and 10. Doesn't matter if Google uses other numbers and does a conversion to that scale, you would never see that so it doesn't matter.

The point of the document too is to make it easy to understand, not start using numbers used in the background that don't relate to what you see in your account. But thanks for the feedback to all.

I don't understand what you are saying about Total QS. There is no such thing. QS is calculated at the keyword-ad level. I think you mean the QS if you have more than one ad.

Regardless, your example goes against what I say on how your CPC is calculated. If you bid $2 and pay $1.93, it's because the advertiser below you has an ad rank (QS times bid) similar to yours. If their ad rank was much lower, you'd pay less.

Agree on improving the QS. I've been preaching that for years. But look at the formula. Bids are not that minor.

[Mark_A]

Incidentally, I just split my ad campaigns so that I have one for search and another for content.

On the search I have QS etc as expected, but on the content ads, despite my use of lots of keywords, QS and the like seem not to show.

Instead on the content only ads it displays "We're only able to provide diagnosis information for keywords in campaigns that are opted in to the search network."

[netmeg]

That's right. They don't show QS for Content Network. The best way you can determine it's working is if your ads appear to be showing and you're racking up impressions (and hopefully clicks)

Now, two reports you are going to want to get real familiar with -

For the Search campaigns, the Search Query report. Run that on any campaign that has broad match keywords. Because that will tell you what keywords your ads are actually running on (vs what you listed in your account) If you have any words that are ambiguous in meaning, or need to be narrowed down to make it clearer (like words with multiple meanings, or aspects - "bags" or "nuts" or whatever) this would show you. You go through that list, and look at all the ways Google is matching ads to your keywords, and if anything looks way off, or are racking you up a ton of spend without making you money, you want to add them as negatives to those campaigns.

For the Content Network, you want to run the Placement Report. That will tell you which Content Network sites your ads are running on. Again - you want to eyeball it and look for stinkers. Anything that's racking up a ton of clicks but doesn't look right (like a social networking site, for example) If you have any doubts, go take a look at the site and see if it seems like a likely place to be advertising your products. Any that are way off - add them as negatives sites to the Content campaigns.

This is how you start to zero in on keywords and sites that will perform for you.

[Mark_A]

Hi netmeg, thanks again for the sage advice. Am looking at reports now. I can't at first glance see a search query report.

I just this week switched all keywords for search into phrase match. CTRs are way up but total clicks are down and the ads have been limited by budget which seems to suggest our average cost per click may have gone up quite a bit. Average quality score is 7/10.

[netmeg]

That's not bad. If you're just getting your feet wet, start with phrase match, and then add the others as you build history (and confidence in what you're doing)

Search Query is actually called Search Query Performance. And there are two reports that use the word Placement - for what I was talking about above, you want the Placement Performance report.

[faust1]

i have heard from a google representative at my side that google adwords system even checks for minor errors such as grammar errors, spelling errors, image html things, html site codes, etc....to determine the landing page quality score ... however, for keyword relevancy one should include the keyword at least 3 - 4 times properly ...

[Eurydice]

Add to Netmeg's notes: GAW doesn't care about CTR in Content Network.

[Mark_A]

Search Query is actually called Search Query Performance.

That is a great report, just what I need right now!

[Mark_A]

That is a great report, just what I need right now!

Famous quote: "Half your advertsing is wasted, the trouble is knowing which half."

Well now I can tell down to individual key terms which users clicked on my ad for, in some cases I have no idea why they clicked on my ad, perhaps they were just trying to waste my money!

On my most important Ad using broad match 29% of keyterms clicked were wasted (irrelevant). On phrase match that waste fell to 14% which is more acceptable. Obviously I know how much money that represents!

On another important ad, 74% of key terms clicked under broad match were wasted / irrelevant, because of the peculiarities of that sector, under phrase match that wastage fell way down to 18%.

[netmeg]

That doesn't mean that broad match won't work for you. You can bid cheaply on broad match, and use that report to find the good ones that you're willing to pay more for, and put those into phrase match. And the ones that are completely off the mark, you can add them to the negative keywords, and keep using broad match.

If you find that you can't figure out why in the world your ads are getting clicked on for some terms, consider trying to use more precise text in your ads to further qualify the clickers. Make sure they have a pretty good idea before they click what they're going to be landing on. That helps narrow down misguided or idle curiosity clicks.

[Mark_A]

Yes, I noticed that some quite specialist three word terms seem to require more money Max CPC to get onto the first page. I am sure they were cheaper in broad match but they probably produced a lot of wasteful clicks and thus wasted money.

One ad was a problem because its key terms were too similar to another much more popular but different product. It was that one that had 74% wasteful clicks and wasted money.

[Mark_A]

I may have to examine the placement performance report because since the split my adsense bound adverts have been racking up a reasonable number of exposures, pretty much all in position 1 but there have been absolutely no clicks. Not a one. Something is afoot!

[Acrill]

When running Content Network ads you should be choosing different keywords than you would use for Google Search ads.

Think of it like this: With Search ads the keywords you select should match search queries that users enter into Google.

With Content Network ads, the keywords you select should reflect the theme of pages where you want your ads to show. They should generally be broader than the keywords you would use for Search ads.

[Mark_A]

Well I am not there yet with my content ad.

First under network it lised some 6 sites that the ads were appearing on, none of them were at all relevant so I excluded them.

But now it is still exposing the ad a few thousand times but not showing me on which websites. And still no clicks (not true, there was one!) but it is pretty ridiculous. We used to have quite a lot of clicks from adwords content before but for some reason I have killed it off!.

[netmeg]

A few thousand times isn't really enough to go on.

What I would do is run some reports from back when you WERE getting clicks from Content and see what sites were delivering them. You can still run reports on deleted or much-modified campaigns. Just pick a time from from before you changed it.

[smallcompany]

Actual CPC = (AdRank to beat/Quality Score of Advertiser) + $.01

And yet you pay $0.75 or more for some keywords even if your ad is the only one in space.
Of course, Google has the explanation for that as well.

My explanation is that each keyword at the given moment has its own price that has been predicted based on the keyword's history (within Google, not your account) and therefore Google's system knows how much it is going to charge for it before the click happens.

Hey, they do budgeting like other companies. . .

[netmeg]

Yea well, Google's not much different from a lot of publishers there. If you mosey on over to the AdSense forum, you will find a lot of publishers who would like to set a threshold for the ads - i.e. if they pay less than X amount, don't show them.

That'll never happen because it would leave AdWords with a lot of low end inventory they wouldn't be able to place.

But it stands to reason that Google probably figures advertising on their property has a minimum value below which they won't display, based on at least in part on quality score.

Reminds me of a joke for which the punch line is "Oh we've already determined WHAT you are, now we're just negotiating price"