Click-Fraud said to be 50% of clicks

[sebastian]

i know i tend to ruffle some feathers...

but, the following article is very disturbing for me:

http://www.mediapost.com/dtls_dsp_news.cfm?newsId=271261

with both companies, (google and overture ...but not yahoo paid inclusion, ...yet), i am experiencing ridiculous click-thru numbers on events where i KNOW i am not only the least expensive, but carry the most inventory.

for example, i am getting hammered on certain keywords and not making the sales. i am the lowest priced in my market and carry more inventory than anyone. it makes no sense to experience this many clicks and not make a sale.

furthermore, on events that we are just "experimenting" with and have less competition, we will make 3 or 4 sales from a very small number clicks. ...many if these shows we don't even carry in-house, so they have a 40% markup associated.

these things just don't add up, and i believe these articles that scream click fraud is growing. it's the nature of the internet for unscrupulous business folks to try any tactic possible to degrade competition.

i also find it odd, that your companies (PPC search enigne companies) do not address this issue more often with your customers. it's an obvious problem, and one that is growing both in truth, fiction and media.

we are enhancing our tracking system to test IPs and the consistencies of certain ones; but how do we handle DHCP IPs coming out of large ISPs. i found one of these click programs and they will instantiate a session, click whatever you specify, then release the IP and ask for another one from the ISP. if it gets the same IP which by the nature of DHCP is quite common, it will release again and keep going until it gets a new IP.

how could you possibly protect against that?

Engine Reps: please advise as to what your companies are doing to protect us, beyond arbitrarily not charging us for all clicks...

thanks.

[MrMackin]

There is a firm that you can hire to document what is happening to you + video of the browser activity you may not see with a clean computer.

[sebastian]

interesting ...however, it seems counter-productive to pay additional money to another comapny that babysits the engines.

the engines should NOT do their own fraud protection. it should be a third party company that has demonstrated quality integrity and functional superiority to tackle this problem. ...an "auditable" company that WOULD relase fraud statistics.

how can the engines be allowed to NOT release fraud statistics?

that information is very important to advertisers and could help advertisers work around it.

this overblown ego approach to business is wrong.

if i sold seed products, and periodically competitiors were coming into my stores, or my distribution partner stores, and cutting open my seed bags, i would demand to know "how many seed bags were being cut"

it's basic logic.

[Mikkel deMib Svendsen]

The best defence from us - all the active advertisers - is to bid on ROI, not on CPC. If the click fraud goes up then our max CPC goes down. Money talks and this will make the engines prioritize it more.

[bakedjake]

Click Fraud is sort of like chargebacks for CC processors. The only one with anything to lose is the advertiser/merchant.

They'll do just enough so they can answer the WSJ in the next conference call.

But clickfraud, just like chargebacks for CC processors, certainly does generate revenue. So it's almost certainly in their interest to maintain some level of clickfraud.

Also, think about it. The intention of most people performing the clickfraud isn't just to cost competitors money...

[seobook]

Also, think about it. The intention of most people performing the clickfraud isn't just to cost competitors money...

mesothelioma, $100, click, try another, $100, click, try another, $100, click, etc. ... looking for a nice uptick in my porfollio next quarter

[fantomaster]

This issue is actually as old as PPC: I can remember it being discussed at
length way back in the times of GoTo (which, for the younger members on
this forum, was later renamed to Overture before Yahoo! finally gobbled it up).

we are enhancing our tracking system to test IPs and the
consistencies of certain ones; but how do we handle DHCP IPs coming out of
large ISPs. i found one of these click programs and they will instantiate a
session, click whatever you specify, then release the IP and ask for another
one from the ISP. if it gets the same IP which by the nature of DHCP is quite
common, it will release again and keep going until it gets a new IP.

how could you possibly protect against that?
You could to some degree but even a fairly weak protection might possibly
cost you more than it would help you save. E.g. if someone's using an ISP
like AOL who will issue a fresh IP for every new browser window opened, it's
fairly easy to track combining a list of IPs and their owners plus time stamps,
etc. so that's fairly trivial.
However, the real fun starts when people resort to daisy wheeling clickbots
across a wide range of unrelated physical servers and IP blocks possibly
spread over several countries and continents.
(There's a legal angle to this as well, btw: if the culprits pick their
jurisdictions with care, it might even be quite legal for them to work this
scheme. To this very day, many countries still don't maintain any serious
checks and balances in their legal systems to combat IT related crime
effectively.)

Still, if there's one thing you can learn from the past as well, it's that PPC
companies seem to react fairly wimpishly if you take a strong stance of
disputing their claims to unique, non-fraudulent traffic generation and the
costs accrued.

In other words: get your stats right, hire a team of good lawyers and go in
for the kill - five will give you twenty that they'll buckle up if only because
their fraud protection tech is so low grade, it simply won't stand up in court.
After all, this is a generic technical problem inherent to all PPC setups:
there just is no 100% secure protection against PPC fraud, so this is the
engines' softest spot by default. (Consider: if push comes to shove, you
might even consider sueing them in criminal court as accessories to the fraud
perpetrated.)

In tandem, none of them are too keen on the PR repercussions such a case
can have. It's happened before and the only surprising thing is that it's not
happening far more frequently. (If it did, you can be sure it would be heavily
publicized at some point.) Once advertiser start getting seriously worried
about this issue (as they well should), down goes the PPC companies'
revenue and ROI.

Mikkel's quite right: money talks - so why not engage in some truly
interesting conversation. :)

[shorebreak]

You can't get rid of click-fraud, but if you manage to ROI you can keep it from costing you more than the industry average.

[optimizemysite]

ROI is the most important metric. For that reason, I would consider doing a CPA(cost per acquistion) program through one of the affiliate marketing companies. That way your publishers bare the risk of click fraud, and you only pay commissions when you make sales.

[halfacat]

If you track the incoming IP's you will see when the same user is clicking through again and again. You then take this to G or O and get real loud. They should play nice and give you an adjusted bill.

At least G should since they "Do no Evil" :D :D

[sebastian]

I think i ruffled some engine feathers ...if i don't post anymore you know i've been 'eliminated'.

:)

seriously, though - this topic is important. fantomaster's post made me even more frightened as i thought of the possibilities within his scenarios.

point is - more and more technical people are getting involved in marketing. while marketing people can say, "Judge it all on ROI", the technical, analytical folks (like myself) balk at this. Just because i am making money doesn't mean it's a fair assesment. Just because the campaigns are CURRENTLY driving more sales than it's costing doesn't mean the problem will not grow. If anything, success with PPC will make the problem grow faster and to further lengths...

Here's the facts:

- click fraud is a huge problem

- ppc comapnies will not address the issues with clients (stating that sharing protection methods would allow frauders to enhance activity)

- due to the nature of the internet, it is possible to make 'click fraud' undetectible ...not saying all fraud is undetectable - BUT IT IS POSSIBLE and worse, not too difficult it seems.

- there are companies out there designed to do nothing more than click ads. they exist as you read this...

will i stop advertising? NO.

will i charge back every odd circumstance i see? NO.

Am i mad at the Engines or think there is a conspiracy? NO.

But i do believe that an approach is being taken by the engines to be quiet about it, thus hoping it will go away. ...but the higher-up techies know deep down that it won't. I wonder how often they actually *think* about it.

i bet a lot.

[halfacat]

But i do believe that an approach is being taken by the engines to be quiet about it, thus hoping it will go away. ...but the higher-up techies know deep down that it won't. I wonder how often they actually *think* about it.

i bet a lot.

naturally a companiy is going to stay a bit quiet about fraud happening on their system. companies stay quiet about shrinkage in retail. Credit Card companies stay quiet about chargebacks and fraud on their system.

your point is valid and taken, but you are only inflaming the issue, which is easily done and is being done very well. what we need to do is propose solutions or understanding.

i made the point earlier that you should be able to check the log files for repeating IP's coming to your system from the various PPC programs you are running. This is called evidence. if you take this to Google or Overture they will look at it (if they dont you could probably start putting together a nice lawsuit).

this will continue to be a problem based on the open nature of the internet. you can deal with it or not.

[fantomaster]

i made the point earlier that you should be able to check
the log files for repeating IP's coming to your system from the various PPC
programs you are running. This is called evidence. if you take this to Google
or Overture they will look at it (if they dont you could probably start putting
together a nice lawsuit).
AFAIK, PPC companies are indeed watching out already for same-IP clicks
within an inordinately small time span in order rule out click fraud. It's a fairly
low tech protection but at least it seems to have been implemented across
the board already.

So offhand I would expect this to be an advertiser's least worry. Rather, it's
the many ISPs assigning dynamic IPs to their users (including every new
browser window) that tend to escalate the issue, one major example being
AOL/CompuServe. I'm not at all sure that PPC setups are protected against
this sort of thing even though it would be fairly easy to achieve.

And on the receiving end, any evidence you can collect within that scenario
will be circumstantial at best. (You cannot actually prove that traffic
generated via different IPs is actually the same person or clickbot, only the
ISP could do that. You can, however, make it seem highly plausible.)
That's why it's generally better to take an assertive stance and create a
stink with the PPC engines before going for litigation: as long as they
cannot be sure that they'll win their case, they will generally negotiate a
settlement with you. Which is a lot safer for both sides than running the risk
of losing in court.

[webanimal]

Click Fraud is sort of like chargebacks for CC processors. The only one with anything to lose is the advertiser/merchant.

They'll do just enough so they can answer the WSJ in the next conference call.

But clickfraud, just like chargebacks for CC processors, certainly does generate revenue. So it's almost certainly in their interest to maintain some level of clickfraud.

Also, think about it. The intention of most people performing the clickfraud isn't just to cost competitors money...


When CC chargebacks and bogus CC transactions are the result of PPC traffic, they add another dimension to PPC fraud. There was a time when I could get a small ROI from retail sales using PPC, if I managed my campaign effectively. Now, I don't believe that is even possible. IME CC fraud has been the direct result of PPC.

Since we are in the Wild Wild West days of the World Wide Web, there is very little accountability or law enforcement. Unless you are a big company like Google or Amazon, you can't make your own rules and enforce them. So,
you are at the mercy of the whole system.

I am wondering how long it will be before the small companies and individuals finally decide "enough is enough" and pool resources to demand and exact
accountability. The only reason PPC search engines make money off our misfortune is because we allow it. The only reason there is identity theft and the CC companies make money off our misfortune is because we allow it. We need accountability from the gatekeepers. This big fat web is run by some big fat monsters, who thrive off our vulnerabilities. The PPC search engine empire is no exception.

[hardball]

SE have a stake in controlling click fraud but not eliminating it. Keep the numbers under the radar and it stays quiet.

The solution is quite simple:

Give advertisers right of refusal via redirect of the click into the trash bin in your CP. The concept of returns is not new, if you don't like the looks of a 10 cent or ten dollar click, you simply refuse it in real time via re-direct.

Server side solutions to click fraud puts the harmed party in control and I would venture to guess it would take a very short time for some highly effective solutions to develop, click fraudsters would be brought to their knees.

[rustybrick]

Do the large PPC companies do a good job of detecting click fraud? I mean, do you always have to bring it to the PPC's attention or will they bring it to the customer's attention?

[Imagjen]

We were the victims to over $5,000 worth of click-through fraud in one month with Google. After requesting a teleconference with their support department we also discovered that eliminating our advertisment for their affiliate sites was the best way to alleviate this occurence.

When we inquired what would happen to the culprit, a vendictive competitor we assume, we were given an extremely vague answer. I agree that an unbiased third party should be in charge of monitoring the fraud occurences so that Google does not overlook these actions. It is our opinion that if the culprit is a substantial advertiser no repercussions will await him and he'll be free to continue to plague our account in addition to others.

[hiero]

Back when Overture began we had problems with huge click spikes with one of our clients. To make a long story short it was never resolved and we ended up eating the excess charges. As said earlier with all of the dynamic IP's out there I'm sure if you intend on committing pay per click fraud it not very hard to do.

All that remains is achieving ROI with your campaign, and in some industries it becomes harder than others. On the positive side it's still cost effective to use PPC for small to medium size web sites advertising on the web than traditional advertising campaigns which can be more expensive and not provide as good of a return.

I look forward to a time when the PPC industry can provide more intelligent software that does a much better job of catching this fraud.

[AussieWebmaster]

You can't get rid of click-fraud, but if you manage to ROI you can keep it from costing you more than the industry average.
Lol... you can trim the fat if you monitor for it... but there are always the few who get away with it.

[sebastian]

FWIW, i hope everyone is noticing that none of the "engine reps" have responded to this post. it's an obvious attention-getter with the number of people reviewing it; yet, the engines see no reason to comment?

i really am not trying to bash 'em as it's important to me to work in tandem with them; however, this lack of ability to step up and share their opinions and what they think should be done is kinda lame. ...but very typical of corporate mentality.

i'd love to see one lone rogue who would jump in here and say, "we know click fraud is a problem, and here is what we intend to do:" ...blah blah blah - and proactively suggest ideas and strategies to help protect against it. i'm sure it would be the same old things we already know or do, but some attention to this matter is just expected.

they won't and they will hide behind the excuse that they are "giving away tactics" ...but in reality, time goes on ...click fraud grows ...and a couple of stanford dudes sit fat-n-happy in cali-forn-i-a.

...maybe i'm just jealous.

cheers.

sebastian "christian"

[Mikkel deMib Svendsen]

...maybe i'm just jealous.

Yes you are - and me too! :) However, that dosn't mean you are not right.

I believe the engines will step up and talk about what they inted to do once it becomes critical for business - for example if a major media - or several, started to focus on this. As long as it's mostly something discussed among geeks in a corner of the web I don't think much will change.

[Papadoc]

Oakey dokey... I get really tired of reporters who are out to do nothing more than make a headline and are willing to twist things around until the point they want to make seems evident. Don't buy into stuff just because some 4th rate reporter wrote it. Unless you work for CBS, you always have to learn to always question your sources and the sources of your sources before you go leaping to any conclusions and then be honest about any bias they might have.

On the basis of the author's article alone, here's why you know this guy is a reporter willing to make the news if he cannot find it:

Can Clicklab give a reasonable assessment to the industry as a whole or just on their clients? Obviously just the later. Just the same as you don't go to a doctor unless you are sick, it seems to me that Clicklab might have a much higher percentage of customers with problems come to them than would be evident in a standard cross-section of the population. That means that this reporter cannot use them to depict the industry as a whole. Did the guy ever mention that Clicklab works for companies that have problems? No. He lets the reader come to their own conclusions and we know that most readers don't think for themselves. He only lists them as a web analytics company, which covers quite a range.

Then he couches his sentences, "...fraudulent clicks can account for more than 50 percent of all advertising fees attributable to certain categories". That's like a 30% chance of a 50% chance of rain.

Note the word "CAN" and "CERTAIN". If a couple of categories of clients with known problems had a 51% fraud click rate on one day, his sentence would hold up. So how does this account for the article title "Pay-Per-Trick: Half Of All Ad Clicks Deemed Fraud" His title makes it sound and gets the reader to assume that half of ALL ad clicks ARE fraud.

It's not until the article is almost ended that the author quotes a Jupiter Research analyst a saying that it is probably only attributable to the pharma and health sectors and that he'd be surprised if even THAT category was anywhere near 50%. Unfortunately that statement didn't support the title though and so had to be buried at the bottom where the vast majority of readers will never read that far.

[fantomaster]

I don't think very many people on this forum have actually bought into those
skewed figures (which indeed are highly selective and merely based on
hearsay, hype and a highly cavalier manner of tweaking sources.) Personally,
I fully agree with your analysis regarding the article per se.

However, that doesn't mean that we're discussing a non-issue here. That
piece, flawed as it is, constituted merely an impulse to kick start a serious
discussion that's long been more than overdue.

Because after all, click fraud is quite real - and so, unfortunately, seems to
be the PPC companies' summary unwillingness to do more than hush it over or
cry wolf the minute someone demands more transparency regarding their
efforts to actually do something effective against it.

[Papadoc]

However, that doesn't mean that we're discussing a non-issue here. That piece, flawed as it is, constituted merely an impulse to kick start a serious discussion that's long been more than overdue.

Because after all, click fraud is quite real - and so, unfortunately, seems to
be the PPC companies' summary unwillingness to do more than hush it over or
cry wolf the minute someone demands more transparency regarding their
efforts to actually do something effective against it.
I would agree. It is far overdue. But I fail to see how giving credence to any chicken little argument will actually help the situation. People that have a legitimate cause don't like others coming around and messing with the facts because it only muddles the situation. This debate gets nowhere when such stuff is posted.

If the situation is real and I believe it is very real, then real numbers will support the argument. It's just a matter of working hard to fish them out.

This reporter proves he is just flat lazy by going to such sources and bogeying the headline instead of taking the time to work with real advertisers to come up with some real figures. What is needed are real facts. How you get the PPC's to the table isn't through a few posts on a few SEO boards or some over-inflated articles. It's through market demand, I would guess brought on by competition with a solid alternative. Right now, the only real potential competitor out there is Yahoo and why they aren't going after this piece of the pie is beyond me unless they are working on some heavy duty anti-fraud stuff so that people will see them as a real alternative when they do come out.

I work both sides of the PPC fence as an advertiser and as a publisher. From both perspectives, transparency and legitimacy would be good. Since I do not click my ads, my revenues drop when fraudsters take high payers off their daily budget and their ads don't show for me. And if fraud were reduced, ROI goes up and advertisers are willing to spend more per valid click. Obviously as an advertiser, I'd far rather eliminate fraud on my clicks too... again good for me.

[fantomaster]

and bogeying the headline instead of taking the time to work with real advertisers to come up with some real figures.
Ah well, without trying to excuse shoddy journalism, do let's remember
that headlines are usually not the reporters' task but their editors' ...

How you get the PPC's to the table isn't through a few posts on a few SEO boards or some over-inflated articles.
Not by a few posts and some basically clueless articles alone, that's for sure.
However, once the stink's been propagated widely enough, there's bound to
be some more serious and viable investigation sooner or later, the more people
chime in and present the issue as absolutely real albeit drawing on a different
factual basis.

As for Yahoo!, my impression is that their current PFI-cum-PPC model is
pretty much self defeating. I don't see many advertiser that are particularly
happy with it and sooner or later they'll have to make up their mind as you
can't really have it both ways: paying for inclusion and paying per click
without any guaranteed rankings whatsoever is nothing but a fool's
game and quite a bummer IMV.

[Edited typo.]

[sebastian]

i agree that much of it could be construed as "sensationalistic'; however it [click fraud] is happening.

while many want to quickly discredit the writer, the facts still remain:

1) click fraud is a problem.
2) companies have developed to do nothing more than 'click' ads.
3) engines are not forthcoming regarding the problem

i find it hard to believe that anyone with 'real finances' invested in PPC programs could or would possibly argue that it's a "writer's" misconception to blame and not the activity of the fraud itself.

if a company spends $2000/day in PPC and learns that possibly 50% of the clicks are fraud, it hurts. ...even 40, or 30 or even 20%. thinking about throwing away $1000 a day for nothing is painful ...even $100 a day. dpending on the actual daily lose, it's comparable to 1-4 employee salaries just flyin' out the window.

only on the internet could these wild west business practices fly. imagine the same model in discount retail, or grocery stores ...it's comparable to stealing.

all i really expect at this point is engine confirmation and dedication of/to the problem. perhaps even assigning a third party 'task force' to really get the bottom of the problem. if you study business communication, you know that companies that have proactively attacked a problem faired much better in public opinion than those that didn 't...

think about it.

christian "sebastian"

[AussieWebmaster]

That is placing way too much faith and trust in a business model... and just so you know.... people eat from the grocery store without paying... just because everyone doesn't do it does not mean it is not going on.

ya lost me there dude ...perhaps my analogy was pretty lame with grocery stores. (although i bet nacho liked it)

anyway - my point still stands and i am adamant about it. THERE IS A PROBLEM ...much larger than is being addressed.

there are companies offshore in which one can contract with - today - and have them clicking competitor ads before tomorrow.

that's spooky. ...and if it wasn't, Google would not have mentioned it in the prospectus when going public. it's a waiting game - and the only time it will be seriously addressed and attacked is whenm that $135/share begins to slide south.

[fantomaster]

that's spooky. ...and if it wasn't, Google would not have mentioned it in the prospectus when going public. it's a waiting game - and the only time it will be seriously addressed and attacked is whenm that $135/share begins to slide south.
South being your way, eh? :)

Of course, you're quite right - it's a ticking time bomb and the sooner it
detonates, the better in the long run for everybody concerned.

Question being: what can we do about it if the engines won't? Set up a task
force? Target a class action suit? The way these things are usually tackled,
it would typically require incorporating a pressure group (for obvious reasons
in the US albeit with international participation, though knowing European
courts with their strong historical bias towards both consumer and small
company protection, the EU might be a pretty nice and effective second
theater of operations as well).
This would obviously require some funding, people who'd cover the PR angle
feeding the media, etc. plus a dedicated web site or two, an independent
forum of its own, high profile sponsors endorsing the agenda, etc. Plus, of
course, some cool, easy-to-remember name and/or abbreviation.

And it's not just Google, either: Yahoo!, Overture (as a self-contained legal
entity, however long it may last), Findwhat/Espotting and dozens of others
are in that very same boat. Though I really wouldn't mind seeing the day that
Google gets grilled by the European Commission for monopolist malpractice
similar to Microsoft. :D

[esoos]

In terms of immediate solutions, I believe Adwords has a feature that allows you to restrict your ads to certain countries by filtering out IP addresses. I spoke to few different customer reps at Google, and they assured me that only users within your specified countries would ever be able to see (or click on) your ads.

There's more at this link:
https://adwords.google.com/support/bin/answer.py?answer=6277&topic=21

This could help to solve the problem of companies being hired to click on your ads. I imagine this would also deter clickbots somewhat, since their IP's should have to come from one of your selected countries. Of course, if the fraudsters are coming through AOL, then I guess they would all have US-based IP addresses.

BTW, it doesn't appear that Overture offers the same kind of feature.

Also, I noticed that there are a number of companies offering click fraud prevention software, such as:

http://www.whosclickingwho.com/
http://www.clicklab.com/

Has anyone had any experience using this kind of software? I agree that the problem ultimately needs to be solved by Google and Yahoo themselves, but in the meantime, how affective do you think this kind of software can be?

[Mikkel deMib Svendsen]

I spoke to few different customer reps at Google, and they assured me that only users within your specified countries would ever be able to see (or click on) your ads.

This is not true. Google does not use IP alone to detect a region but a number of parameters. Thats all fine. But what I found and documented is that if you set your region to one country it will also be served - and possible clicked on, by users in other regions if Google can not determine the location of the user!

The reason I found out is that I suddenly started seeing ads that I was managing for a large US client that was targeted to US only. By asking Google I found out that they had my IP detected as "no country, no region" even though coming from the major ISP here (with a +50% market share!).

So, I would naturally have assumed that no ads would be shown to me, because Google did not know where I was from but instead I found that they served me what looked like random ads - mostly US, and ads that I was managing and therefore KNEW was segmented to US only.

[sebastian]

Greetings. "Click-Fraud" is getting more attention. This is from WIRED:

WIRED Article on Click Fraud (http://www.wired.com/news/culture/0,1284,65324,00.html?tw=wn_tophead_4
)

This is something that i like to see happening. the more people write about it and expose it's potential, the more the engines are forced to take notice and perhaps comment as to what we can expect for future protection.

Cheers!

sebastian "christian"

[AussieWebmaster]

That Wired Article is so old... it seems that there are a couple of good press releases going around about this and they are getting publicity.

[Papadoc]

That Wired Article is so old...
October 13th, 2004 is old? AAAARRRGGHH! I need SPEEEEEED!!! Give me MEGAHERTZ, lots and lots of MEGAHERTS. I NEED megahertz, Aussie NEEEEEDS megahertz. Pleeeease gimme gimme gimme... I need more SPEED!!! Johnny 5 needs INPUT!

[AussieWebmaster]

October 13th, 2004 is old? AAAARRRGGHH! I need SPEEEEEED!!! Give me MEGAHERTZ, lots and lots of MEGAHERTS. I NEED megahertz, Aussie NEEEEEDS megahertz. Pleeeease gimme gimme gimme... I need more SPEED!!! Johnny 5 needs INPUT!
I think it is a republished story... have seen the exact article a couple of months ago or else a big chunk of it was previously published.

[AussieWebmaster]

Also we are a day ahead of the US!!!!!!!!!!! (actually I am in NYC... but I get stuff sent from home ... pity the track results could not be time delayed).

[The Generator]

Hey rustybrick,

No, it has been my experience that PPC companies will not call attention to click fraud. However, if you have your own good tracking software and find a large discrepency between your reporting and the PPC company's reporting, they will normally refund some money into your account. In other words, it's the squeeky wheel gets the oil syndrome.

Of course customer service is a factor in whether or not you will get refunded, and with time it becomes pretty obvious which engines are good at addressing requests... and which are not.

[Papadoc]

No, it has been my experience that PPC companies will not call attention to click fraud. However, if you have your own good tracking software and find a large discrepency between your reporting and the PPC company's reporting, they will normally refund some money into your account.
Quite true. Few PPC clients are going to raise a public fuss. There is a bit of a fear that doing so could cause some negative reaction from the PPC and cost them more in the long run. Besides, if they get their refund, they consider no harm done and it might just work in their favor if the competition is still paying for click-fraud.

From those who have gone this route, I've learned that there is very little problem getting a refund as long as you provide data. No PPC wants this one to go to court and generate the high profile that would result. They would have two options, cede the case and take the necessarily negative publicity and tidal wave of subsequent complaints, or in order to counter any such claim, the PPC would have to enter their documents into the public record. That would very likely release a bunch of proprietary information about how and what data they collect.

[projectphp]

The opportunity here is a mass market tracking and management tool with extensive click fraud detection tools built in, and marketted to many companies. There is no other way that the cost of Click fraud detection will ever justify the expense in programming and other costs, even for the massive players.

Getting a refund on click fraud is also an ROI nightmare IMHO. Imagine the time it takes to achieve this goal. If a person on $30 an hour spends 8 hours to get a refund, in reading reports, contacting teh SE rep etc, that means that the refund needs to be in excess of $240 dollars to make it worthwhile. At $0.50 a clkick, that is 480 click that need to be found to make this ROI positive, and this figure excludes the fixed costs of developing the detection program, and all the months in which reports are checked and there is nothing out of the ordinary.

IMHO, there are some costs a business just has to wear, and click fraud is probably one of these. As long as a campaign has an effective ROI, what does it really matter?

[Papadoc]

Getting a refund on click fraud is also an ROI nightmare IMHO. Imagine the time it takes to achieve this goal. If a person on $30 an hour spends 8 hours to get a refund, in reading reports, contacting teh SE rep etc, that means that the refund needs to be in excess of $240 dollars to make it worthwhile. I'd say that if you are paying anyone $60K per year, and it takes them 8 hours to read reports generated by click fraud detection software and contact the rep, you might oughta think about replacing the $60K per year guy AND the company supplying the supposed anti-fraud software.

At 480 clicks and given the estimate that's been used here of 50%, that's still no more than 1000 total clicks to analyze. How many ways do you want to look at 1000 clicks? The whole point of the software is to do much of the analysis for you.

[Mikkel deMib Svendsen]

IMHO, there are some costs a business just has to wear, and click fraud is probably one of these. As long as a campaign has an effective ROI, what does it really matter?

So, are you really saying you think it's ok for me to run my perfect little click fraud bot on your clients - just to give me that extra edge and to prove savings and better ROI to MY clients? :rolleyes:

Fraud is fraud - but I agree that it dosen't allways pay to investigate it and claim a refund. However, that dosen't chance the fact that the engines should be a lot more concerned about this (well, maybe they are but are just wise enough not to talk about it .... ). After all, if the "fraud industry" is allowed to evolve and grow too much this could hit the engines very hard.

With just a few major medias featuring (and maybe over-blowing) the issue thousands of business could chose to pull teir ads right away. The log-in is weak - and the engines know that. It only takes a few minutes to set up an AdWords account - but it takes even less to pause it and back out. Major media coverage of the fraud issue could potentially hurt very bad if a large percentage of (especially small and mid-sized) business dropped out over night.

I honestly hope that the engines have a back-up strategy for this event because I do think that the risk of this issue blowing up in the medias real soon is very real...

[projectphp]

I'd say that if you are paying anyone $60K per year, and it takes them 8 hours to read reports generated by click fraud detection software and contact the rep, you might oughta think about replacing the $60K per year guy AND the company supplying the supposed anti-fraud software...The whole point of the software is to do much of the analysis for you.
I dunno. I have never seen any software that made any process 100% automated.

It seems to me that if you checked click fraud once a week for half an hour, not unreasonable after logging in and havinga quick look, and then actually discovered something, put it in a report and fired it off to a search engine, that is a minimum of 5 hours work. That is still $150 you need in refunds to make it a break even proposition. Throw in a half hour call from the rep and its 5 and a half hours.

If an SEM had multiple clients, the figure may be shorter per account, but still, you gotta find something and get something back to make this worth the effort. Depending on how this is billed, it may never be cost effective.

People rarely factor time into ROI equations, and even if 50% of clicks are fraudulent, how many of those can be positively identified? And how many need to be, on a monthly basis, to make the whole project worth the effort?

So, are you really saying you think it's ok for me to run my perfect little click fraud bot on your clients - just to give me that extra edge and to prove savings and better ROI to MY clients?
I can tell you exactly what I am saying: spend resources where they have the best effect. Simple idea, but often lost in the flow.

This makes the question: what is the expected ROI on click fraud detection? Not is click fraud ok, not how to stop click fraud, but is it worth spending to detect it. If you do choose to walk that road, are there effective counter measures that are ROI positive?

This is not a question of absolutes. As an analogy, the question isn't is it OK that Microsoft engages in monopolistic practices, but what can a competitor do about these practises, and waht is the most effective? Sometimes right and wrong aren't the issue, but what practical steps can individuals take concerning said rights and wrongs.

After all, if the "fraud industry" is allowed to evolve and grow too much this could hit the engines very hard.
So what are the best ways to deal with this? Employ detailed click fraud analysis at a cost to a business for "the greater good"? Start lobbying for the engines to do this for businesses? Let it go, and focus on your own business and maximising ROI? I really don't know, but a lot of comments in forums on this issue don't seem to address any practical, useful and cost effective solution as the starting point, and just assume it is worth doing.

I do think, though, that for a large agency or SEM vendor, click fraud detection software rolled out accross many campaigns could be cost effective, and a mighty good USP, whether it actually works or not ;)

For the SE themselves, tightened click fraud detection will become an increasingly important issue as a marketing and PR issue. I have no doubt they have already started writing their propoganda on this for when the S**t does hit the fan.

[Stevesko]

I too was tapped by Google for 10K over 2 months. I discovered it through my web analytics where I saw a surge in traffic on several days on one particular keyword from JUST netscape.

While I did get a refund and Google claims that the sudden surge of traffic (that didn't convert at all) it is due to "high traffic keywords" showing up as a quick search on the home page of Netscape, it sounds like AOL's army of programers are hard at work making their numbers. I have since searated my content and search terms into different buys so I can better track and stop this keyword on my content group.

[projectphp]

Hold on, did this issue cost you $10K, or was that your total spend?

... it sounds like AOL's army of programers are hard at work making their numbers.
That is a bit unfair. So I had a look into it. You said Google claimed:
..it (the high traffgic no results) is due to "high traffic keywords" showing up as a quick search on the home page of Netscape
At this very minute, http://netscape.com/ lists the hot searches as:
Hot Searches:
1. Electoral Map
2. George Bush
3. John Kerry
4. Election Results
5. Oprah Winfrey
That sort of feature, a quick link to popular searches, is bound to attract attention. If the searches were commercial, they would get clicks. NNow, that isn't click fraud whatsoever.

In fact, I would personally say this is an example of extraordinary customer service, as Google paid you back for clicks you received that were legitimate, but just didn't convert. Man, how good is that? Count your lucky stars. I bet a newspaper wouldn't be so kind.

There are so many ways to minimise the effects of this sort of thing (starting with not showing ads outside of Google), and in an automated system, Google can't be held absolutely responsible if people don't have a full grasp of the system and its nuances.

IMHO, thumbs up to Google for a nice bit of good old fashioned customer service, and not sticking to the hardline, letter of the law and screwing a customer over.

[Stevesko]

10K over 2 months was the "fraud"... I spend close to 50K/month so I am a good customer... and yes, Google does have good customer service and this is an example of that, however for you to state the I don't understand the nuance of the system...

As I stated, the Netscape quick search was Google's guess, they offered no proof to back it up, nothing to say that for sure that on those days with significantly higher traffic on that keywork from Netscape that it WAS a quick search item... just a guess. Had they provided me with that sort of info to back up the conjecture I would be more beliving. But guess what, since the refund in the summer it has NOT happened again. They moved on to someone else... I was watching.

I may have been harsh on AOL, however I think Google is going be less critical of someone like AOL than on your local content provider. It just seemed like an excuse. And if the quick search is generating all these hits and no converts it is a money machine for AOL/Google. That got an extra $1,000/day out of me each time they did it.

As for breaking out the content and search to separate groups, I have done that and it allows me to "shut off" the keyword for the content group if I need to. I needed to break it out because the keyword is to important in search and normally converts at 5-10% on content. Google and Overture have no way of just shutting off one content provider, this would be more useful and would give media buyers greater control over this issue, find a content provider that is wacking you.. black list them for your keywords and keep everyone else.

[projectphp]

10K over 2 months was the "fraud"
Even in quotes, that is still the wrong word, and, IMHO, this is still the wrong thread. In many ways, it deserves its own thread.

..the Netscape quick search was Google's guess, they offered no proof to back it up, nothing to say that for sure that on those days with significantly higher traffic on that keywork from Netscape that it WAS a quick search item...
So, do you believe them? I personally would tend to. Which again, comes back to unbelieveable customer service. I know of very few other advertising models that give money back simply for low conversions.

nd if the quick search is generating all these hits and no converts it is a money machine for AOL/Google. That got an extra $1,000/day out of me each time they did it.
The reason SEO/M is so good and converts so well is because it is a conscious decision. People search for what they want. Someone seraching for "Buy plane ticket online" wants to do just that. 5 "Hot Searches" links change that paradigm. Now, people who are just bored or lazy will click on the link, and probably (being bored and lazy) the first result, a CPC AdWord link. Tire kickers never convert, and a aimless link follower is the ultimate tire kicker.

Now, while I get your concerns, this issue is far more productive to everyone as, instead of a complaint / "fraud" issue, we address it as a tip and something to look out for. And what a tip: Netscape's quick Searches, if they are words you bid on, can cause budgets to absolutely blow out, and fast. In your case, it was $1,000 per day. That is a massive ammount to potentially save, and a tip well worth knowing!!

Google and Overture have no way of just shutting off one content provider...
And that, I agree 100%, would be an excellent feature.

[AussieWebmaster]

Would love a few more details about how this occurred etc.
Could you detail it with links to examples etc?

[ClickFraud]

Hi,

I recently worked with two separate clients that had experienced click fraud in excess of 70% of click-through traffic, they were both advertising on primary ad networks.

The quality of content network partners is a major factor in both of the cases above, refunds were made due to bad affiliates milking the system (similar in methodology to the Google law suit announced recently) and the ad network failed to detect the fraud even though the volumes of referrals were simply outrageous.

I will document the aboves cases on http://www.clickfraud.com/ (free click fraud info site) if I can get permission to do so from the affected businesses, there may be search network NDA's that prevent me but I will look into it.

Adam

[AussieWebmaster]

Thanks Adam....