AdWords For Domains Garbage Traffic

[dannysullivan]

While I have you -- did you want to elaborate on the "as Yahoo pretty much does" part?

So let's go back to the days before we had contextual ads from Google. Those who want a catch up, see my Google Throws Hat Into The Contextual Advertising Ring article from 2003.

Google didn't start contextual advertising but they certainly helped blow it up into a major separate channel. Before them, we software that did contextual (Gator, eZula) and companies doing placement like Applied Semantics and Sprinks (ah, Sprinks, remember Sprinks?).

But both Google and Overture (at the time) were doing contextual without formal programs. They experimented with putting search ads in newsletters, or putting them into "directory" formats on parked domains or into directory pages at AOL, generating ads based on the directory topic. None of this was search targeting. But neither was it done through contextual programs that you could buy. It happened as part of your search buy.

When Google launched AdSense, suddenly the contextual stuff moved into its own channel though the pricing for that wasn't in its own marketplace. Those who've never seen Andrew speak have missed a real treat as for years he and others on his contextual panel had to painfully waste time explaining to marketers how to better segregate contextual from search and turn it into its own marketplace in spite of Google. Then we got SmartPricing or whatever and finally what should have happened from the beginning, the ability to bid independently. We waited so long because initially Google thought we'd be too confused over having such options. I guess (sarcasm intended) we're all now grown up enough to buy them separately.

So back to what you asked, Andrew. As you know, Yahoo fairly recently rolled out its own AdSense rival in the Yahoo Publishers Network. Rather than do the Google thing, they made YPN an independent buy from search, or at least that's my understanding.

The "pretty much" part comes in because clearly domain traffic hasn't been put into YPN. Plus, I suspect there's probably some other traffic such as those old style put ads on a directory page types of things that are more contextual but still count as search.

[Chris_D]

Ottaviani wrote:

Google sent the following response to our complaint about bogus/fraudlent traffic. In the response they did not even address a single issue we compained about and did not reference a single piece of data. They have been unwilling to refund even a penny of our money.

Thats a strange position for a company with a mantra of 'do no evil' to take - isn't it?

Especially when Google also recommend that when you deal with independent 3rd party SEO firms - you should:

Make sure you're protected legally.
For your own safety, you should insist on a full and unconditional money-back guarantee. Don't be afraid to request a refund if you're unsatisfied for any reason...

http://www.google.com/intl/en/webmasters/seo.html

Maybe its time for Google to start drinking its own Kool Aid.... or SEO firms to recommend to clients that they seek a direct GOOGLE contract for SEM services offering a "full and unconditional money-back guarantee... and a refund if you're unsatisfied for any reason"

[Ottaviani]

During this whole saga we were told by two different AdWords reps that to solve our problem we should create a seperate campaign just for the "Search Network" lisiting. This is basically a run around because Google has not seperated our the program like the eventually did with contextual ads. By taking their advice I am not running four different campaigns because they cannot organize their own programs properly.

My Campaigns:

Google Search
Google Contextual Ads
Google Site Ads
(now) Google Search Network Ads

Oh but wait... YOU CAN'T EVEN CREATE A UNIQUE Search Network Campaign. When you try to do it you must click on the Google search box in or to get the network partners but you can't get the network partners without the search. Which means you can't even do what google recommends.

I feel like I am in the twighlight zone... meanwhile google still has our money they overcharged us for and if it was not for the fact that I have now opted out of their search parnter network I would still be getting charged for parked domain traffic somewhere outside of the US with a ROI of 600 clicks per sale. Compared to google search which is 7 clicks per sale.

[Discovery]

It's obvious that if google or Y! let us eliminate ALL poor performing domain sites and content partner sites they would lose a great deal of revenue.

30-50% click fraud? No, certainly not in my industry. 10-15% may be more like it, but that still represents close to 200k a year for us and our clients. That's an amount I am willing to fight hard for.

The pressure for them to change is building, but the current corp. response is that they would certainly remove poor content partners if they could determine who they are. So we tell them individually, then they throw up a huge smoke screen such as send me a ton of log data and we will research it for the next 10 days. Ultimately, they give you a vague explaination such as we found some questionable traffic therefore we're going to give you a token $500 refund. (most of us have lost well over $500 in a day on this traffic).

It is easy for these companies to clutter, confuse and dilute the issue to the point of no resolution when they are only dealing with us on an individual basis.

So my thought is this:
Create a solution that benefits all parties.
Use our collective buying power/influence to leverage change.
Force the issue by presenting our cummulative data in a clear and concise report.
Simplify the process by collecting our data and submitting one report monthly of the top content/domain offenders.

In summary, if they are unwilling to compile a list of poor performers and eliminate them, we need to cut through the smoke screen and force the issue by generating and presenting our own cummulative performance results. It would be hard for the bigs to ignore such a highly visible and well formed report.

The Criteria for poor performers could be:

Minimum of 250 clicks
A conversion of under 2%
and a gaurantee that there were no server/website issues on our end.

I'm not sure how to implement something like this, but I certainly would be glad to help build it, submit my data and sign off on it's submission to Google.. and Y! for that matter. If the solution works perhaps it could grow into a type of impartial watchdog service.

Do you think this idea has merit?

Discovery

[Discovery]

The data we compile could also have other huge benefits. Such as if we have a database of all these websites we could then also easily display a list of which sites belong to which advertising program.

ihaveacaseoftheDTCshakes.com = Overture Sitematch
pu.com = Miva all
Iparkedmydomainformoney.com = Google content

Discovery

[Mel66]

LOL Discovery!

[andrewgoodman]

Originally Posted by dannysullivan

So let's go back to the days before we had contextual ads from Google. Those who want a catch up, see my Google Throws Hat Into The Contextual Advertising Ring article from 2003.

Google didn't start contextual advertising but they certainly helped blow it up into a major separate channel. Before them, we software that did contextual (Gator, eZula) and companies doing placement like Applied Semantics and Sprinks (ah, Sprinks, remember Sprinks?).

But both Google and Overture (at the time) were doing contextual without formal programs. They experimented with putting search ads in newsletters, or putting them into "directory" formats on parked domains or into directory pages at AOL, generating ads based on the directory topic. None of this was search targeting. But neither was it done through contextual programs that you could buy. It happened as part of your search buy.

When Google launched AdSense, suddenly the contextual stuff moved into its own channel though the pricing for that wasn't in its own marketplace. Those who've never seen Andrew speak have missed a real treat as for years he and others on his contextual panel had to painfully waste time explaining to marketers how to better segregate contextual from search and turn it into its own marketplace in spite of Google. Then we got SmartPricing or whatever and finally what should have happened from the beginning, the ability to bid independently. We waited so long because initially Google thought we'd be too confused over having such options. I guess (sarcasm intended) we're all now grown up enough to buy them separately.

So back to what you asked, Andrew. As you know, Yahoo fairly recently rolled out its own AdSense rival in the Yahoo Publishers Network. Rather than do the Google thing, they made YPN an independent buy from search, or at least that's my understanding.

The "pretty much" part comes in because clearly domain traffic hasn't been put into YPN. Plus, I suspect there's probably some other traffic such as those old style put ads on a directory page types of things that are more contextual but still count as search.

Danny, the picture unfortunately is not as clear as one would like over at Yahoo, if you're a buyer of traffic. Content Match is opt-out within the interface. But a variety of oddball types of traffic have always been mixed in on both sides of the ledger. We've seen this in our own accounts, but have done quite a bit of research on what others are saying about their experience, and the frustration with Yahoo seems pretty widespread.

They do a poor job of disclosing content partners for Content Match. But more importantly Yahoo has been very up & down about the types of sites they allowed to be part of the "search" or "precision match" side. Sometimes they are crap or fraud sites, there's no other way to describe them.

When you sit on the phone with your (large financial institution) client with Y! on the phone as well, it's beneath everyone on the phone to be having to discuss $2k worth of nonconverting clicks that came from a site of such laughable ethical standards (and confused consumers into thinking it was a government site, when the owners weren't the ones generating the clicks). Like G, Y will take refund requests on a case by case basis. But when you have obviously deceptive sites being treated as normal traffic, why did they get approved in the first place? I hear this complaint again and again from advertisers who are shutting down their accounts because they have no confidence in what they're doing over there, and costs are out of control. Obviously those problems don't affect everyone (and it's sad to say that the persistent long-timers in the biz get their case heard eventually which just means that less vigilant advertisers may be getting even more of the bad stuff because someone's gotta get it).

The YPN is a brand new deal and will cause more headaches if not well-administered. We are still dealing with all the old headaches.

It shouldn't be a game of who's "worse" here. That'll wind up being a matter of opinion, or spin. But just don't let the opinion or spin come from either Y or G. Let it be direct information from server logs about the referring URL's. Constant pressure from the advertiser community has reduced the degree of shell-gameyness -- without the years of complaints, we'd still be getting GoTo.com-quality affiliate traffic.

To be clear: the rant about content is fairly standard. But the rant about SEARCH including non-search traffic is counterintuitive and few advertisers grasp this. Both Y and G do this, but IMHO Y is worse. In point of fact, G allows you to opt out of the "search network," thus giving you the OPTION of getting GOOGLE SEARCH PLACEMENT ONLY. Y does not allow this. Ergo, it looks very much to me like G is the better platform, if you understand how to opt out of the things you don't want.

[dannysullivan]

It shouldn't be a game of who's "worse" here.

Believe me, that's not what I intended. My original article came off a review that Microsoft researchers did looking at domain traffic driven by Google. My response was that Google, as well as Yahoo and likely Microsoft in the future will all need to correct this.

I think advertisers should have full control over where ads will run, and it shouldn't be a guessing game as to whether something is contextual or search or whatever at Google, Yahoo or any other provider. And I agree, there's plenty of that guessing at both places right now.

[ApogeeWebLLC]

Originally Posted by Ottaviani

2. The proportion of the search traffic is way out of line (Google 20%, domainsponsor 29%, information 41%)

FYI, more than a year later I'm seeing this sort of garbage traffic. For a single, exact match keyword, I'm seeing up to 72% of the clicks coming from:

searchportal.information.com

When I reported the unusual search traffic to Google, I received a similar, canned response. I've done some research and posted the findings on my blog. I won't drop a link. If you want more details, my blog is called Apogee Weblog. You can also find it on the Business or Marketing channels of 9rules.

Check your server logs for this domain! This was only on the Search network. The Content network was turned off. Nasty stuff. Google's really, really irritating me right now. Make sure you're not paying for this garbage traffic!

The company that owns searchportal.information.com is Oversee.net and they also run DomainSponsor, a parked domain program. It also appears like they are the creators of SearchAndBrowse, a spyware/adware program. That's probably how they're able to drive so much traffic. Oversee.net is on a buying spree of domain names, so there'll likely be more garbage traffic now than there was a year ago. Again, check your server logs!

[Discovery]

Thanks Apogee for the update. I can't believe that it has been a year (or more) that we have been fighting for more transparency into the search and content networks. Yes G has upped their exclusion limit to 500 domains, but 1 junk trafficker can easily create 500 junk sites, and we could spend all our time every day just figuring out what domains they are using today to exclude them. And if the SE's are cloaking much of their traffic we have to manually figure out whos who.

Search engines need to work with us, wean themselves off the money from junk clicks and defeat click fraud.

Being that it is a new year I think this should be the year of transparency.
I envision seeing before the end of 2007 headlines like

"Peer reviewed networks defeat fraudsters"
"The open search engine; transparency kills the junk click"

For now its still a pipe dream, but if this dream does not come true in 07 we may find that PPC will turn foul just as email marketing fell to spam marketing.

Discovery

[ApogeeWebLLC]

Originally Posted by Discovery

Yes G has upped their exclusion limit to 500 domains

Yes, but the site exclusion tool only applies to the Content network. I've seen threads here on SEW and elsewhere that suggest the tool also works for parked domains on the Search network, but that's simply not true. I tried to exclude the garbage traffic when I started seeing spikes on the Search network from searchportal.information.com which appears to power a great number of parked domains. If the traffic was on the Content network, I could live with that. I'd simply block them. Note that I contacted Google's click quality team when I saw this traffic. They denied that the traffic was bogus. Issued a canned response that included this nugget:

"If you aren't satisfied with the value of the traffic from searchportal.information.com, please reply to this email. We can then assist you in preventing your ad from appearing on these pages."

I'd really like to know the origin of the traffic. My hunch is that's it's either coming from the SearchAndBrowse spyware/adware that the company that owns searchportal.information.com appears to distribute. Either that or traffic exchange programs. Google's so focused on automated clickbots that they don't realize that human clicking can also, actually, be artificial.

[Discovery]

Spot on Apogee, there was some talk about blocking being implemented on the search side but Adwords rep cleared that up with a definate "NO BLOCKING ON SEARCH" answer.

Google - Block the content partners if you can decipher who the heck they are. Thats why they cloak the traffic. You have to eat everything on the search side.

MSN - Eat everything - No blocking, period
YSM - Eat everything - No blocking, period

Discovery

[ApogeeWebLLC]

Originally Posted by Discovery

Google - Block the content partners if you can decipher who the heck they are. Thats why they cloak the traffic. You have to eat everything on the search side.

Discovery, I know you've seen my free software tool that at least lets you see where your Google Content network clicks are coming from. That's a start. ;-)

BTW, any objection to me posting a link to my blog post where I give a detailed timeline of the garbage traffic problem from searchportal.information.com?

[Discovery]

Apogee, can your software actually see who is behind the google syndication clicks? I thought that the click went from the "content" partner to googles servers who then washes the information clean then passes that click on to the advertisers website. Thus there is no data to be seen or analyzed within your logs other than syndication.google.com?

I would love to understand how one might be able to gain information deeper than this.

I'm not the MOD here, but if the blog post has worthy, related info and not a sales pitch then I dont see a problem linking.

Discovery

[ApogeeWebLLC]

Okay, here is my post about recent Search network garbage traffic.

Yes, you can see the actual, original source of the content clicks. I'm working on a specific example to show you. Will post in that info in a few minutes...

[Mel66]

Originally Posted by Discovery

Spot on Apogee, there was some talk about blocking being implemented on the search side but Adwords rep cleared that up with a definate "NO BLOCKING ON SEARCH" answer.
Discovery

Actually, this is what AWR said:

Originally Posted by AdWordsRep

So, to clarify - one can exclude any site in our Content network and any parked domain in either the Search or Content network but cannot exclude other sites in the Search network.

However, since this didn't work for Apogee, I would infer that you need to block the individual parked domains, e.g. ibuildmfasitesformoney.com et al, instead of their root or parent domain searchportal.information.com. I don't know this for a fact - just guessing.

It still bugs me that an Adwords representative said at SES that this type of traffic can be blocked if that is not really the case. He even used Oingo as an example - he said you could put them in the site exclusion tool, and you'd stop getting traffic from them. Why would he say that if it's not true? Why would he say "there is not a single site that you cannot exclude from Adwords" when it isn't true? I'm not buying the "well, he was talking about domain traffic in the search network" argument. He knew exactly what people were talking about when they asked about blocking these search partners, because people gave specific examples and named names such as Oingo and even, IIRC, searchportal.information.com, and he said "you can block those now." Either the guy was misinformed (in which case, maybe he shouldn't have been on the panel), or something has changed. Or rather, perhaps something was supposed to change, and then G realized how much $$$ it was costing them?

Something is rotten in Mountain View here.

Melissa

[Mel66]

Originally Posted by ApogeeWebLLC

Okay, here is my post about recent Search network garbage traffic.

Great blog post, Apogee.

Holy h*ll - something IS rotten in Mountain View. Or at least something evil is afoot. This is big, big money for these people, at our expense.

Melissa

[Discovery]

You go after 'em Mel.

Bottom line there is no straight answer, and their is no logical reason that the Bigs cant provide advertisers a way to view all traffic and block any site from search or content. It would just cost them too much money in the short term. Until a disrupter SE comes about with a decent network, and offers just that the bigs will not change.

Now if we had tools like Apogee is talking about and we could clearly identify whos who, advertisers could compile some pretty compelling data to force their hand.

Discovery

[ApogeeWebLLC]

Discovery, See the demo here:

http://www.apogee-web-consulting.com...ContentAds.cgi

Here's an example of one of the googlesyndication URLs from the log file from that demo. I'm manually inserting line breaks after each piece of query data to make it easier to understand. These are long, ugly URLs:

pagead2.googlesyndication.com/pagead/ads?
client=ca-pub-3280446741348688&dt=1128043249746&
lmt=1128043249&
alt_color=ffffff&
prev_fmts=728x90_as&
format=300x250_as&
output=html&
channel=4554257847&
pv_ch=4554257847%2B&
url=http%3A%2F%2Fwww.websitepublisher.net%2Farticl e%2Fhtml_seo%2F&
color_bg=FFFFFF&
color_text=000000&
color_link=DD0A2D&
color_url=DD0A2D&
color_border=FFFFFF&
ad_type=text_image&
ref=http%3A%2F%2Fwww.websitepublisher.net%2Fsucces sful_website%2F&
cc=100&
u_h=1024&
u_w=1280&
u_ah=996&u_aw=1280&
u_cd=32&
u_tz=-240&
u_his=18&
u_java=true

What you need to do is extract the "url=" piece of the query and then urldecode it:

urlencoded:
http%3A%2F%2Fwww.websitepublisher.net%2Farticle%2F html_seo%2F

urldecoded (I'm dropping of the leading http):
://www.websitepublisher.net/article/html_seo/

That's how you can see where the actual refer came from. Free, open source tool is here:

http://www.apogee-web-consulting.com...ntent_ads.html

Does that help?

[Mel66]

But why is the burden of proof on us? That's what the Sausage Manifesto asks, and I agree. The engines should be proving to *us* that their traffic is legit, not the other way around.

However, since most advertisers are more than willing to just keep paying the bills, there is no incentive for the engines to change their ways.

Melissa