Re: False Clicks/Click Fraud

From: Brad Jensen <brad.jensen_at_laservault.com>
Date: Wed 16 Aug 2006 14:08:52 -0500

Adam Audette wrote:

> Google is a business after all, and their AdWords and AdSense
> programs represent very large incomes for the company. IAB
> and PricewaterhouseCoopers released data on Internet advertising
> recently: $3.9 billion in first quarter 2006. Google's share?
> A fat 58%.
>
> It's obvious but needs restating: with such massive
> financial incentives, Google faces a direct conflict of
> interest between eliminating click fraud and hurting their
> bottom line.

I do hope you are kidding.

Every business is based on trust, and the last thing that
would help Google's bottom line is click fraud. They aren't in
business for one week, or one quarter.

> So what's Google going to do about the turmoil - be it real,

> imagined, or somewhere in between? Detlev Johnson reported
> today in SearchReturn that Google prepared and delivered a
> report "from the stage at Search Engine Strategies last week

> [that] took panelists completely by surprise - and without
> time to respond." Google called the methodology of third
> party monitoring services "very, very flawed."
> http://directmag.com/news/google_clickfraud_surprise/
>
> Detlev doesn't have today's copy of SR up yet, so I'm going
> to quote a block from it here:
>
> ----------------
> "Google stopped short of saying there was intentional
> wrongdoing, but using a live platform without informing other
> panelists and sharing the report ahead of time did not allow

> the proper discourse about the problem as the room needed.
> Panelists could only offer a delayed response after the
> chance to read the paper. The truth is somewhere in between.

> What is true is that search engines take fair measures
> against click fraud, and the problem is far worse than they think.

I am still looking for evidence of click fraud 'being far
worse than they think'. That is not an objective measurement
that can be evaluated, confirmed or refuted. And while I
personally tend to believe in some psychic phenomena, I don't
think it is a good basis for a business evaluation.

> "For example, Google produces a statistic that 800 clicks
> reported as fraud converted at nearly the same rate as 24,000
> clicks that were not in dispute. A conversion does not
> automatically disprove fraud. Conversion tracking may be
> designed to calculate signups, and fake signup numbers can
> exceed what Google reports as fraud.
> The most pernicious type of click fraud from PTR (Paid To
> Read) gangs of clicksters is well suited for actually
> completing the signup conversion process with completely fake
> information.
>
> "The search giant either discovers bad clicks and doesn't
> charge the advertiser for suspicious clicks - even those that
> lead to conversions, or they wrongly assume a conversion is
> automatic proof of a good click, even when it certainly is
> not. The thing about conversion fraud is that it really is
> fraud. PTR users that take the time to complete a form to
> show fake conversions are those that do so specifically to
> fool the search engine. When calculated together, these fake

> clicks and conversions at times can total to high click fraud
> estimates that Google criticizes."

Oh please. So far all the 'evidence' of click fraud that I
have heard here is based on a) appeals to authority -
worthless in my mind as evidence, particularly when the
authority is faceless and quoted secondhand or third hand or
b) based on hypothetical and mysterious processes that don't
make rational sense

> (www.searchreturn.com)
> ----------------
>
> Is Google underestimating fraud? The SES incident makes me
> wonder if "don't be evil" includes "don't be equivocal" in
> this case. If so, that kind of closed authority is only going
> to widen the gulf between publishers and the ad giant.
>
> -Adam
>
>
> Adam Audette, Publisher
> LED Digest | the email discussion list
> http://www.led-digest.com/

There is no question that Google has a massive PR problem
here, and I think their approach of being forthright at this
conference is a good start on answering their critics, even if
those critics respond by blowing smoke (they didn't give ME
and advanced copy!)

Stop and think of this idea of armies of faceless gnomes
falsely clicking ads and even filling out conversion forms!

So if my average click is worth a dollar, I hire one thousand
people to click on one hundred different sites. Each of those
thousand people earns one hundred dollars for me and I give
them half. However, $50 isn't enough money for anybody in the
USA to do this, so I have to hire one thousand people in , say
China, to do it. But wait, Google would notice that all those
clicks on all those sites that have my AdSense code are coming
from the same thousand people, or are coming from China, and
they would filter them out, revoke my AdSense PID, and there
would be no click fraud. Meanwhile I have to sign up a
thousand people, pay a thousand people, and hope not one of
them turns me in to Google or the FBI.

It doesn't make economic or operational sense.

Of course, it would be far easier to just use AdSense in a
legitimate fashion to make money...

If 30% of Google's clicks are fraudulent, there must be
millions of individuals engaged in full time click fraud.
Since they can't be in foreign countries, and there have to be
a very large number of them to blend in with the rest of us,
and they are willing to work for basically no money, this must
be a religious cult or something, and one of them must live
with two doors of your home. I'll bet they come in your house
pretending to give you a religious tract, then run over and
click on AdSense ads on your computer, when you are in the
kitchen getting them iced tea. Be afraid. Be very afraid.

Not.

I'm sorry, friends. I don't mean to be sarcastic. I just find
it hard to believe that any of you believe in magic click
fraud. I am a web programmer. I understand web browsers and
TCP/IP and such fairly well. None of the scenarios I have
heard here make any sense at all to me. They are all easily
detectable, or economically or operationally unfeasible.

I've been disappointed with my AdWords results at times, also.
I don't think any of my competitors is so afraid of me, that
they would bet their whole business future or scuttling my pay
per click ads. They have too much else to do, and far too much
to lose.

IFR someone can come up with a pay-per-click fraud method that
could possibly be widespread enough to make a significant
percentage of Google's clicks fraudulent - say 10% - while
being undetectable or even difficult to detect by Google, I am
open to hearing about it.

I don't think Google is naive, stupid, or indifferent to click
fraud.

Brad Jensen



Received on Wed Aug 16 2006 - 14:08:52 CDT