Not challenging the complaints about click fraud from Botswana by pointing out that AdWords can be geo targeted is a perfect example of incomplete reporting.

Is there actually any technical difficulty with a computer in Botswana committing click fraud on your US-only ads? Geotargetting controls (we hope) how Google displays ads, but is there any proof that Google actually supplies the other half of the solution by applying the same geotargetting filter to clicks that come in?

The article doesn't let the truth get in the way of a good story in places, that's for sure - not least with its vague definition of "click fraud".

Eh? Couldn't follow that objection whatsoever. Most people know fraud when they smell it, and the lovely variety of techniques presented in the article is both well-known to exist and quite obviously fraud.

When AdWords started, part of what made them a real innovation was a commitment to quality. Over time, they retained the lip service to quality, while their practices clearly grew to contradict that.

• Before: You can't put ads on non-content pages. Now: Put ads everywhere! Let's see if we can make some bucks off your contact page!
  
• Before: Your website has to have some useful content. Now: Everybody dive in! The water's great! We love parked domains and MFAs!
  
• Before: If you can't write an ad that gets better than .5% CTR, then we're not going to let you annoy users with it. Now: Hey, we'll let you "fix" that "low-quality" ad by just paying us more per click! And of course, if you're a big, generic advertiser, you can splat your crud up for every single search term in the known universe! Searching for "rape prevention"? Here's an ad that says "Find more rape on *bay!".
  

If Google is not willing to forgo some short-term revenue and return to a commitment to quality, then AdSense will be the big long-term loser. C'mon Google -- Act now, supplies of credibility are limited!

Is there actually any technical difficulty with a computer in Botswana committing click fraud on your US-only ads?

If Google can't tell that a computer is in Botswana and thus shouldn't see US-only ads, then the advertiser can't tell that the computer is Botswana.

Conversely, if the advertiser can see (via IP address) that a computer is Botswana, so could Google and thus if the advertiser was using geo targeting then that computer wouldn't have gotten the advertiser's ads.

Geotargetting controls (we hope) how Google displays ads, but is there any proof that Google actually supplies the other half of the solution by applying the same geotargetting filter to clicks that come in?

I don't know about the clicks, but Google does geo-target who gets to see the ads in the first place. Geo-targeting the clicks would be a redundant feature that could be useful to help prevent spammers from trying to use tricks to skirt around geo-targeting of the ads being displayed.

ronburk - you say you can't see a problem with how click fraud is defined in the article. How about this then?

"dummy Web addresses like insurance1472.com, which display lists of ads and little if anything else. When somebody clicks on these recycled ads, marketers such as MostChoice get billed"

That isnīt click fraud. When a journalist starts speaking nonsense like "dummy web addresses" you can only question how much they know about click fraud and legitimacy.

I have over 500 parked domains, all attract type-in high quality traffic, and all display ads. If someone clicks an ad where's the fraud?

I have over 500 parked domains, all attract type-in high quality traffic, and all display ads.

The quality of that traffic is a subjective opinion. Personally I look at parked domain traffic as complete junk and it is one of the reasons I stopped using AdWords.

If someone clicks an ad where's the fraud?

Agreed. I might not think that parked domain traffic is worth anything, but in general it is not fraud, it just isn't traffic I'd want.

I'm at the Atlanta airport and just walked out of one of those book/magazine junk stores. Yup, all over the front page of Business Week "The dark side of on-line Advertising". Couldn't miss it, they are serious about this story.

"The dark side of on-line Advertising"

hidden subtext: The darker side of business journalism.

Not that they'll notice but BW is no longer linked in my investment section.

[edited by: Andreals at 6:11 pm (utc) on Sep. 24, 2006]

The whole article is hyped up rubbish. The line
"A well-executed click-fraud attack is nearly impossible, if not impossible, to detect"
is quite obviously untrue. You require absolutely no computer or business knowledge to detect click-fraud. Simply stop advertising for a week every couple of months and check your income. If your income is down, then continue to advertise, if it stays the same then stop paying for the advertising. It's that simple. To believe that a 40 year old entrepreneur would spend $2 million without frequently checking if the adverts were working is not believable. Sure there is click fraud out there but not to the extent that the article suggests.

hidden subtext: The darker side of business journalism.

Not that they'll notice but BW is no longer linked in my investment section.

Do you have a problem with the messenger or the message?

There is a "dark side" to online advertising. Why shouldn't it be mentioned? Why shouldn't the ad networks and publishers who contribute to the problem be held accountable?

If it's any consolation, cost-per-click advertising, like direct mail, has one advantage over more traditional forms of advertising: Advertisers can track their expenditures and revenues, and if they're smart enough to keep their emotions from getting the better of them, they can base their future CPC ad-buying decisions on measured ROI.

the validity of his clicks, for which he pays up to $8 apiece, has become an obsession. Every day he pores over fresh spreadsheets of click analysis. "I told Yahoo years ago," he says, "'If this was costing you money instead of making you money, you would have stopped this."'

this part of article says I told Yahoo years ago.YPN is just 1 year old.not so?

OMG.... is AdSense is doomed, all because some illinformed third rate piece of what... fiction? I think not.

Meanwhile, I'll be sitting in the corner tallying up the contents of my bags of recycled quarters.

Do you have a problem with the messenger or the message?

Mostly the messenger because the messenger has an ulterior motive and has crafted the message to serve that motive.

The purpose of this piece is not to inform the public but to further the interests of the messenger and this is done in a dishonest way. It will work only to the extent that the public is gullible--but the messenger is highly skilled and probably will fool many.

It is not presented as a dagger to the competition (which it is) but as a warning to the public (which it isn't due to hidden agenda).

The purpose of this piece is not to inform the public but to further the interests of the messenger

Do you have evidence to support that claim? Are you willing to take responsibility for that allegation by identifying yourself, as BUSINESS WEEK and its writers did when they published the article?

If Google can't tell that a computer is in Botswana and thus shouldn't see US-only ads, then the advertiser can't tell that the computer is Botswana.

My point is not "does Google have the data needed", my point is "is there any data to prove that Google has implemented the extra step required".

a) when IP address aaa.bbb.ccc.ddd needs ads, eliminate ads that exclude its country from the pool of possible display ads.

b) when a click comes via aaa.bbb.ccc.ddd (possibly much, much later in time from when the ad was displayed), confirm that the ad clicked on was actually ever displayed to that IP address. (Or, more weakly, check again that the IP address is on the list of approved countries for that advertiser.)

The simplest and most profitable algorithm, of course, is to assume that if you get a click, then it must have come from an approved country because of step (a). Much like Google used to make advertisers eat the cost of double clicks.

I can prove to a reasonable degree by inspection that Google does (a). I don't know how to easily prove that Google does (b). Most advertisers do not attempt to detect that AdWords traffic comes from approved countries and, in the case of click fraud, the fraud software may transmit the click to Google without ever fetching a page from the defrauded advertiser, further reducing the odds that a failure of Google to perform step (b) would be widely and immediately noticed (especially if tastefully distributed by a discerning thief).

Kenb

"The quality of that traffic is a subjective opinion. Personally I look at parked domain traffic as complete junk and it is one of the reasons I stopped using AdWords."

The quality of that traffic is not subjective at all. It can be measured. You may consider it complete junk, but that's an opinion, nothing more. Type-in domains deliver very targeted traffic, particularly in the niches I work in.

The purpose of this piece is not to inform the public but to further the interests of the messenger

Do you have evidence to support that claim?

It is hard to believe BW would attack online advertising (read $GOOG & $YHOO) with such a semi-ignorant article, without more than an informative purpose.

Do you have evidence to support that claim? Are you willing to take responsibility for that allegation by identifying yourself, as BUSINESS WEEK and its writers did when they published the article?

That's an aburd question, I am not a "respected" business periodical or even a journalist. I am making a supposition based on what appears to me to be blatantly obvious. I do have inside information but I sure won't parade it here for you. Your credentials and your evidence are no better than mine. What nerve you have, asking for my identity.

The quality of that traffic is not subjective at all. It can be measured. You may consider it complete junk, but that's an opinion, nothing more. Type-in domains deliver very targeted traffic, particularly in the niches I work in.

Even if what you say about the "quality" of your domains is true, this is not true for all parked domains. The parked domains that I had traced clicks to for my ads were not the result of type-ins they were junk and may have resulted from arbitrage.

Regardless of the quality issue, parked domains need to be treated separately from regular content sites. Advertisers deserve the ability to opt out of parked domains independently from content sites if they so desire.

I also believe that advertisers should be informed of the source of all clicks on their ads. This would go a long way towards reassuring advertisers about content sites and would give them the ability opt out of sites that they find disagreeable.

I am making a supposition based on what appears to me to be blatantly obvious. I do have inside information but I sure won't parade it here for you. Your credentials and your evidence are no better than mine. What nerve you have, asking for my identity.

1) If it was a supposition, why didn't you present it that way?

2) I'm not asking for your identity; I'm merely asking that you behave responsibly.

Trying to imagine how much that article will affect G in terms of lost adwords revenue - alot I guess.

My impression was that the article did seem contrived - the photo of the couple involved in the click fraud looked so staged that it was almost comical - like out of the onion or something.

In any case, I don't think the subject is exactly new to current advertisers. Even Time Magazine and Forbes have written on this exact subject ages ago, not to mention there was a whole class action suit over the matter, so there can't be too many advertisers for whom this comes as a shock.

Not that a feel-bad article in a major publication is going to increase confidence in the product, but there's nothing new in the article and next week BW will be inventing a different crisis, and this will mostly be forgotten.

I don't think Google has to worry that current advertisers will bail out of AdSense, because most current advertisers are aware of the risks and their ROI. Click fraud is like shoplifting in stores or junk mail that goes directly into the wastebasket: It may be annoying, but if the advertiser is making money, it's an annoyance that the advertiser may be willing to live with.

I think the bigger challenge for Google is in how to attract mainstream advertisers and media buyers who want to target audiences, not just keywords. Articles about click fraud in TIME, BUSINESS WEEK, the NEW YORK TIMES (today), etc. can't be helpful and may be harmful to Google's efforts in that area.

Everything else aside, I think the bottom business line is very simple.

I can run an ad, for example, in a print magazine and pay $X for the ad.

I can run some PPC ads and spend $X on the campaign. Some of what I spend on the PPC campaign goes to fraud clicks.

As long as I make more money with my PPC campaign than from my magazine ad campaign, I'm going to continue with the PPC campaign, regardless of the fraud. Or regardless of the "news" articles that focus on the fraud.

FarmBoy

1) If it was a supposition, why didn't you present it that way?

2) I'm not asking for your identity; I'm merely asking that you behave responsibly.

I presented it as my opinion, what's your complaint with that? I do not apologize for my "behavior," it is entirely responsible. You are holding me to a far higher standard than you set for yourself.

The motives of the BW editors and writers on the other hand are less than honorable, that's my opinion and I'm sticking to it.

Type-in domains deliver very targeted traffic, particularly in the niches I work in.

the purest targeted traffic one can achieve. or complete junk.

Regardless of the quality issue, parked domains need to be treated separately from regular content sites. Advertisers deserve the ability to opt out of parked domains independently from content sites if they so desire.

if "what kind of websites are in my advertising portfolio?" is relevant for the advertiser, absolutely. even if some parked domains deliver above average conversions. we need to build confidence in the content publisher network. so google should urgently do something in that direction.

in case of "what makes the most money?" then it's indeed only a question of participating and watching the bottom line.

You require absolutely no computer or business knowledge to detect click-fraud. Simply stop advertising for a week every couple of months and check your income. If your income is down, then continue to advertise, if it stays the same then stop paying for the advertising. It's that simple.

This is not detection of click fraud. It's adjustment of ad spend. From a formal standpoint, to detect click fraud, one would have to be able to discern from the clickstream the intent of its originators.

To believe that a 40 year old entrepreneur would spend $2 million without frequently checking if the adverts were working is not believable.

Actually, this is quite believable. There have been numerous stories of people who've dropped huge sums on PPC advertising. They claim they are not worried about click fraud; they say they don't have time to worry about it; they say they're given a budget and have discretion to spend it however they wish as long as they're getting good ROI (as compared to print or broadcast advertising). Such individuals could very well be click fraud victims. But they won't consider themselves victims until they start getting poorer ROI than those individuals who pay closer attention to their spend.

Sure there is click fraud out there but not to the extent that the article suggests.

Without being able to determine the intent of the clickers, this is not possible to know.

I think the bigger challenge for Google is in how to attract mainstream advertisers and media buyers who want to target audiences, not just keywords. Articles about click fraud in TIME, BUSINESS WEEK, the NEW YORK TIMES (today), etc. can't be helpful and may be harmful to Google's efforts in that area.

Since G wants to attract mainstream advertisers, and extend its reach into other media (e.g. TV, print, radio), it will have to come up with justification as to why advertisers should expect to get better ROI with them (click fraud or not).

b) when a click comes via aaa.bbb.ccc.ddd (possibly much, much later in time from when the ad was displayed), confirm that the ad clicked on was actually ever displayed to that IP address. (Or, more weakly, check again that the IP address is on the list of approved countries for that advertiser.)

The ad may have been displayed to the same machine with a different IP address (due to DHCP and other technologies that can cause an IP address to change).

Most advertisers do not attempt to detect that AdWords traffic comes from approved countries and, in the case of click fraud, the fraud software may transmit the click to Google without ever fetching a page from the defrauded advertiser, further reducing the odds that a failure of Google to perform step (b) would be widely and immediately noticed (especially if tastefully distributed by a discerning thief).

The click transmitted to Google and the request of the page from the advertiser (if done) are different steps involving separate HTTP connections. So it is possible for a fraudster to submit clicks to G which the advertiser is charged for, but no traffic ever appears on the advertiser's site. All G is responsible for is sending the redirect command back to the requesting HTTP client.

KenB - Your comment that "Even if what you say about the "quality" of your domains is true, this is not true for all parked domains" is equally true of websites. What matters is the quality of the traffic.

Moti - "the purest targeted traffic one can achieve. or complete junk." Or one of the zillion shades of grey inbetween.

KenB - Your comment that "Even if what you say about the "quality" of your domains is true, this is not true for all parked domains" is equally true of websites. What matters is the quality of the traffic.

Absolutely; hence (as an AdSense publisher) I advocate Google disclosing to AdWords advertisers the source of every click. Sure maybe in the short run I'd be hurt by some AdWords advertisers opting out of my site, but I think in the long run Google and myself would profit by more advertisers becoming more confident in AdWords advertising and with advertising on the content network.

Google should also allow AdWords advertisers and AdSense publishers exclude several times more sites from their respective accounts than is allowed currently.

Giving advertisers more control of their advertising would alleviate a great deal of the concern of click fraud and junk leads. It is the advertiser's money to spend they should have more control over how it is spent.

Again this is about long term profits not immediate revenues.

Blocked Keywords Filter

AdSense for Content

AdSense for Search

AdSense for Content filters

Enter keywords to filter from ads on your content pages.

.info
about
allthe
best
blog
cash
cheap
directory
discount
easy
find
free
gift
great
hot
less
link
money
quick
quote
result
ring
search
sites
top

Save changes