Google Stumps Up $90 Million To Settle Click Fraud Action

I'm seeing it at Reuter's as well.

Doesn't this open the door for other lawyers to zero in on Google for some quick bucks? Not that they shouldn't have to pay up for their lack of effort in stopping click fraud. Either way, the stock will take a shit tomorrow.

I have a JD and after reading about this settlement I think it is pretty bogus (except for the lawyers and for Google, of course). I just wrote a quick blog post on it here: http://blogation.blogspot.com/2006/03/google-click-fraud-settlement-lives-up.html. Would love to hear people's comments.

if you would love to hear comments on your opinion then it would be refreshing to do more than a link drop in every post, eh?

especially when u make the linkdrop URL a broken link.

You don't have to read it there if my analysis is enough for you.

David says he has a JD (that's a basic law degree..2 years typical). He also says he's a search marketer working at a class action law firm, which seems particularly relevant here. He doesn't like the implications of this settlement for search marketers.

He notes that Google may have to pay up to $90 million, depending on how much click fraud gets verified to be click fraud. According to David, if I summarize correctly, Google reserves the right to decide what is and is not click fraud, as well as a right to keep that definition secret. I aree with him on that being a bad idea, but I am not sure I agree when David states that

I don't see it that way, but I don't have a JD and I don't work at a class action law firm.

David does correctly (in my non-legal opinion) tell us that class action lawsuits suck (my word) for consumers because they force everyone to either participate or opt-out, where particpation means accept the settlement as is, and opt-out means if you want to sue Google you will have to do so on your own. In a very real sense, some class action lawsuits are gifts to the corporations because, like in this case, they can lump up every possible click fraud claim for the past 4 years into one well-engineered settlement, in advance. David calls it a "win-win for Google".

David also reminds us that the legal fees will likely eat up a few million of the payout, and that there is no telling in advance how many claimants will participate in the suit. It's a win for the lawyers. I hope David gets a real good salary for his search marketing efforts for such a lucrative industry.

Excellent analysis John!

Sorry for the short post - I wrote it at work so I didn't have time for much composition. By the way, I *used* to work at a class action law firm, now I work at a lead gen company. No more law for me!

Here's the salient parts from my blog posting:

1. Google Isn't Obligated to Pay Anyone, Anything
Read the announcement carefully and you'll note that Google has agreed to pay "up to and no more than" $90 million to settle click fraud claims from 2002 to present. As noted on Google's blog: "For all eligible invalid clicks, we will offer credits which can be used to purchase new advertising with Google." So, this isn't a cash payment, it's limited to $90 million, and it only applies to "eligible invalid clicks." What does "eligible" mean? Read on.

2. The Definition of "Click Fraud" Remains a Black Box Defined by Google
As has always been the case - and as will continue after this settlement - the definition of click fraud remains a closely-guarded secret inside Google. Google - and only Google - determines whether your clicks were fraudulent. They are not obligated to provide an explanation of their reasoning (and generally do not) and their decision is final. This settlement doesn't change that. So anyone who previously submitted a click fraud claim and was rejected will likely get rejected again, meaning that virtually no one will actually receive a credit from this lawsuit.

3. Advertisers Basically Have No Choice But to Accept the Settlement
One of the downsides of class actions for consumers is that it enables companies to "freeze" future claims. In other words, if the judge approves the settlement, you have two choices: accept the terms of the settlement, or "opt out" of the agreement and pursue action against the defendant independently. Since the average consumer (or small business) can't afford to pursue action independently, you essentially have no choice but to accept the settlement. So few companies will actually get any money by opting-in to the settlement and even fewer will have the legal resources to opt-out. A win-win . . . for Google.

4. Class Action Suits Mostly Benefit Lawyers
So why would the plaintiff make such a bad settlement? Again, read the fine print: "We do not know how many will apply and receive credits, but under the agreement, the total amount of credits, plus attorneys fees, will not exceed $90 million." Ah, attorneys fees. Lawyers who file successful class action lawsuits usually get a huge payment as part of the negotiation. The payment includes the cost of their attorneys' time, as well as a lump sum of pure profit. My guess is that a $90 million lawsuit will probably return at least $1-2 million to the law firm that filed the suit.

So let's review: this suit doesn't obligate Google to do anything, leaves the determination of click fraud up to Google, is unlikely to result in any new settlements, puts the onus on the marketer to recover any money, precludes further lawsuits, and only financially benefits the lawyers filing the suit.

They've prolly already got that ($90m) in the bank already from the closing and forfieting of commissions due on accounts arbitrarily closed over the last year or so... with little to no supporting evidence.

Nice summary davidzhawk....

This is very bad PR though for Google. I saw the story on BBC World business news over my cornflakes. The presenter spent the best part of the section explaining what click fraud was and how Google was susceptible. Large diagram of the SERPs explaining that unscrupulous people can simple “click the adverts themselves or design robots to their work”.

Irrespective of the specifics - that's a lot of promotion of the issue. I would agree with Cornwall - definite pandora's box. Far easier to keep investors happy when you didn't admit the problem.

Current results for click fraud

This is a classic Laywer doing what lawyers do best...

My guess is that a $90 million lawsuit will probably return at least $1-2 million to the law firm that filed the suit.

More to the tune of 30 million will go to the legal team that maybe put 100k into the project while the 60 million is suppost to cover 4 years of fraud/low quality refunds for all the US consumers.

At every milestone Google lost in court, they were being backed into the corner. This could've cracked the egg wide open but...

The entire reason for this quick settlement was the judge pretty much told Google you are going to have to let some click fraud experts look at your data and audit it to verify the 1% click fraud number. Google didn't want that so they threw the money at the legal team and while they laugh all the way to the bank the consumer gets the way short end of the stick.

It gets better. Things are still in the works for some of the data to still get reviewed but not by click fraud specialists. Right now the audit may go to Neilson Ratings who has 1. no clue what to look for when it comes to click fraud / low quality clicks 2. doesn't know all the avenues for abuse 3. has no clue that this system is best milked over weeks and months across 100's of sites. 4. doesn't have access to current fraud data...

You see the fun thing is to have access to data like I do with Vericlix.com. Vericlix was meant to pick up OBVIOUS infractions like 1 ip clicking the add 30 times a day or a noob that allowed himself to get cookied even though hes switching his ip after every click via a proxy server or a parked domain from some no name registrar that chews up someones entire budget etc. The sad thing is that the obvious fraud is not only detected its detected at such low levels of logical threshholds it makes you cringe. To see these low quality clicks still pass the buck and to see them draw money from an advertiser's account is really sad.

All this leads us to believe if grossly obvious can make it past the doctorates over at Gplex then this matter is something that gets swept under the rug, and if they can keep their data away from prying eyes then we have to just to take their word that the 1% fraud/low quality clicks is the truth. Sorry G if all the world was a Brett Tabke forum maybe we'd believe you.

When it comes down to it, a bad system doesn't get fixed, the lawyers get richer the consumer gets screwed...

If you take the refund today and then login to adwords tomorrow and see that some assclown blew your spend in less than 15 minutes you're not eligable for a refund.

Whats the solution? To take away the pay per peformance model and make it a blanket advertising model. Sure you can still allow users to bid but let them bid per hour or per day vs bidding on the per click basis. When you blanket an hours worth of time you don't need half-ass algorithms to check for low quality clicks.

The choice not to use Google as the opportunity for undetected click fraud is just too easy. Give me a list of 10,000 proxy servers and I'll show you just how easy it is to unload your ad buget with nothing to show for it and no way to trace it except for the unusually high volume of clicks with no ROI.

If that Google's fault?

Hell no, it's the defrauders fault, and the fact that the internet if full of more holes than swiss cheese.

The problem is that Google naively doesn't (or didn't) disclose that such fraud is possible when you sign up for AdWords and force you to use it "AT YOUR OWN RISK". Such statements would obviously make people think twice about just how much money they were willing to risk in the first place.

Plus, Google could very easily implement a tracking system that only counts click-thrus that actually make it to the advertisers site and stay more than 5-10 seconds, or something similar which eliminates simple click attacks as visitors would have to actually land on the other page before it completes the circuit of being classified as a click.

They could also easily track and discredit the same click to same site from the same computer within an hour or half-hour, which I know they don't as found that out testing my own ad campaign once.

Doesn't matter, coulda, shoulda, woulda, they seem content to pony up a settlement without addressing the issues whatsoever so the statue quo remains until the next settlement, and the next, and the loss of faith by advertisers, etc.

The situation as I see it is they have created this huge cash cow and are willing to milk it to death and either ignore the inherent risks to maximize profits or they do have the next generation sitting in the wings that fixes some of these problems and are merely biding their time to release.

Who knows, only Google, and they aren't talking.

>Sure you can still allow users to bid but let them bid per hour or per day vs bidding on the per click basis.

The old SearchKing search engine had a pay-per-day bidding model for precisely that reason. This was 5 years ago and far ahead of it's time, IMO.

One of the directory scripts I use has a pay per day sponsorship model but it is flat fee not bidded.

Exactseek's ISEDN advertising network works on time period rather than per-click so there is no click fraud there either.

Frankly these are all good ideas that nobody implements because they are all to busy copying YOverture! and Google to think for themselves.

Actually it is their fault. It is their system, their responsibility. If they want to avoid being at fault, they need to display actual click fraud statistics from independent studies so that users can make informed decisions. Not some bullshit fuzzy math from the Plex.

When you claim click fraud is 1% and users are getting 10% consistently, something is wrong. There is a fine line between ignorance and straight out fraud.

After reading incrediBILL's "Forget Click Fraud, Fix Technical Money Loss", I am curious about something in relation to: "..the conversion tracking bug could have an OnUnload event component that reports back the total time spent viewing a page so that accidental clicks resulting in less than perhaps a 3 second threshold of visitor time spent on a page could be detected and deducted as well."

Forgive my technical ignorance here, but can anyone tell me why doing a CPA ( http://www.marketingterms.com/dictionary/cost_per_action/ ) model instead of a PPC model would not work to at least greatly reduce click fraud?

In ClickZ's "Click Fraud Settlement Viewed as Win-Win by Some" it said: "The problem, according to Jessie Stricchiola, president of SEO firm Alchemist Media and a subject-matter expert for the plaintiff on the Lane's case, is that search engines are making a decision about what constitutes a valid click without having some of the more crucial information required to make that decision, including site visitor behavior, clickstream, and conversion data." So, would the "conversion tracking bug" give Google the above "crucial information" so that CPA could work?