<?xml version="1.0" encoding="UTF-8"?><!-- generator="wordpress/2.0.12-alpha" -->
<rss version="2.0" 
	xmlns:content="http://purl.org/rss/1.0/modules/content/">
<channel>
	<title>Comments on: AdWords - Relevance - WTF?</title>
	<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/</link>
	<description>Effective Internet Marketing Strategy and Technique Through Experiments, Measurement and Audit</description>
	<pubDate>Thu, 24 Jul 2008 18:20:37 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.0.12-alpha</generator>

	<item>
		<title>by: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-44211</link>
		<pubDate>Thu, 03 Apr 2008 18:12:40 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-44211</guid>
					<description>@Dan - thinking further... The key difference between this new and still immature market in keywords and impressions and the older media, is third party oversight. &lt;a href="http://www.abce.org.uk/" rel="nofollow"&gt;ABC&lt;/a&gt; and similar entities provide a third party assurance to advertiser and publisher. Google doesn't have that - &lt;a href="http://www.comScore.com/" rel="nofollow"&gt;comScore&lt;/a&gt; and the like sample the data, and the data sets are so huge that effective oversight and fairness guarantees simply can't be offered - with the industry organised as it is. It needs advertisers to force Google to come to the table and deal straightly with them, rather than hiding behind magical utterances of power. "It's the Quality Score, guv" is not an adequate explanation, nor does it allow the market to optimise... I'll probably think more about this... thanks...</description>
		<content:encoded><![CDATA[<p>@Dan - thinking further&#8230; The key difference between this new and still immature market in keywords and impressions and the older media, is third party oversight. <a href="http://www.abce.org.uk/" rel="nofollow">ABC</a> and similar entities provide a third party assurance to advertiser and publisher. Google doesn&#8217;t have that - <a href="http://www.comScore.com/" rel="nofollow">comScore</a> and the like sample the data, and the data sets are so huge that effective oversight and fairness guarantees simply can&#8217;t be offered - with the industry organised as it is. It needs advertisers to force Google to come to the table and deal straightly with them, rather than hiding behind magical utterances of power. &#8220;It&#8217;s the Quality Score, guv&#8221; is not an adequate explanation, nor does it allow the market to optimise&#8230; I&#8217;ll probably think more about this&#8230; thanks&#8230;
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-44209</link>
		<pubDate>Thu, 03 Apr 2008 18:07:31 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-44209</guid>
					<description>@Dan - Hmm - I'm offering you techniques to use the existing system. Not advocating that it is right. If you read the rest of this blog, you'll see that I rant against assumptions that Google makes for how advertisers and would-be purchasers use it. There seems to be no proper understanding of how searches evolve and the interaction of search with the buying process, etc. 

I tend to agree with you. I've got clients who have seen impressions crash, but who were getting leads at a viable rate. After the latest round of changes, they don't get the business, because they aren't being shown on relevant searches, at Google's whim. Most of those suffering worst from the changes have niche markets, and look for less than 1% of the audience that uses a specific keyword. They have a built in low CTR, but can pay a lot, because the service is valuable to that segment - so they used to appear. That low intrinsic CTR doesn't invalidate the advert, and does help the small percentage of searchers who want to see that one niche advert. That this skinny subsegment of the audience want to see it, is borne out by the high conversion rate. However Google apparently only looks at the clicks.  Clicks are revenue, and IMO, Google appear to be trying to optimise revenue, not sales or businesses. It's the Nash Equilibrium, played out in advertising - they seem to think it is a numbers game, not a psychology game - Macroeconomics rather than microeconomics. 

I've already started pushing my clients towards alternatives - I expect to write up some of the experiences. ;)

Hope to hear your lucid thoughts here again :)</description>
		<content:encoded><![CDATA[<p>@Dan - Hmm - I&#8217;m offering you techniques to use the existing system. Not advocating that it is right. If you read the rest of this blog, you&#8217;ll see that I rant against assumptions that Google makes for how advertisers and would-be purchasers use it. There seems to be no proper understanding of how searches evolve and the interaction of search with the buying process, etc. </p>
<p>I tend to agree with you. I&#8217;ve got clients who have seen impressions crash, but who were getting leads at a viable rate. After the latest round of changes, they don&#8217;t get the business, because they aren&#8217;t being shown on relevant searches, at Google&#8217;s whim. Most of those suffering worst from the changes have niche markets, and look for less than 1% of the audience that uses a specific keyword. They have a built in low CTR, but can pay a lot, because the service is valuable to that segment - so they used to appear. That low intrinsic CTR doesn&#8217;t invalidate the advert, and does help the small percentage of searchers who want to see that one niche advert. That this skinny subsegment of the audience want to see it, is borne out by the high conversion rate. However Google apparently only looks at the clicks.  Clicks are revenue, and IMO, Google appear to be trying to optimise revenue, not sales or businesses. It&#8217;s the Nash Equilibrium, played out in advertising - they seem to think it is a numbers game, not a psychology game - Macroeconomics rather than microeconomics. </p>
<p>I&#8217;ve already started pushing my clients towards alternatives - I expect to write up some of the experiences. ;)</p>
<p>Hope to hear your lucid thoughts here again :)
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Dan</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-44168</link>
		<pubDate>Thu, 03 Apr 2008 15:22:24 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-44168</guid>
					<description>I appreciate your thoughts, but you have to realize that while your technical background is most extensive and impressive, you do have consider the marketer's point of view.

In every other medium, the advertiser controls the message, who sees, what price they pay. The vendor, whether it be newspaper, radio, tv, outdoor - has no place or business in deciding  what ad is worthy of space, and removing ads because they feel it's not relevant. This is the height of arrogance.

Imagine if a newspaper said "No, you cannot advertise in the sports section, because your ad isn't relevant to people who read sports."; or perhaps a a radio station executive tell a person "your ads cannot run on our talk station, your product has nothing to do with the content of our talk shows.".

Similarly, can you imagine a radio station or tv station saying "Well, we are going to increase your ad rates on all future campaigns because those ads you just ran didn't produce enough results."?  Crazy! Yet Google is doing just that. Any Sales Manager in the country for any medium would be fired instantly for saying such a thing. While the vendor does care about your success, they won't punish your or make you reopen your doors if you have a campaign that performs poorly.

And finally, what mass media company would have the audacity to tell a business how to design, decorate, stock or even locate their store? Again, Google does.

It's entirely absurd. Google doesn't know the first thing about my business, our marketing plans, or who we want to reach.  Google hasn't spent the last 15 years fine-tuning marketing campaigns that are demographically and psychographically targeted to specific markets. They are a channel through which we can reach potential clients.

Google's stock has dropped nearly 40% in the last six months, from over $700 dollars to just $450. This decline will continue until they stop pretending that they know everything about everything. I, like most Marketing people, don't have the time to hassle with a company that is so self-righteous and egotistical. There are plenty of mediums out that would love a piece of the pie; and now thanks to Google's policies, they're going to get it.

Best wishes to your Jeremy.</description>
		<content:encoded><![CDATA[<p>I appreciate your thoughts, but you have to realize that while your technical background is most extensive and impressive, you do have consider the marketer&#8217;s point of view.</p>
<p>In every other medium, the advertiser controls the message, who sees, what price they pay. The vendor, whether it be newspaper, radio, tv, outdoor - has no place or business in deciding  what ad is worthy of space, and removing ads because they feel it&#8217;s not relevant. This is the height of arrogance.</p>
<p>Imagine if a newspaper said &#8220;No, you cannot advertise in the sports section, because your ad isn&#8217;t relevant to people who read sports.&#8221;; or perhaps a a radio station executive tell a person &#8220;your ads cannot run on our talk station, your product has nothing to do with the content of our talk shows.&#8221;.</p>
<p>Similarly, can you imagine a radio station or tv station saying &#8220;Well, we are going to increase your ad rates on all future campaigns because those ads you just ran didn&#8217;t produce enough results.&#8221;?  Crazy! Yet Google is doing just that. Any Sales Manager in the country for any medium would be fired instantly for saying such a thing. While the vendor does care about your success, they won&#8217;t punish your or make you reopen your doors if you have a campaign that performs poorly.</p>
<p>And finally, what mass media company would have the audacity to tell a business how to design, decorate, stock or even locate their store? Again, Google does.</p>
<p>It&#8217;s entirely absurd. Google doesn&#8217;t know the first thing about my business, our marketing plans, or who we want to reach.  Google hasn&#8217;t spent the last 15 years fine-tuning marketing campaigns that are demographically and psychographically targeted to specific markets. They are a channel through which we can reach potential clients.</p>
<p>Google&#8217;s stock has dropped nearly 40% in the last six months, from over $700 dollars to just $450. This decline will continue until they stop pretending that they know everything about everything. I, like most Marketing people, don&#8217;t have the time to hassle with a company that is so self-righteous and egotistical. There are plenty of mediums out that would love a piece of the pie; and now thanks to Google&#8217;s policies, they&#8217;re going to get it.</p>
<p>Best wishes to your Jeremy.
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-43564</link>
		<pubDate>Wed, 02 Apr 2008 21:30:21 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-43564</guid>
					<description>Hi Dan, if you have a sizeable budget, you do what other companies with your size do... open several accounts and migrate the damaged keywords to a new, clean account. While there does appear to be some stain affecting low performing web sites, a poor history in an account attaches more to the keyword. This performance factor has been around since at least 2005 - it isn't a new feature. I've been in the strange position of having two clients with a shared ownership, tackling different segments in the same market - so I've been able to see differences, even at the level of impression rates for the same keywords, based on the account performance.

You aren't going to get transparency. How Google controls the value extraction from the market is partially a result of their asymmetric information holding. They are the gateway to value. Making changes that can cause a single keyword to vary in impression volume by a factor of five over a few weeks, when other market signals indicate that the demand is unchanged,  without any announcements and even offering denials of changes, is quite reprehensible, IMO.

I am aware of the statistical flux of the Internet. I've been using it since the 1980's. I have a degree in Cybernetics. I've run large scale experiments with noisy data (200+ houses, with more than a thousand sensors, for several years). I know noise. I *like* noise and signal extraction. There's a signal here, which is that the Black Box has changed the transfer function.</description>
		<content:encoded><![CDATA[<p>Hi Dan, if you have a sizeable budget, you do what other companies with your size do&#8230; open several accounts and migrate the damaged keywords to a new, clean account. While there does appear to be some stain affecting low performing web sites, a poor history in an account attaches more to the keyword. This performance factor has been around since at least 2005 - it isn&#8217;t a new feature. I&#8217;ve been in the strange position of having two clients with a shared ownership, tackling different segments in the same market - so I&#8217;ve been able to see differences, even at the level of impression rates for the same keywords, based on the account performance.</p>
<p>You aren&#8217;t going to get transparency. How Google controls the value extraction from the market is partially a result of their asymmetric information holding. They are the gateway to value. Making changes that can cause a single keyword to vary in impression volume by a factor of five over a few weeks, when other market signals indicate that the demand is unchanged,  without any announcements and even offering denials of changes, is quite reprehensible, IMO.</p>
<p>I am aware of the statistical flux of the Internet. I&#8217;ve been using it since the 1980&#8217;s. I have a degree in Cybernetics. I&#8217;ve run large scale experiments with noisy data (200+ houses, with more than a thousand sensors, for several years). I know noise. I *like* noise and signal extraction. There&#8217;s a signal here, which is that the Black Box has changed the transfer function.
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Dan</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-42696</link>
		<pubDate>Tue, 01 Apr 2008 21:32:23 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-42696</guid>
					<description>Great story. 

Google won't see a penny more of my Company's million-dollar plus marketing budget until they allow ME to decide what's relevant to my business and what's not.

And to punish/poison an AdWords client for experimenting with keywords is inexcusable. How are we supposed to learn?

Google's ego has gotten out of hand.</description>
		<content:encoded><![CDATA[<p>Great story. </p>
<p>Google won&#8217;t see a penny more of my Company&#8217;s million-dollar plus marketing budget until they allow ME to decide what&#8217;s relevant to my business and what&#8217;s not.</p>
<p>And to punish/poison an AdWords client for experimenting with keywords is inexcusable. How are we supposed to learn?</p>
<p>Google&#8217;s ego has gotten out of hand.
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-42579</link>
		<pubDate>Tue, 01 Apr 2008 08:44:58 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-42579</guid>
					<description>Ok, I've been seeking permission from my clients to aggregate data for publication. I see a decline. It varies by client, and it varies according to segmentation. Large clients are generally least affected. So the sample used implies the conclusions. I'm trying to get a representative sample of my clients. I probably could find a set where the combination resulted in increased impressions - but that would be an unrepresentative sample. On the basis of the evidence that I look at, I'd say that SearchIgnite had a biased sample set :)

You haven't made any argument against my analysis of the buying process and the impact that Search History Permutation has on that. There's not *one* thing that Google has changed, it is a bunch of factors. Sometimes, for some clients, they combine to dreadful effect. For others, the changes are visible but the results are more complex to describe.</description>
		<content:encoded><![CDATA[<p>Ok, I&#8217;ve been seeking permission from my clients to aggregate data for publication. I see a decline. It varies by client, and it varies according to segmentation. Large clients are generally least affected. So the sample used implies the conclusions. I&#8217;m trying to get a representative sample of my clients. I probably could find a set where the combination resulted in increased impressions - but that would be an unrepresentative sample. On the basis of the evidence that I look at, I&#8217;d say that SearchIgnite had a biased sample set :)</p>
<p>You haven&#8217;t made any argument against my analysis of the buying process and the impact that Search History Permutation has on that. There&#8217;s not *one* thing that Google has changed, it is a bunch of factors. Sometimes, for some clients, they combine to dreadful effect. For others, the changes are visible but the results are more complex to describe.
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Epik Paul</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-41817</link>
		<pubDate>Thu, 27 Mar 2008 19:45:13 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-41817</guid>
					<description>Hi Jeremy,

You commented that clicks are declining,  "...This is borne out by comScore reports - declining click volumes, declining count of adverts..." I presume you mean Google Clicks.  That said the comScore report is suspect at best, SearchIgnite found the exact opposite to be true. http://www.mediaweek.com/mw/news/recent_display.jsp?vnu_content_id=1003717516

Additionally, initial reports from eMarketer etc. suggest that Search Advertising is as healthy as ever despite a U.S. economic downturn.</description>
		<content:encoded><![CDATA[<p>Hi Jeremy,</p>
<p>You commented that clicks are declining,  &#8220;&#8230;This is borne out by comScore reports - declining click volumes, declining count of adverts&#8230;&#8221; I presume you mean Google Clicks.  That said the comScore report is suspect at best, SearchIgnite found the exact opposite to be true. <a href="http://www.mediaweek.com/mw/news/recent_display.jsp?vnu_content_id=1003717516" rel="nofollow">http://www.mediaweek.com/mw/news/recent_display.jsp?vnu_content_id=1003717516</a></p>
<p>Additionally, initial reports from eMarketer etc. suggest that Search Advertising is as healthy as ever despite a U.S. economic downturn.
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-41753</link>
		<pubDate>Thu, 27 Mar 2008 13:35:18 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-41753</guid>
					<description>@Paul - Thanks. Yes - generally you'll find education via advertising will need deep pockets. In marketing terms, you may be catching people in late phases as they get ready to buy, and shifting them backwards in the buying process, where they are researching again. You might consider changing the way that the site works, or use a different site - there's no reason to have one-business/one-site. If you appeal to different segments or find that walk-ons work differently from paid search, then you do what you need to, to win the business.</description>
		<content:encoded><![CDATA[<p>@Paul - Thanks. Yes - generally you&#8217;ll find education via advertising will need deep pockets. In marketing terms, you may be catching people in late phases as they get ready to buy, and shifting them backwards in the buying process, where they are researching again. You might consider changing the way that the site works, or use a different site - there&#8217;s no reason to have one-business/one-site. If you appeal to different segments or find that walk-ons work differently from paid search, then you do what you need to, to win the business.
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Paul Evans</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-41737</link>
		<pubDate>Thu, 27 Mar 2008 12:09:46 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-41737</guid>
					<description>Jeremy,

Apologies on the link being there - I hadn't noticed that names linked directly to the website - feel free to remove it from my previous post.

I don't believe I have poisoned my account - I did have a play in a different account before my current one - but it could well be worth a try again.

You could well be right about some of my competitors relying on the captured customer idea. In part because their products are so poorly made that they break.

We are trying to sell quality over quantity, which would suggest that we are probably incorrect in advertising on the generic keywords, and we should retreat into a quality/luxury niche.

I think what I have found is that AdWords is not a good place to try and educate people. If they already have a fixed idea of what they're after, even if you are offering something much better, they probably won't want to listen.

I think I will target those specific areas that I am trying to appeal to and also try out other areas such as MSN.

Paul</description>
		<content:encoded><![CDATA[<p>Jeremy,</p>
<p>Apologies on the link being there - I hadn&#8217;t noticed that names linked directly to the website - feel free to remove it from my previous post.</p>
<p>I don&#8217;t believe I have poisoned my account - I did have a play in a different account before my current one - but it could well be worth a try again.</p>
<p>You could well be right about some of my competitors relying on the captured customer idea. In part because their products are so poorly made that they break.</p>
<p>We are trying to sell quality over quantity, which would suggest that we are probably incorrect in advertising on the generic keywords, and we should retreat into a quality/luxury niche.</p>
<p>I think what I have found is that AdWords is not a good place to try and educate people. If they already have a fixed idea of what they&#8217;re after, even if you are offering something much better, they probably won&#8217;t want to listen.</p>
<p>I think I will target those specific areas that I am trying to appeal to and also try out other areas such as MSN.</p>
<p>Paul
</p>
]]></content:encoded>
				</item>
	<item>
		<title>by: Jeremy Chatfield</title>
		<link>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-41718</link>
		<pubDate>Thu, 27 Mar 2008 09:49:32 +0000</pubDate>
		<guid>http://blog.merjis.com/2008/03/24/adwords-relevance-wtf/#comment-41718</guid>
					<description>@Paul - your link is "not child safe" - I should add a WP tag to show this... Hmmm....

Yes, I know the industry. Suggestion - are you registering people for email? I suspect that many of your competitors are relying on a "captured customer" to get repeat sales and amortising the value over that, rather than a cost-for-one-sale, that is, use the lifetime value of conversion. This can justify a higher-than-you'd expect click cost.

Also, have you "poisoned" your account? If you had a period, typically when learning, when you had a low CTR, you might have driven up the base price you pay. Annoyingly, the only real way to tell is to see how a new account, starting with your best shot, performs. Large businesses will migrate keywords to new accounts, a process sometimes known as "keyword rehabilitation" or "keyword recovery" to get a better AvCPC. Small businesses - harder to do. 

You might also want to consider MSN (esp targeting MSN Personals), Yahoo, Advertising Exchanges and Domain Parks. There's stuff to do after you've exhausted those, too :)</description>
		<content:encoded><![CDATA[<p>@Paul - your link is &#8220;not child safe&#8221; - I should add a WP tag to show this&#8230; Hmmm&#8230;.</p>
<p>Yes, I know the industry. Suggestion - are you registering people for email? I suspect that many of your competitors are relying on a &#8220;captured customer&#8221; to get repeat sales and amortising the value over that, rather than a cost-for-one-sale, that is, use the lifetime value of conversion. This can justify a higher-than-you&#8217;d expect click cost.</p>
<p>Also, have you &#8220;poisoned&#8221; your account? If you had a period, typically when learning, when you had a low CTR, you might have driven up the base price you pay. Annoyingly, the only real way to tell is to see how a new account, starting with your best shot, performs. Large businesses will migrate keywords to new accounts, a process sometimes known as &#8220;keyword rehabilitation&#8221; or &#8220;keyword recovery&#8221; to get a better AvCPC. Small businesses - harder to do. </p>
<p>You might also want to consider MSN (esp targeting MSN Personals), Yahoo, Advertising Exchanges and Domain Parks. There&#8217;s stuff to do after you&#8217;ve exhausted those, too :)
</p>
]]></content:encoded>
				</item>
</channel>
</rss>
