homepage Welcome to WebmasterWorld Guest from 54.196.62.132
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Become a Pro Member
Home / Forums Index / Google / Google AdSense
Forum Library, Charter, Moderators: incrediBILL & jatar k & martinibuster

Google AdSense Forum

This 56 message thread spans 2 pages: 56 ( [1] 2 > >     
Reasons Why AdSense Won't Implement a Whitelist
There Must be a Reason, Let's Figure it Out
BigDave




msg:3210631
 7:28 pm on Jan 6, 2007 (gmt 0)

It seems like barely a week goes by that the subject of a whitelist doesn't come up in one thread or another. Google MUST know by now that it is a feature that is very much desired by a certain category of publishers.

Instead of looking at it only from our own points of view, let's try and figure out why Google hasn't given it to us yet. I have a few ideas that might explain part of the problem.

- It's an engineering problem. The current system might not be designed so that the whitelist can easily fit in. Remember a few years ago when google made the huge change to the guts of the search algo, and how painful that was as they rolled in the changes? Can they afford something that painful when it comes to their revenue generation.

There isn't much of a solution to this one other than keep up the pressure.

- They don't want to lose ad revenue due to clueless publishers. Let's face it, some people with no understanding of the system will turn on their whitelist and not populate it fully. They might forget that they have a site and not add it to the list. They might add the adsense code to a new site and not add the site to their whitelist. They might blame Google for something that is their own fault.

I have a solution to this, but most publishers would still complain. Make it an option that has to be turned on. To stay on it must be reviewed at least once a month. Every domain that has served ads using that publisher ID is automatically entered and must be manually deleted.

It would also need the "new sites", that tried to serve ads but failed, to be listed. These would have to be OKed or blocked within a month or the whitelist turns off.

- They just don't believe the rank and file AdSenser would use it.

- Dealing with whitelist issues will add a large manpower requirement. Both hunting down infringing sites and helping the clueless webmasters.

---

Those are the possible reasons that I came up with. Can you think of any others or possible solutions *from Google's point of view*?

 

Hobbs




msg:3210639
 7:34 pm on Jan 6, 2007 (gmt 0)

As I said in the other thread:
[webmasterworld.com...]
I see no reason why they do not listen to publishers and implement the whitelist, none of the above possible reasons makes sense to me, but I hope to hear more possible reasons.

BigDave




msg:3210643
 7:41 pm on Jan 6, 2007 (gmt 0)

They make no sense to you, but do they make sense to Google? That is what matters.

If they implement this, they have to do it right or it could cost them millions of dollars a day. And if it is an engineering design problem, it would have to be a VERY high priority before it gets fixed.

martinibuster




msg:3210648
 7:45 pm on Jan 6, 2007 (gmt 0)

They just don't believe the rank and file AdSenser would use it.

That's an interesting point. Related to that is if the instances of ID Jacking are so rare, it would be hard to justify to upper management the devotion of engineering resources to solve a problem that isn't perceived as a problem by Google.

If it's the case that ID Jacking is a rare occurrence, creating a solution for something that isn't perveived as a problem might send the message to the public and partners that there is a problem. If they're offering a solution, then the natural presumption (at least for me) is that there is a problem. So could it be that Google doesn't feel there is a problem?

[off topic]
I personally feel that a whitelist would make me feel more secure that at least there's one bad thing that wouldn't be affecting my account. It's good to feel secure, that things are locked down. Much like leaving the lights on in the patio makes me feel secure.
[/off topic]

How often ID Jacking happens, and how many innocent webmasters have been booted or inconvenienced is unknown. But the answer to that may be a key to understanding why Google has not instituted a whitelist.

But if there is no problem, then this tool would not be useful to the rank and file, the rank and file may not be interested in it, and the tool may backfire by alarming the rank and file.

Hobbs




msg:3210651
 7:57 pm on Jan 6, 2007 (gmt 0)

So there are 2 ways the whitelist can be implemented:

a) It is proven that the lack of it affects their bottom line or is requested by advertisers.

b) The problem blows out of proportion, gets discussed in the media.. and takes higher priority on their list.

hunderdown




msg:3210653
 8:07 pm on Jan 6, 2007 (gmt 0)

Another possibility may be that from Google's perspective, they don't HAVE a problem with ID hijacking.

As martinibuster suggested, it may be rare, and when it does occur, they may have a way to deal with it.

And those publishers who think ID hijacking has caused them to be banned are simply mistaken....

If the above is true, then why would Google want to go to the effort to "fix" a non-problem, especially, as noted above, putting a solution in place would alarm a lot of people (and might lead to negative media coverage)?

jimbeetle




msg:3210767
 11:09 pm on Jan 6, 2007 (gmt 0)

Another possibility may be that from Google's perspective, they don't HAVE a problem with ID hijacking.

It's a possibility, along with everything else you said. I don't doubt that can be the case from Google's point of view.

But maybe G sticking to it's point of view isn't the smartest way to run a business. A heck of a lot of advertisers are very leery about the quality of clicks coming from the content network. That's a problem. Many publishers are worried about ID hijacking and any other methods that might get them booted from the program through no fault of their own. That's a problem.

Two out of three of the constituencies involved in Adsense believe there are problems. And both believe that Google's very inadequate communication concerning the perceived problems, well, that's another problem.

So, I really don't care if Google, from its point of view believes it has problems or not. Fact is, if advertisers don't want to spend bucks, and if good publishers get booted, then Google's fancy-schmancy ad delivery system is not worth as much as it could be.

From G's point of view that should be considered a problem.

europeforvisitors




msg:3210781
 11:34 pm on Jan 6, 2007 (gmt 0)

As martinibuster suggested, it may be rare, and when it does occur, they may have a way to deal with it.

That's what I suspect. (Google is obviously in a better position to know if it's a widespread problem than we are.)

I can also imagine problems with a whitelist, namely:

1) Getting publishers to use it, as has been discussed elsewhere in this thread;

2) More server overhead.

The fact that this "whitelist" discussion is new (or that, if it isn't new, it's gone unnoticed by most of us on the forum) would suggest that there isn't widespread demand for such a feature by publishers, advertisers, or Google's own fraud-control team.

Play_Bach




msg:3210788
 11:42 pm on Jan 6, 2007 (gmt 0)

I posted about this last year - it's certainly not new to me.
[webmasterworld.com...]

BigDave




msg:3210794
 11:57 pm on Jan 6, 2007 (gmt 0)

So, I really don't care if Google, from its point of view believes it has problems or not.

That is the sort of attitude that keeps disputes from being resolved. You need to care about their point of view if you want to effectively communicate your own point of view. That's Negotiation 101.

If you don't care if Google believes it has problems, then that means that you don't care if there is ever a resolution to what you perceive as a problem. Since any resolution would include Google as the primary player, you had better care what they believe.

celgins




msg:3210882
 2:21 am on Jan 7, 2007 (gmt 0)

Whether there is widespread demand for such a feature by publishers and advertisers or not, I can think of a few technical/logical challenges Google could face if implementing a whitelist:

For example, what if my publisher ID (00000000) is associated with my domain name (www.widgets1.com). Let's say another site owner (www.widgets2.com), publisher ID (111111111), creates a new site (www.widgets3.com) and doesn't add it to his whitelist.

1. What happens if I add his domain name (www.widgets3.com) to my own whitelist?

2. How would Google determine who the site belongs to, since their system now believes my publisher ID (000000000) is tied to (www.widgets3.com)?

3. What if the other site owner suddenly remembers to add his new domain (www.widgets3.com) to his whitelist? Google would need the proper programming/scripts in place to automatically determine if the domain he's adding (www.widgets3.com) is already in someone else's list.

4. How much database space should Google allow for whitelists? Some site owners need to add 100 domains, while some only need to add 1.

5. What if another site owner (publisher ID 2222222222) adds my domain name (www.widgets1.com) to his own white list? How does Google determine which publisher's ID actually belongs with the domain name? Could they use the ... 'this site owner added the domain name first, so let's kick the second guy' ... scenario?

Maybe, but what if it's a case where #1 and #2 are happening?

Seems like a lot of additional server/personnel overhead, and maybe Google has already thought of it.

Think about it: In the beginning, Google did not offer to implement an Adsense system where they allowed publishers to create 50 Adsense accounts - because it was easier for them to allow you to place the same Adsense account on 50 different websites.

With the former option, they would have to maintain and track 50 different sites (which they would probably setup as channels under one account).

With the latter option, they only have to track 1 site (1 Adsense publisher ID) and hope you're adding it to your own websites.

KenB




msg:3210943
 3:41 am on Jan 7, 2007 (gmt 0)

I suspect the primary reason Google doesn't implement a white list is that they don't see how it would improve their bottom line. Although I really want a white list ability, I'd have to agree it is hard to justify adding such a feature from Google's economic standpoint. Really it would be more of a goodwill feature for publishers and since Google has no real competition, they don't have a need a feature like this to create a goodwill competitive edge.

RonS




msg:3210997
 5:24 am on Jan 7, 2007 (gmt 0)

God the simplest way to do this would be for google to email the advertiser the first time a new doamin shows up with an impression, and give them a links to either OK or REJECT the domain.

Of course all of the CACHING search engines provide a complexity, don't they. I'm sure with just a little bit of work Google could figure that one out...Seems liek they figured out their own robots.txt caching issue. :)

Anyway, they can also make the whitelist/blacklist of domains a function within the ADSense system, but this would be a quick way to automate it and to reduce the risk of a webmaster not adding a site and to make the rejection of a site automagic.

It would also allow a webmaster to not show ads on chached pages -- or even to turn off a domain for a while, say during a click attack or a slashdotting.

fredw




msg:3211001
 5:36 am on Jan 7, 2007 (gmt 0)

How about, instead of a whitelist of domain names, we had the ability to set our web server's IP address? Then webmasters with many domain names but only one server would only have one setting to enter.

mzanzig




msg:3211044
 7:03 am on Jan 7, 2007 (gmt 0)

I think the core question is - what does Google GAIN if they implemented such a feature? Would they earn a single buck more, or would it be just implementation cost to keep publishers happy?

From a short-term perspective, they gain virtually nothing by implementing this. Agreed, a few bucks here and there might roll in, but all-in-all, I just see implementation cost. Combined with the technical/logical problems outlined above, I think they just do not touch this issue. They have fine-tuned their algos to alert them when things seem to go wrong, and then they deal with it (i.e. either de-activating the account or have a human look at the issue).

From a long-term perspective, it might be worthwhile to improve publisher satisfaction.

But now put yourself in the shoes of the Über-Google Head of Product Management. You see, it's always the same. Even Google has limited resources. They can not implement ALL features. Not enough skilled programmers or certain design limitations. They can't. And the tech folks will tell him this - "You can't have everything, make a decision." So he (or someone from his team) needs to prioritize the features to be implemented in 2007. He will probably ask his product manager, how important is this feature to Google in the long-term? Now, ask yourself: how important is it, really? Would you rather have, say, an improved filter or the domain whitelist?

And truth is, even =we= (at WW) would have problems prioritzing this feature high enough to get it done.

europeforvisitors




msg:3211048
 7:07 am on Jan 7, 2007 (gmt 0)

It isn't just the cost of implementing the feature; it's also the extra microseconds of time for each ad unit served.

Hobbs




msg:3211063
 7:38 am on Jan 7, 2007 (gmt 0)

No, it does not have to delay ad serving as it can be done in the back end before rewarding the publisher for the click, let the ads be served but charge advertiser and reward publisher only for verified clicks and impressions, which they do one way or the other now anyway, they don't show us clicks and earnings real time.

I can't believe people actually could fail to enter the name of their domain name correctly, they are webmasters and managed to open an account, enter their name, address, email, password.. but when it comes to entering a domain we expect mistakes? Yes the algo has to make conditions for when this happens, but I don't think it is too complicated for Google to achieve or for a novice webmaster to enter his own domain name.

Hobbs




msg:3211065
 8:13 am on Jan 7, 2007 (gmt 0)

By the way, they have the verify domain ownership bits working well in Google Analytics.

Pengi




msg:3211086
 9:03 am on Jan 7, 2007 (gmt 0)

I think celgins' post shows how there may be a number of practical issues that would need to be well thought through before implementing a white list. Maybe not just the issues identified; I'm sure there could be lots of different scenarios that would need to be identified, properly analysed and catered for by the solution.

As EFV points out, it is not just the cost of implementing a solution that needs to be considered, it is its impact on the performance of the AdSense service. Therefore Hobbs is quite right, IMHO, in indicating that this must be an "after the event" activity rather than a check conducted in real time.

What exactly is the problem? I'm struggling to understand the motivation for the type of ID highjacking we're discussion here.
- My ID on someone else's domain? Surely this directs revenue to my account. Why would anyone else want to do this? The only reasons I have been able to do this, are as a possibly means of the ID owned disguising the ownership of fraudulent clicks (i.e. disown the domain "someone else put my ID there") or as a deliberate attempt to get another account holder banned.
- Someone else's ID on my domain? Yes I can see lots of motivation for this, but it should be relatively easy to trace.

swa66




msg:3211132
 10:34 am on Jan 7, 2007 (gmt 0)


- My ID on someone else's domain? Surely this directs revenue to my account. Why would anyone else want to do this? The only reasons I have been able to do this, are as a possibly means of the ID owned disguising the ownership of fraudulent clicks (i.e. disown the domain "someone else put my ID there") or as a deliberate attempt to get another account holder banned.

Some competitor might try to ban you, and while at it, might send loads of click that do not covert and thus cause your real revenue to be smartpriced.


- Someone else's ID on my domain? Yes I can see lots of motivation for this, but it should be relatively easy to trace.

This is much more common that you might think: visitors having their machine infected with spyware and worse malware can have bits installed that change the publisher ID on the fly.

The solutions seems obvious: Google already has a proof of website ownership in production. Tie it in that we can opt our account to only work on sites we've proven are ours to our adsense (avoiding problem #1 above).

Tie it in that a domain that's proven to be from somebody can only have ads from that adsense account (avoiding problem #2 above).

As far as bad press goes: fixing things for the better isn;t bad press, that's good press.

frox




msg:3211235
 2:00 pm on Jan 7, 2007 (gmt 0)

First of all, I see the whitelist mechanism as a quite simple one:

[X] Activate whitelist:
Allowed domains:
wwww.a.com
www.b.com
www.c.com

The ads are displayed anyway, but if you activate the whitelist you only get money if the Pub-id is yours and the domain is listed.

They don't want to lose ad revenue due to clueless publishers.

They might even EARN ad revenue...

As a publisher, I would be happy to say: "if it's not in the whitelist, Google can take all the money"

Also, think of the small but constant share f income you get from google cache....

Processing power: That's very little.
Think about the complex algos they implement now (keyword targeting, ads ranking, fraud filtering etc.) and the triviality of this problem (referrer domain in list).
Also, it could be done not in real-ime, i.e. at the end of the months my clicks get discounted of all the clicks coming from non-allowed domains

Celgins, the problems you list are fictitious: you only get money if the pub-id is yours AND the domain is listed.
So, there is not way another webmaster can "steal" your clicks, ans the pub-id is yours anyway.

The only possible explanation is that they don't feel it as a real problem.

A strong problem cold also be that the only way to see "on what domain the ads are being shown" would be to look at the referrer for the HTTP GET request for the javascript.

As we all know, the referres are VERY easy to forge, so this would not protect against malicious attacks done on purpose to hurt the publisher.

celgins




msg:3211283
 3:34 pm on Jan 7, 2007 (gmt 0)

God the simplest way to do this would be for google to email the advertiser the first time a new domain shows up with an impression, and give them a links to either OK or REJECT the domain.

I'm guessing you meant to say: "Email the publisher the first time a new domain shows up ..." Either way, a person entering your domain name into his whitelist will simply answer his email by saying, "Yes... some other person's domain name is actually mine and should belong in my whitelist"

the problems you list are fictitious: you only get money if the pub-id is yours AND the domain is listed. So, there is not way another webmaster can "steal" your clicks, ans the pub-id is yours anyway.

This still doesn't solve the: "first come, first serve" problem. If I register a new domain name, but never add it to my existing Adsense accounts' whitelist:

What if someone else whitelists your domain name before you do? How does Google know who the domain name belongs to? Will they assume that the first person to add it to a whitelist is the true owner? Do they conduct WHOIS queries to lookup technical contact info? How would they determine that you're the true domain owner when 3 months later, you try to add it to your own whitelist?

This goes back to #3 in my first post:

"3. What if the other site owner suddenly remembers to add his new domain (www.widgets3.com) to his whitelist? Google would need the proper programming/scripts in place to automatically determine if the domain he's adding (www.widgets3.com) is already in someone else's list."

europeforvisitors




msg:3211289
 3:49 pm on Jan 7, 2007 (gmt 0)

The only possible explanation is that they don't feel it as a real problem.

Given their data and experience in dealing with publishers from all walks of life, they're certainly in a better position than we are to know if (a) publisher ID theft is a widespread problem, and (b) the cure is likely to be more disruptive than the disease.

Hobbs




msg:3211304
 4:31 pm on Jan 7, 2007 (gmt 0)

What if someone else whitelists your domain name before you do? How does Google know who the domain name belongs to?

I am guessing you don't use Google Analytics, they have a way to do it & it is already cleanly sorted out.

jimbeetle




msg:3211357
 5:13 pm on Jan 7, 2007 (gmt 0)

If you don't care if Google believes it has problems, then that means that you don't care if there is ever a resolution to what you perceive as a problem. Since any resolution would include Google as the primary player, you had better care what they believe.

Yeah, a throwaway sentence on my part, not to good in helping to get my point across.

I do care about what Google believes, but I do believe that "maybe G sticking to it's point of view isn't the smartest way to run a business".

The path to resolution is first identifying a problem. Advertisers perceive problems, publishers perceive problems. Google continues to ignore the issues both parties have raised. You can't have resolution without communication.

The ball's in Google's court.

europeforvisitors




msg:3211419
 6:28 pm on Jan 7, 2007 (gmt 0)

The path to resolution is first identifying a problem. Advertisers perceive problems, publishers perceive problems.

Yes, but Google doesn't have to rely on perceptions--it's the only one of the three parties (advertisers, publishers, and the ad network) that can actually determine whether there's a real problem.

Also, let's be honest: If Google were to publicly say, "Our data indicates that publisher ID theft isn't a significant problem," how many (if any) of the vocal whitelist proponents would say "OK, thanks" and move on?

Hobbs




msg:3211435
 6:48 pm on Jan 7, 2007 (gmt 0)

zero.

But none of us got banned for it, it would feel very different if any of us did, how long till the word spreads that this is an easy way to switch your competition to shoot the monkey ads?

BigDave




msg:3211443
 7:16 pm on Jan 7, 2007 (gmt 0)

Google has a very good reputation when it comes to dealing with widespread *actual* problems. Their reputation on communicating with advertisers and publishers is not as good as it could be, but it is better than just about any other giant companies that deal with so many small fry.

As for the idea of serving only to a site that claims ownership of the domain, I just don't see that happening. Google makes too much money from revenue sharing sites, including their own. Hell, that is what the AdSense API is all about.

fredw




msg:3211689
 12:59 am on Jan 8, 2007 (gmt 0)

You know, I always thought a domain whitelist was a really good idea. But celgins brought up some points that I think are unsolvable and unavoidable.

What's to prevent someone from putting YOUR domain in THEIR whitelist? And the first to put it their whitelist can't be depended on either, what if someone puts your domain their whitelist before you get a chance to put it in yours?

I HAVE AN ALTERNATE IDEA:

What if, as an opt in feature, adsense publishers had the ability to put a small code-named file on their server? This code-name would be known only to you and adsense (would not be your adsense publisher id), thus no one could ever retrieve it or duplicate it on their server because they don't know the code-name for the file. Then, if opted in, and adsense does not find this code-named file on the server of the referrer pulling the ads, the ads could be denied. It would not incur a lot of badwidth because once adsense ties the referer and the code-named file together, it never has to get it again.

celgins




msg:3211691
 1:10 am on Jan 8, 2007 (gmt 0)

I am guessing you don't use Google Analytics, they have a way to do it & it is already cleanly sorted out.

Actually, I do have Analytics and after reading fredw's post, I think I see what you mean about a possible way Google could handle it.

Similar to Sitemaps and Analytics, they could use an XML-configured file or something similar for verification.

There may still be some technical concerns with that scenario.

This 56 message thread spans 2 pages: 56 ( [1] 2 > >
Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Google / Google AdSense
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved