Yahoo approach to click fraud

[Martinscholes]

Martin,

an intelligent approach. I can see your point. In this particular case the suspect searches all came from a Yahoo partner search engine. I have a very good log of six hours of activity. 31 searches for the term were made via 18 different IP addresses. The term is so obscure that an exact match analysis on Google Adwords shows only 2 searches for the term over 5 days.

I have sent the data to Yahoo Search Marketing in an Excel file. I suspect they haven't even opened the file.

Ah! Now with that information, it is clearly fraud. But by whom? It might even be by the partner search engine.

I wonder if this is like a fraud my wife and I spotted in a branch of a discount supermarket?

It turned out that the staff were running a scam. They would key the last item of the preceding shopper as the first item on the following shopper in the queue.

As nothing in the shop was much over a £1.00 or so, nobody would notice. They were not using itemised till receipts, so nobody would have noticed. And if there's an extra 50p on the bill, how would you know?

How they were found out was that my wife was doing the shopping of an elderly friend, I was doing our shopping. The last item on her shopping was a bottle of brandy, about £8.00.

Now when an item for £8.00 suddenly appeared on my bill, we went back to the store to complain. The manager went white, and handed the money over without demurring. Odd in itself.

It turned out that they were able to keep a running tab with their till system, and all of the money that was extra was kept and shared out by some of the staff.

It's possible that someone at the partner search engine runs a similar scam with every -say- 10th search. But they hit a problem in that this time they chose to use a very obscure term. With someone on the case as sharp eyed as you.

The only reason I can think that Yahoo would not seem to properly investigate such a problem is if they asked the partner search engine who would lie and tell Yahoo everything was fine.

Either that or Yahoo do not care once they have your client's money.

Of course, the above might be a load of crap, but someone defrauded your client and the people who should have investigated could not be arsed to do it. It does make you suspicious, doesn't it?

[cw1865]

31 searches for the term were made via 18 different IP addresses. The term is so obscure that an exact match analysis on Google Adwords shows only 2 searches for the term over 5 days.

It looks to me like the click frauders are teaming up communicating and collectively defrauding.

[dburdon]

Martin, CW,

taking an objective look at this could reveal that both Yahoo and the partner engine both enjoy a short term benefit from Yahoo's less than vigilant approach to spurious clicks. Long term Yahoo loses as its brand is associated with theft.

Kjell,

I agree with your point. I've withdrawn most of my client sites from the Yahoo programme. Mainly due to the inferiority of their conversion costs. Sadly, I'm unhappy to be so dependent on Google.

[cw1865]

You would even think that Yahoo would be suffering in the short term as well only because advertisers should be intuitively adding the cost of click fraud into their calculations.

During Yahoo's click fraud class action, I was actually an objector to the proposed settlement and the basis of my objection was that any settlement must include an option on the part of the advertiser to preclude advertising on Yahoo's partner sites. The settlement was approved over my objection of course (I'm in NJ and I wasn't about to fly out to the Federal court in California)

[Martinscholes]

It looks to me like the click frauders are teaming up communicating and collectively defrauding.

Could they have spoofed the IP records at the partner search engine? To make it look as if they were getting the hits from 18 different IPs?

[dburdon]

Could they have spoofed the IP records at the partner search engine? To make it look as if they were getting the hits from 18 different IPs?

Martin,

I guess I could investigate further. But I'm fairly convinced on the basis of keyword data only that the partner search engines of Yahoo are being allowed to indulge in click fraud. Yahoo themselves benefit from this activity. I have a good keyword log, right back to the page and search term used to access my client's site. The search term used rarely ever appears in the much larger Google Adwords sample despite a much larger budget. Yet the little Yahoo partner engine keeps generating traffic for this term.

[dburdon]

You would even think that Yahoo would be suffering in the short term as well only because advertisers should be intuitively adding the cost of click fraud into their calculations.

During Yahoo's click fraud class action, I was actually an objector to the proposed settlement and the basis of my objection was that any settlement must include an option on the part of the advertiser to preclude advertising on Yahoo's partner sites. The settlement was approved over my objection of course (I'm in NJ and I wasn't about to fly out to the Federal court in California)

CW,

if only Yahoo would allow advertisers to exclude all the rubbish. Firstly Google allows you to excluded partner engines. Secondly, the partner engines on Google - AOL, Ask etc - are of a much better quality than their Yahoo equivalents.

[cw1865]

Yahoo Lawsuit Attorney | Lawyer Checkmate Settles with Yahoo!

This was the 'Checkmate' class-action settlement. With respect to Click Fraud yahoo agreed to do the below (roman numerals) - My objection, which went nowhere, the class-action attorneys in California, I am sure, wanted their fees, was to do it the way google does, permit the user to block the partners.


i) Yahoo! shall launch an online traffic quality center, which will be available to its advertisers within 90 calendar days of the Effective Date of the settlement. The traffic quality center will include a resource center which will contain FAQs, best practices documents, traffic quality articles, enforcement guidelines, and an advice column.
ii) Yahoo! shall designate a Yahoo! employee as a traffic quality advocate who will be part of a traffic quality group to fulfill the function of fielding advertisers’ concerns regarding traffic quality, including its click fraud prevention efforts, within 90 calendar days of the Effective Date of the settlement.

[dburdon]

CW,

the ability to block specific engine partners would be better.

Anyone with a good analytics programme that measures sources of conversions can see that the partner websites deliver at dubious levels.

In the particular case I've brought to attention of Yahoo I find it amazing that they refuse to analyse the log I've sent them. The client in question was one of my last that spent any money with Yahoo - £2,000 ($4,000) per month. Now its zero.

[cw1865]

Anyone with a good analytics programme that measures sources of conversions can see that the partner websites deliver at dubious levels.

That is not surprising, I am curious, what differential in conversion rates are you experiencing between a PPC from the yahoo site vis-a-vis their partner network?