Google Showing Click Fraud Numbers In Adwords

[AussieWebmaster]

This was an interesting find - thank you Adotas.

Seems that Google has added this a little while ago without announcing it - or at least in anything I had seen.

We are seeing 10% plus Click Fraud numbers - but these are the ones Google has already zapped out of the system.

So with the 14% number floating around lately plus this 10% - the 25% number may have been very real over the last 4 years.

You can access it through Reports and Campaign Reports options - it is well hidden.

[AdWordsRep]

Glad this is getting notice - thanks for posting about it AussieWebmaster.

Seems that Google has added this a little while ago without announcing it - or at least in anything I had seen.

The Inside AdWords blog is the best place to look for this sort of thing, as many new launches are mentioned there. It's worth checking out every day or two, IMO - but then I am not exactly impartial, as I have several friends on that team.

Inside AdWords blog: http://adwords.blogspot.com/

This post in particuar: http://adwords.blogspot.com/2006/07/...id-clicks.html

Not sure if it is OK to quote the post on this subject (which was posted last evening) so I won't quote the whole thing. However, here is an excerpt, minus a screenshot, a bunch of text, and links:

...Today, we’re announcing the launch of a new AdWords feature enabling advertisers to have a much more detailed picture of invalid click activity in their account. The metrics of “invalid clicks” and “invalid clicks rate” will show virtually all the invalid clicks affecting an account.

These clicks are filtered in real-time by our systems before advertisers are charged for them. The resulting data will of course differ from one advertiser to the next. In addition, a much smaller number of invalid clicks may also be credited to advertisers’ accounts after-the-fact, as the result of a publisher being terminated from the AdSense program for invalid click activity. These will appear as account-level credits.

One of the most controversial issues related to the topic of click fraud has been estimating how big the problem is. Estimates from third-parties (usually from consultants who have a financial incentive to make the problem seem very large) have been both inconsistent and greatly exaggerated due to their methodologies. Advertisers have always been able to compare their log data with their AdWords charges to calculate an estimate of the number of invalid clicks in their own account. This new tool will make estimating invalid click activity much easier...

I hope everyone will read the blog from time to time. Tell 'em AWR sent you.

AWR

[Discovery]

AWR, I am so very glad to see this direction being taken by Google.

I hope that one day Google will utilize a very valuable asset, the advertiser, to help them monitor and control click fraud at the account level.

Millions of advertisers disabling toolbar partners, junk click IP's and Fraudulent clicks from domains will give Google the data necessary to pin point exactly where the "bad" clicks are coming from and kick them to the curb. I know this presents a large technical challenge and financial risk, but the idea is a powerful one.

Keep going Google, we know there is an industry wide problem here, we dont want to point fingures, we would much rather have you work with us to solve it. To me this little addition to adwords is a small signal to say you are doing just that.

Discovery

[Jeff Martin]

Its a step in the right direction. With the settlement being close to wrapped and with Yahoo setting up the traffic center this was something Google needed to do to stay competitive when they announced it a few days ago.

Keep in mind the number reported is supposed to be what Google has caught, the actual amount of click fruad is still undetermined. As Dr. Tuzhilin states in his report "“…when their methods work reasonably well, the invalid click detection problem is ‘under control." Of course Dr. Tuzhilin points out that this is derived from indirect data as their is no hard data to support his summary. A big question is when does it work and how often?

Thanks for the helful tip.

[Alexis]

This is great news! And I've just run a report showing invalid clicks. My question is, when are these clicks accounted for? Are they removed from the data in the Campaign Management screen or simply reflected in this new reporting?

Also, other than manually, is there a way to determine how much money was spent on invalid clicks so that reporting can be adjusted? This could potentially affect CPL, CPA and so on....

[AussieWebmaster]

Originally Posted by Alexis

This is great news! And I've just run a report showing invalid clicks. My question is, when are these clicks accounted for? Are they removed from the data in the Campaign Management screen or simply reflected in this new reporting?

Also, other than manually, is there a way to determine how much money was spent on invalid clicks so that reporting can be adjusted? This could potentially affect CPL, CPA and so on....

I think that is a part of why they are giving you the numbers.... remember these are ones they caught and did not charge you for... but you did get the clicks.

[Alexis]

So, the invalid clicks in the report exist independently from your total clicks and the cost associated with the total, correct?

[AussieWebmaster]

Yes they are the precharge clicks Google did not think you should be chraged for....

[Jeff Martin]

Originally Posted by AdWordsRep

Advertisers have always been able to compare their log data with their AdWords charges to calculate an estimate of the number of invalid clicks in their own account.

John Slade from Yahoo tried to say this also on the last "PPC Auditing and Click Fruad Issues" panel I sat in on. This isnt realistic.

1. How many advertisers have time to pour through their log data line by line? Many advertisers I deal with dont even know of the existence of these files and most certainly can't make heads or tails of them.

2. Its difficult for a log file analyzer to be able to pull out what is AdWords traffic and what is organic traffic.

3. Analytic software usually give different numbers when using the same log data. Different analytic software packages have different metholdoligies for reporting. For example what constitues a unique visitor? A user who hasnt visited but once that month? Once a Day? 20 min ago? Doesnt have a cookie? etc...

4. Log data can also be incomplete due to network, server, client or Google errors.

The new column is a nice gesture, although somewhat hidden away, however it's up to each advertiser to police their ppc spending for fraud wether you use a 3rd party service or build one in-house.

[The Generator]

I'm sure you already know this tidbit about Google and false ad clicks, but here is the NY times article anyway (you have to log into NYT or register for free).

http://www.nytimes.com/2006/07/27/te...ref=technology

[AussieWebmaster]

The Generator..... have not seen you on the board in a long time... welcome back.

[The Generator]

Thanks AussieWebmaster, it's good to be back ; )

[sheseltine]

The Internet's leading search engines are teaming up with an advertising trade group to find a better way to identify and measure "click fraud," a scam that has raised doubts about the Web's trustworthiness as a marketing vehicle.

The initiative, announced Wednesday by the Interactive Advertising Bureau, will draw upon the expertise of Google Inc., Yahoo Inc. and Microsoft Corp., the owners of the top online search engines, to attack a problem threatening to erode their profits. Combined, the three companies control 86 percent of the lucrative U.S. search engine market, according to comScore Media Metrix.

Two smaller search engines, InterActiveCorp.'s Ask.com and LookSmart Ltd., also have joined the alliance along with the Media Rating Council, a nonprofit group formed 42 years ago at the urging of Congress to help track and validate the sizes of advertising audiences.

Full article here

[AussieWebmaster]

Yeah I saw that earlier today but have not had time to fully readit... good addition

[AdWordsRep]

Thanks for spotting that sheseltine.

Since I do love a good chance to mention the Inside AdWords blog as a excellent source of AdWords related info, I thought I'd mention that this was posted about there as well:

Creating standards for clicks
http://adwords.blogspot.com/2006/08/...or-clicks.html

Inside AdWords blog : http://adwords.blogspot.com/

AWR

I’m new here and not a wack. I set a trap with three intentional no traffic exact phrase keywords at a high price around 4.00. Phrases words within a industry not used in that way. Then for the other exact phrase keywords that had high traffic, i kept reducing the price of the high traffic keywords until $.30. I had ten campaigns running. After a month of NO traffic for the trap keywords, the high traffic words slowed down on clicks. Then one night 6 clicks on each of the trap words. NOBODY can convince me this is possible. SO, is someone in my account looking for juicy priced keywords and clicking. I accused Google of this and in all the Emil I received from them, they won't mention my accusations I have reinitiated three times. I am Ignored. In my mind I’m thinking what are the odds in this, and do I know enough to call this. One in a billion. I think their dirty.

[The Generator]

Originally Posted by Mobyhunr

I’m new here and not a wack. I set a trap with three intentional no traffic exact phrase keywords at a high price around 4.00. Phrases words within a industry not used in that way. Then for the other exact phrase keywords that had high traffic, i kept reducing the price of the high traffic keywords until $.30. I had ten campaigns running. After a month of NO traffic for the trap keywords, the high traffic words slowed down on clicks. Then one night 6 clicks on each of the trap words. NOBODY can convince me this is possible. SO, is someone in my account looking for juicy priced keywords and clicking. I accused Google of this and in all the Emil I received from them, they won't mention my accusations I have reinitiated three times. I am Ignored. In my mind I’m thinking what are the odds in this, and do I know enough to call this. One in a billion. I think their dirty.

Well, that certainly is some experiment. Google dirty though? What did they mention in their emails as responses to your complaint? Also, were these words set to search and not to content (that's important because content match can trigger certain keywords that no one would otherwise search for)?

[lizcamps]

Thanks for tipping us off to this, Aussiewebmaster. I've been keen on watching my click fraud stats ever since, and have a few thoughts.

One might ask that if we are not getting charged for the invalid clicks, what value is there to examining it so closely?

Invalid charge avoidance is one thing. But what about invalid *impressions*? If I'm running a campaign at less than Google's Recommended Daily Budget using Standard Delivery, my client's ads will only show a proportional part of the time in any 24-hour period. If say 10% of those times are fraudulent, competitor or bot-driven searches, then 10% of my client's daily impressions are being wasted. By eliminating invalid clicks I will effectively get our ads to show to a 10% better qualified audience. A 10% boost in ROI would therefore result.

Modifications you can make include: Lowering Content Targeted bids to $1.00-0.00/click; targeting US- or more focused regional areas; targeting English-only; shutting off problematic words for a few days; using day-parting to only run your ads during hours when the problem doesn't seem to be occurring; analyzing problem issues using log and script-based metrics software.

I wonder whether the proportion of "caught" invalid clicks to "uncaught" invalid clicks is constant across the ad groups for a particular campaign. Because if the ratio is constant, it would behoove any Pay-Per-Click advertiser to analyze invalid clicks individually for each campaign, ad group, keyword, creative, etc.