<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	>
<channel>
	<title>Comments on: AdWords Search Query Reports</title>
	<atom:link href="http://blog.merjis.com/2008/01/14/adwords-search-query-reports/feed/" rel="self" type="application/rss+xml" />
	<link>http://blog.merjis.com/2008/01/14/adwords-search-query-reports/</link>
	<description>Effective Internet Marketing Strategy and Technique Through Experiments, Measurement and Audit</description>
	<pubDate>Thu, 11 Mar 2010 23:01:28 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.5.1</generator>
		<item>
		<title>By: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-34690</link>
		<dc:creator>Jeremy Chatfield</dc:creator>
		<pubDate>Mon, 28 Jan 2008 13:14:50 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-34690</guid>
		<description>Hi Gerda,

Drat. I'm obviously making a real hash of getting the point over. Let me see if I can boil this down to a few sentences. Oh, and thanks for pointing to an error - I should have said "at least one impression" - put it down to haste (I was trying to get a rusted out exhaust pipe replaced at the same time, thanks to the wonders of the Vodafone USB Modem and the Kwik-Fit customer waiting area) and note that I've edited the comment to correct this. 

Super summary:

Google takes adverts from advertisers. Google publishes the adverts, but will often keep back relevant information from advertisers, for reasons that satisfy Google, but that do nothing positive for, and in some cases will hurt advertisers.

The mechanics of reporting are such that for some reports, you need at least an impression. For other reports, you need at least a click. The Search Query reports are in the latter category - you need a click, not an impression. I wasn't trying to go report by report and demonstrate which need a click and which need an impression... though that's probably something that should be explicitly documented. 

This requirement for either impressions or clicks are design decisions by Google. They could have made different choices, which would give advertisers less cause for concern about click fraud, and improve the ability to target adverts where they are most effective *for the advertiser*.  

Thanks for your challenges. Keep it up :)

Cheers, JeremyC.</description>
		<content:encoded><![CDATA[<p>Hi Gerda,</p>
<p>Drat. I&#8217;m obviously making a real hash of getting the point over. Let me see if I can boil this down to a few sentences. Oh, and thanks for pointing to an error - I should have said &#8220;at least one impression&#8221; - put it down to haste (I was trying to get a rusted out exhaust pipe replaced at the same time, thanks to the wonders of the Vodafone USB Modem and the Kwik-Fit customer waiting area) and note that I&#8217;ve edited the comment to correct this. </p>
<p>Super summary:</p>
<p>Google takes adverts from advertisers. Google publishes the adverts, but will often keep back relevant information from advertisers, for reasons that satisfy Google, but that do nothing positive for, and in some cases will hurt advertisers.</p>
<p>The mechanics of reporting are such that for some reports, you need at least an impression. For other reports, you need at least a click. The Search Query reports are in the latter category - you need a click, not an impression. I wasn&#8217;t trying to go report by report and demonstrate which need a click and which need an impression&#8230; though that&#8217;s probably something that should be explicitly documented. </p>
<p>This requirement for either impressions or clicks are design decisions by Google. They could have made different choices, which would give advertisers less cause for concern about click fraud, and improve the ability to target adverts where they are most effective *for the advertiser*.  </p>
<p>Thanks for your challenges. Keep it up :)</p>
<p>Cheers, JeremyC.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Gerda Arts</title>
		<link>http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-34682</link>
		<dc:creator>Gerda Arts</dc:creator>
		<pubDate>Mon, 28 Jan 2008 12:29:12 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-34682</guid>
		<description>Hi Jeremy,

Thank you for your quick reply.

I did not mis the point you were trying to make, I was merely pointing out there is another possible source of these impressions, which you had failed to mention. However, if you are saying it was the only point you were trying to make, and you were not trying to prove that the impressions are from "non-keyword-search" sites, then I did fail to see that, and I'm looking forward to a modified version of your article, clarifying the conclusions from your research.

I would like to ask you a question following from your comment. What do you mean by saying “Google’s reports *all* require a click to become effective.”? Maybe this time I'm missing what you are trying to say by putting all between aterisks? Not all Google reports require the items (e.g. keywords or ads) to generate a click before appearing in the report. In fact, for most reports the only requirement for items to be reported is that they have at least one impression. Is that what you were trying to say? 

Looking forward to your reply,

Gerda</description>
		<content:encoded><![CDATA[<p>Hi Jeremy,</p>
<p>Thank you for your quick reply.</p>
<p>I did not mis the point you were trying to make, I was merely pointing out there is another possible source of these impressions, which you had failed to mention. However, if you are saying it was the only point you were trying to make, and you were not trying to prove that the impressions are from &#8220;non-keyword-search&#8221; sites, then I did fail to see that, and I&#8217;m looking forward to a modified version of your article, clarifying the conclusions from your research.</p>
<p>I would like to ask you a question following from your comment. What do you mean by saying “Google’s reports *all* require a click to become effective.”? Maybe this time I&#8217;m missing what you are trying to say by putting all between aterisks? Not all Google reports require the items (e.g. keywords or ads) to generate a click before appearing in the report. In fact, for most reports the only requirement for items to be reported is that they have at least one impression. Is that what you were trying to say? </p>
<p>Looking forward to your reply,</p>
<p>Gerda</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-34468</link>
		<dc:creator>Jeremy Chatfield</dc:creator>
		<pubDate>Sat, 26 Jan 2008 15:12:32 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-34468</guid>
		<description>Hi Gerda

Thanks for the comment. You missed the point - which is my fault. 

The point is that Google is not telling advertisers where the impressions are going. That's really useful information. If the adverts are appearing on unsuitable sites or searches, that can account for low click rates - which in turn would allow adjusting keywords, bids, budgets and excluded sites to improve performance. The fact that Google doesn't present this information to advertisers is the entire point, rather than an omission.

Google's reports *all* require at least one impression to become effective. It's been one of the long lasting complaints about the AdWords API, since at least 2006. So called "zero-impression" reporting would allow AdWords API users to extract whole account data via reports - but Google's reporting system *only* collects data for reporting when at least one impression is involved. IMO, this is a serious design failure. 

The impact on API users is that in order to collect whole account information, especially when there may be user-managed (as well as automation management) changes, is a drastic increase in costs to collect the information (1000 quota used per report - would be enough to query a single account for a handful of AdGroups and keywords).

I should have explicitly pointed out the Google design decision in the article, in retrospect. I may well amend the article to explicitly bring this out... but there's some other research in progress that I'll probably publish first. 

Cheers, JeremyC.

[edited to reflect that Google's reports currently require at least one impression before they report anything, and some require at least one click - previously an erroneous statement that all reports required at least a click.]</description>
		<content:encoded><![CDATA[<p>Hi Gerda</p>
<p>Thanks for the comment. You missed the point - which is my fault. </p>
<p>The point is that Google is not telling advertisers where the impressions are going. That&#8217;s really useful information. If the adverts are appearing on unsuitable sites or searches, that can account for low click rates - which in turn would allow adjusting keywords, bids, budgets and excluded sites to improve performance. The fact that Google doesn&#8217;t present this information to advertisers is the entire point, rather than an omission.</p>
<p>Google&#8217;s reports *all* require at least one impression to become effective. It&#8217;s been one of the long lasting complaints about the AdWords API, since at least 2006. So called &#8220;zero-impression&#8221; reporting would allow AdWords API users to extract whole account data via reports - but Google&#8217;s reporting system *only* collects data for reporting when at least one impression is involved. IMO, this is a serious design failure. </p>
<p>The impact on API users is that in order to collect whole account information, especially when there may be user-managed (as well as automation management) changes, is a drastic increase in costs to collect the information (1000 quota used per report - would be enough to query a single account for a handful of AdGroups and keywords).</p>
<p>I should have explicitly pointed out the Google design decision in the article, in retrospect. I may well amend the article to explicitly bring this out&#8230; but there&#8217;s some other research in progress that I&#8217;ll probably publish first. </p>
<p>Cheers, JeremyC.</p>
<p>[edited to reflect that Google's reports currently require at least one impression before they report anything, and some require at least one click - previously an erroneous statement that all reports required at least a click.]</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Gerda Arts</title>
		<link>http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-34379</link>
		<dc:creator>Gerda Arts</dc:creator>
		<pubDate>Fri, 25 Jan 2008 21:24:33 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-34379</guid>
		<description>I'm very disappointed with this article, and the fact that the author did not take more time to find out what the Search Query Report actually provides. When I read the article, I was surprised by such a high CTR as reported by the search query report used in the article. As no other information was available (it could have been because of a high position of the ad and a lack of competitors ads showing for the search queries), I obtained a report for one of our own clients. Surprisingly, I also found a CTR reported by the Search Query Report of over 80%. I found this hard to believe because of the competitive market of the keywords, and the fact that the average position was only 3. I decided to have a look at what data the report actually provides, as I thought it was a bit suspicious that all (groups of) search queries resulted in at least 1 click. Looking at the Adwords Help Centre, it is clearly stated that: "A Search Query Performance report shows performance data for the search queries that triggered the ads which appeared after receiving clicks". In other words, it will not report any impressions of search queries that did not result in a click. So the "99 missing impressions" are probably just impressions of search queries that did not result in a click, and there is no evidence that these are "non-search impressions". I do have to agree with the author though, that the Search Query Performance report is of very limited use at the moment.</description>
		<content:encoded><![CDATA[<p>I&#8217;m very disappointed with this article, and the fact that the author did not take more time to find out what the Search Query Report actually provides. When I read the article, I was surprised by such a high CTR as reported by the search query report used in the article. As no other information was available (it could have been because of a high position of the ad and a lack of competitors ads showing for the search queries), I obtained a report for one of our own clients. Surprisingly, I also found a CTR reported by the Search Query Report of over 80%. I found this hard to believe because of the competitive market of the keywords, and the fact that the average position was only 3. I decided to have a look at what data the report actually provides, as I thought it was a bit suspicious that all (groups of) search queries resulted in at least 1 click. Looking at the Adwords Help Centre, it is clearly stated that: &#8220;A Search Query Performance report shows performance data for the search queries that triggered the ads which appeared after receiving clicks&#8221;. In other words, it will not report any impressions of search queries that did not result in a click. So the &#8220;99 missing impressions&#8221; are probably just impressions of search queries that did not result in a click, and there is no evidence that these are &#8220;non-search impressions&#8221;. I do have to agree with the author though, that the Search Query Performance report is of very limited use at the moment.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Richard Ball</title>
		<link>http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-32774</link>
		<dc:creator>Richard Ball</dc:creator>
		<pubDate>Mon, 14 Jan 2008 13:52:28 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/2008/01/14/adwords-search-query-reports/#comment-32774</guid>
		<description>Thanks for the mention - and good to see you blogging again!</description>
		<content:encoded><![CDATA[<p>Thanks for the mention - and good to see you blogging again!</p>
]]></content:encoded>
	</item>
</channel>
</rss>
