We are running multiple campaigns for multiple product lines. Our
average CPC is ~ 1.00 for each keyword. This leads to anywhere between $1000 to $1500 that we spend on adwords every month.
Our logs show multiple enteries / short visit / exit for the same ip
addresses on a daily basis. We have logged all this and sent it to
google. Their response was that we 'were not charged for these clicks and google takes the issue of clickfraud very seriously blah blah blah'
I asked them then whether there is somewhere that I can see for which clicks I was charged? This should be really easy and I can compare the logs against our own logs. However, their response was an adament 'No' because the adwords is a propietry technology!
I am not sure what that had to do with my question. If I am able to see for what I was charged, wouldn't it help both my company and google to combat click fraud better?
I already have my logs and can see the ip address of various visitors and where the click was originated from. Wouldn't the ip of the clicks that google charges me for be a subset of my logs? So what is propietry about that?!?
Why would google be afraid of supplying me with that info in the first place? How can I ensure that I am actually being charged for the correct number of clicks let alone valid or fradulant if the company wont supply me with details? Are we just to trust google and payup whatever they charge even though the charge may be erronous?
Sometime ago there was a story about a phone company mistakenly sending a subscriber a bill for an enormous sum of money. The customer was able to fight the charges successfully because they had access to detail information of the charges. I am sure that the bill was created by a computer. So it wasn't human error either.
The question is should we put blind faith on google and trust them to charge us accurately?
I am interested in your thoughts...
what is multiple? twice the same ip? or more like 100 times the same?and what if a person searches for your site everytime they want to visit your site, and clicks the ad? is that googles fault? they don't charge you for fraudulent clicks, and imo above case isn't fraudulent.
Google doesn't charge for "invalid" clicks. They may charge for fraudulent clicks, if their click quality team doesn't catch them.
The descriptions you gave could be legit or fraudulent, depending on the intent of the clicker.
Why would google be afraid of supplying me with that info in the first place?
That someone may not be you, but Google has no way of knowing that. And even if you are not that someone, if they give you the information, then they have to give the information to everyone else who asks for it. And one of those people may be a someone who figures out a bypass.
No, you don't have to blindly trust Google. You are perfectly free to stop advertising with them.
To say that I "don't have to blindly trust Google. You are perfectly free to stop advertising with them" is similar to saying you do not have to buy Microsfof products or you do not have to buy an iPod. When something becomes a monopoly or something very near a monopoly, you most definitely need to use them to some extent.
I somewhat agree with your point abou some malicious user using that info to get around the click fraud algorithm google uses. However, that is the case with any information. Someone may use information maliciously.
The google's mission is to index the information on the net and provide that info freely to the masses. I do not see google not indexing a site say about ways of committing suicide because someone may use it to commit suicide.
The click information in fact is a subset of my logs so in a way I own it as well. If I ask a company what I am being charged for, they can certainly come back and say they wont tell me and as you said, I can discontinue using their services. But if you are an online company, I am not sure you can survive without using google (a near monopoly on search)
We have a much better ROI with Yahoo's PI than AdWords. I have never owned an iPod and don't feel any pressing need to. Other than an OS on a new computer, I haven't bought a Microsoft product in years. (And I don't mean that I've been pirating software- I haven't bought, traded, hacked, or otherwise acquired any Microsoft product for probably 5 years, except for the afore-mentioned OS and the free version of SQL Server 2005.) Although I bought Office 2000 years ago, I have no plans to upgrade. More likely I'll look into some of the free Office alternatives. I haven't looked into them before now because I already have Office 2000, which is basically free to use at this point.
The google's mission is to index the information on the net and provide that info freely to the masses.
Also, as I mentioned before, if they have to provide you with the logs, then they have to do that for everyone else who asks. With their thousands of advertisers, that creates a huge potential overhead of work that has to be performed and staffed. Staff members will spend all their time processing log requests and having to answer questions from Joe Advertiser about one $.05 click on May 16 last year and then listen to Joe's tirade about why he doesn't think he should have to pay for that $.05 click. Not a good business decision.
[edited by: LifeinAsia at 5:05 pm (utc) on Dec. 13, 2006]
ClickTracks and some other web stats packages combine log file data with the actual data provided by Google AdWords and Yahoo! Search Marketing, so it might be helpful to use software to compare your logs with the actual reports for keywords you're being charged for.
How much do you make out of the clicks minus how much you pay. That's it.
There will always be fraud and 'strange' clicks and also clicks that do not convert at all. Try to minimize that and come up with the best ROI measuring system possible, especially if you pay $1 per click.
For instance: take how many UNIQUE weekly visitors come to you from adwords, how they convert and see if it is OK to bid $1 for them.
Fraud and repeated clicks are just part of the nature of Internet adverstising. It is probably at least 10% of what you spend.
Now, if you see VERY strange non-converting numbers, one of the first things to look into is the Adwords content network.
One other thing. A conversion should be more that just click on a link. That is far too easy to reproduce for fraudsters.
Also, as I mentioned before, if they have to provide you with the logs, then they have to do that for everyone else who asks. With their thousands of advertisers, that creates a huge potential overhead of work that has to be performed and staffed. Staff members will spend all their time processing log requests and having to answer questions from Joe Advertiser about one $.05 click on May 16 last year and then listen to Joe's tirade about why he doesn't think he should have to pay for that $.05 click. Not a good business decision.
Arguably, banks, phone companies, credit card companies, and many other institutions that handle financial transactions are able to provide both detailed transaction logs and customer support for a far higher number of customers. They also have to deal with fraud.
Arguably, banks, phone companies, credit card companies, and many other institutions that handle financial transactions are able to provide both detailed transaction logs and customer support for a far higher number of customers. They also have to deal with fraud.
All those companies have large customer service departments that cost them a lot of money. (Google's made the decision not to spend money that way. Google's a lean tech company, not a bloated customer service company.) And all have very clear-cut, non-proprietary definitions of fraud and how to identify it.
With a few exceptions, all of them are reactive and YOU have to notify the credit card company or bank or phone company of suspected fraud. Yet Google is proactive and notifies you about fraud.
- Clicks from fraudsters (10%?)
- Clicks from (network content) sites that do not convert (50%)
I am more concerned about the second.
With a few exceptions, all of them are reactive and YOU have to notify the credit card company or bank or phone company of suspected fraud. Yet Google is proactive and notifies you about fraud.
Well, they notify you about the "invalid" clicks and the fraud that they know about.
What's the difference between:
- Clicks from fraudsters (10%?)
- Clicks from (network content) sites that do not convert (50%)I am more concerned about the second.
The latter may or may not be fraudulent. It isn't always possible to tell. From G's perspective, the former could be determined if they actually discovered or learned of some fraudulent activity, such as a botnet that targeted certain sites.
If none of the data coming over in the HTTP exchanges was altered, one could determine from the referring URL where clicks came from, and take further action if necessary, such as excluding the content network.
I see it in this way: you pay X to get Y visitors, out of which you make Z. You decide what X is.
If G allows too much fraud, or non-converting clicks, X will go down. It is really a G problem.
The goal for advertisers is to be able to measure Z properly. Most advertisers are still not able to do that.
G is giving many tools to improve Z. You can exclude sites from the content network. You can exclude countries. You can decide when to show your ads. The more the better! (Please let us exclude Search sites!) :=
I see some sites that may not be fraud, technically. Still, I have days like yesterday.
Nothing worse to see these spikes in traffic, daily budget is used up, missed a lot of quality spend, check the logs and find out the spike came from parked domains.
Maybe I am lucky, my ads are SOOO good that any fat fingers parked domain visitor sees my ad and is sooo compelled to click on the ad that they do...and there are sooo many fat finger parked domain visitors...that having nothing better to do on xmas eve...that must be what happened.
Anyone asked google about the parked domain traffic and tried to get a refund? I don't believe that it is classified as fraud.
