<?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: Privacy and AdWords Extended Search Query Reports</title>
	<atom:link href="http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/feed/" rel="self" type="application/rss+xml" />
	<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/</link>
	<description>Effective Internet Marketing Strategy and Technique Through Experiments, Measurement and Audit</description>
	<pubDate>Fri, 12 Mar 2010 17:37:27 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.5.1</generator>
		<item>
		<title>By: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-142757</link>
		<dc:creator>Jeremy Chatfield</dc:creator>
		<pubDate>Mon, 31 Aug 2009 09:31:02 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-142757</guid>
		<description>Dr Acne  - great name, BTW, how original! [sarcasm] I *must* change my name to "Internet Marketing Strategy" and post on a lot of nofollowed blog comment areas. [/sarcasm] Yes, and I see you, too, Ms Coach Bags. Sheesh. Install some FireFox plugins or set up GreaseMonkey/GreaseKit to highlight blogs that aren't worth trying to seed with links... 

Anyway... Dr Acne misses the key point. Keyword Quality Score is apparently determined by CTR on the exact match, on Google's search results only. It isn't affected by CTR on partner sites or by phrase and broad matches... Google is already offering a substantial fraction of the data needed to manage the CTR driven portion of the QS. It's the *extra* range of augmented enhanced super-wide investor-pleasing revenue-projection-matching Broad Match that is being obfuscated.</description>
		<content:encoded><![CDATA[<p>Dr Acne  - great name, BTW, how original! [sarcasm] I *must* change my name to &#8220;Internet Marketing Strategy&#8221; and post on a lot of nofollowed blog comment areas. [/sarcasm] Yes, and I see you, too, Ms Coach Bags. Sheesh. Install some FireFox plugins or set up GreaseMonkey/GreaseKit to highlight blogs that aren&#8217;t worth trying to seed with links&#8230; </p>
<p>Anyway&#8230; Dr Acne misses the key point. Keyword Quality Score is apparently determined by CTR on the exact match, on Google&#8217;s search results only. It isn&#8217;t affected by CTR on partner sites or by phrase and broad matches&#8230; Google is already offering a substantial fraction of the data needed to manage the CTR driven portion of the QS. It&#8217;s the *extra* range of augmented enhanced super-wide investor-pleasing revenue-projection-matching Broad Match that is being obfuscated.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Dr Acne</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-142679</link>
		<dc:creator>Dr Acne</dc:creator>
		<pubDate>Wed, 26 Aug 2009 20:53:41 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-142679</guid>
		<description>It is crucial for determining how to extend Broad Match. Low CTRs on poorly chosen search queries will result in the search query no longer being matche</description>
		<content:encoded><![CDATA[<p>It is crucial for determining how to extend Broad Match. Low CTRs on poorly chosen search queries will result in the search query no longer being matche</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Coach Bags</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-142616</link>
		<dc:creator>Coach Bags</dc:creator>
		<pubDate>Tue, 25 Aug 2009 16:21:34 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-142616</guid>
		<description>I completely agree with the person above that "I doubt it’s about privacy. My guess is that G prefers to hide this info but figured that the policy in his case is not effective due to the ability to get this info from your server logs or analytics package." 
So Google is now letting us know what we could already know, and it is helpful for us in managing our web server logfiles properly.


&lt;a href="http://www.clearlychic.com" rel="nofollow"&gt;coach bags&lt;/a&gt;</description>
		<content:encoded><![CDATA[<p>I completely agree with the person above that &#8220;I doubt it’s about privacy. My guess is that G prefers to hide this info but figured that the policy in his case is not effective due to the ability to get this info from your server logs or analytics package.&#8221;<br />
So Google is now letting us know what we could already know, and it is helpful for us in managing our web server logfiles properly.</p>
<p><a href="http://www.clearlychic.com" rel="nofollow">coach bags</a></p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Richard Ball</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-140172</link>
		<dc:creator>Richard Ball</dc:creator>
		<pubDate>Thu, 28 May 2009 19:49:45 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-140172</guid>
		<description>My goodness!  For the sake of privacy, Yahoo had better drop OVRAW from their tracking URLs.  ;-)</description>
		<content:encoded><![CDATA[<p>My goodness!  For the sake of privacy, Yahoo had better drop OVRAW from their tracking URLs.  ;-)</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-140171</link>
		<dc:creator>Jeremy Chatfield</dc:creator>
		<pubDate>Thu, 28 May 2009 19:31:18 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-140171</guid>
		<description>I can understand the privacy thing. WIth a geek background, you worry about information flow; use of globals, leakage, security. Most people haven't got a clue, all they know is that someone else knows something, and that is a threat for many people. The unknown is always fearful. Look at the fuss over cookies, for example. 

Marketing dictum: perception is reality. So Google has to deal with the reality that they are perceived as interfering with privacy. But with an already fully anonymised, aggregated search query report - WTF is the issue?</description>
		<content:encoded><![CDATA[<p>I can understand the privacy thing. WIth a geek background, you worry about information flow; use of globals, leakage, security. Most people haven&#8217;t got a clue, all they know is that someone else knows something, and that is a threat for many people. The unknown is always fearful. Look at the fuss over cookies, for example. </p>
<p>Marketing dictum: perception is reality. So Google has to deal with the reality that they are perceived as interfering with privacy. But with an already fully anonymised, aggregated search query report - WTF is the issue?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Robert Brady</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-140165</link>
		<dc:creator>Robert Brady</dc:creator>
		<pubDate>Thu, 28 May 2009 16:48:20 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-140165</guid>
		<description>I think the whole privacy thing is blown out of proportion. If you think Google is violating your privacy, STOP USING GOOGLE. This isn't like a doctor or lawyer where laws exist to protect privacy. You chose to use Google, so you also choose to accept the possible consequences. 

As for Google using broad match to increase their revenues, I totally agree. I manage several accounts and I noticed a significant change in impressions and CTR a few months ago that would indicate a change in the algorithm that AdWords uses. Conspicuous how it matches with their recent earnings slump isn't it?</description>
		<content:encoded><![CDATA[<p>I think the whole privacy thing is blown out of proportion. If you think Google is violating your privacy, STOP USING GOOGLE. This isn&#8217;t like a doctor or lawyer where laws exist to protect privacy. You chose to use Google, so you also choose to accept the possible consequences. </p>
<p>As for Google using broad match to increase their revenues, I totally agree. I manage several accounts and I noticed a significant change in impressions and CTR a few months ago that would indicate a change in the algorithm that AdWords uses. Conspicuous how it matches with their recent earnings slump isn&#8217;t it?</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-140152</link>
		<dc:creator>Jeremy Chatfield</dc:creator>
		<pubDate>Thu, 28 May 2009 10:50:08 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-140152</guid>
		<description>Yeah, I'll definitely agree that there's at least five "layers" - matching service, impression server, click service, reporting/billing and invalid click service. Could add conversion tracking to that, I guess.

However, I'm fairly sure that the matching service runs a regular training update based on the search query matching and clicking. That layer learns - you can see "twitches" typical of an AI with a new training set, every so often, and it adjusts to "low volume keywords", suddenly presenting them as distinct impressions when the CTR is high enough.

And yes, age of the system seems relevant. The AdWords API took forever to engage with zero-impression reporting; hugely useful and more cost effective. Probably because it needed to develop hooks deeper into the system. 

I shall mutter to myself about the likely system design again.  Thanks for the provocation :)

It still seems that "User Privacy" is a distinctly red herring, though.</description>
		<content:encoded><![CDATA[<p>Yeah, I&#8217;ll definitely agree that there&#8217;s at least five &#8220;layers&#8221; - matching service, impression server, click service, reporting/billing and invalid click service. Could add conversion tracking to that, I guess.</p>
<p>However, I&#8217;m fairly sure that the matching service runs a regular training update based on the search query matching and clicking. That layer learns - you can see &#8220;twitches&#8221; typical of an AI with a new training set, every so often, and it adjusts to &#8220;low volume keywords&#8221;, suddenly presenting them as distinct impressions when the CTR is high enough.</p>
<p>And yes, age of the system seems relevant. The AdWords API took forever to engage with zero-impression reporting; hugely useful and more cost effective. Probably because it needed to develop hooks deeper into the system. </p>
<p>I shall mutter to myself about the likely system design again.  Thanks for the provocation :)</p>
<p>It still seems that &#8220;User Privacy&#8221; is a distinctly red herring, though.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Tomas</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-140108</link>
		<dc:creator>Tomas</dc:creator>
		<pubDate>Wed, 27 May 2009 11:22:46 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-140108</guid>
		<description>Hi Jeremy,

I agree with you that Google is most definitely collecting that data during matching. But from what i've seen and experienced ove the years (and i'm sure you have too) adwords is a fairly disconnected system that collates data from various sources (e.g. clicks and impressions are not recorded/stored at the same moment cause there's sometimes a time delay between them - i.e. clicks get updated but impressions aren't).

the search query report is relatively late addition to adwords and i think it's just another layer/module they put on top of the core. That's why i think it may be just be mining the referer urls instead of talking to the matching algorithm (which would prob be a lot more complicated/risky to implement)</description>
		<content:encoded><![CDATA[<p>Hi Jeremy,</p>
<p>I agree with you that Google is most definitely collecting that data during matching. But from what i&#8217;ve seen and experienced ove the years (and i&#8217;m sure you have too) adwords is a fairly disconnected system that collates data from various sources (e.g. clicks and impressions are not recorded/stored at the same moment cause there&#8217;s sometimes a time delay between them - i.e. clicks get updated but impressions aren&#8217;t).</p>
<p>the search query report is relatively late addition to adwords and i think it&#8217;s just another layer/module they put on top of the core. That&#8217;s why i think it may be just be mining the referer urls instead of talking to the matching algorithm (which would prob be a lot more complicated/risky to implement)</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-140104</link>
		<dc:creator>Jeremy Chatfield</dc:creator>
		<pubDate>Wed, 27 May 2009 07:22:11 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-140104</guid>
		<description>Hi Tomas - I'm pretty sure that G does collect this data. It is crucial for determining how to extend Broad Match. Low CTRs on poorly chosen search queries will result in the search query no longer being matched, IMO. That means that G is recording the search query... doesn't it?

Your general point about unexpected simple optimisations is appreciated though. Seen it :)</description>
		<content:encoded><![CDATA[<p>Hi Tomas - I&#8217;m pretty sure that G does collect this data. It is crucial for determining how to extend Broad Match. Low CTRs on poorly chosen search queries will result in the search query no longer being matched, IMO. That means that G is recording the search query&#8230; doesn&#8217;t it?</p>
<p>Your general point about unexpected simple optimisations is appreciated though. Seen it :)</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Tomas</title>
		<link>http://blog.merjis.com/2009/05/23/privacy-and-adwords-extended-search-query-reports/#comment-140052</link>
		<dc:creator>Tomas</dc:creator>
		<pubDate>Tue, 26 May 2009 08:46:54 +0000</pubDate>
		<guid isPermaLink="false">http://blog.merjis.com/?p=293#comment-140052</guid>
		<description>other alternative: 
it could be that those reports are not generated from the matching engine, but are (like external tools) based on the actual referer field.

you'll notice that the first hop of every ad goes back to google, and it could be that they use that hop to extract the query from the referer url. 

It would be odd if they did it that way, but it wouldn't be the first time engineers come up with a shortcut in order to achieve something. Perhaps doing it that way is technically simpler or it is done on purpose in order not to expose the matching algorithm.</description>
		<content:encoded><![CDATA[<p>other alternative:<br />
it could be that those reports are not generated from the matching engine, but are (like external tools) based on the actual referer field.</p>
<p>you&#8217;ll notice that the first hop of every ad goes back to google, and it could be that they use that hop to extract the query from the referer url. </p>
<p>It would be odd if they did it that way, but it wouldn&#8217;t be the first time engineers come up with a shortcut in order to achieve something. Perhaps doing it that way is technically simpler or it is done on purpose in order not to expose the matching algorithm.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
