Forum Moderators: martinibuster

Message Too Old, No Replies

Competitor Click Attack Solution (CCA)

AdWords CTR's come to the rescue

         

androidtech

4:48 pm on Feb 29, 2004 (gmt 0)

10+ Year Member



To all regarding competitor click attacks (CCAs) worries, which I've been reading about for quite some time across multiple threads:

There is a way to stop these.

For the discussion, let's rule out the scenario of a competitor being able to pull off a CCA that is human driven. For example, a competitor e-mails a hundred of his friends to all sit around on their computers, take the time to read the Ads on your web pages, and then click on your site's AdSense ads. This is a very unlikely scenario.

The competitor would have to use a fairly sophisticated bot. He can't do it himself, because there would be a concentration of clicks coming from one or a few IP addresses which would be a dead giveaway.

A bot that spoofs IP addresses, or hops through open proxy servers would be necessary.

Every so often Google, let's say 1 out of 100 times, Google serves ads that, based on the content of your page, are known to have an absolutely terrible CTR.

If the CTR for these terrible CTR ads jumps on your site, they know it's a bot doing the clicking. They would know this, because only a bot would click on it because bot's can process text, but they don't understand the content.

Second challenge. How does Google know they are serving up 'bot test' ads that have a terrible CTR?

That's the easy part. The 'bot test' ads are the very same ads that Google is disabling on the AdWords side, due to a terrible CTR for the AdWords advertiser. It's a self-generating system.

In addition, since the ads are real ads with proven bad CTR's from actual AdWords campaign data, it would be almost impossible for a highly motivated and highly skilled competitor, to try and modify the bot so that it checks the ad keywords for relevancy against the web page's text.

Even though the 'bait ads' have a terrible CTR, they were still written by some hapless AdWords advertiser that was trying write a relevant ad. The ad just failed, that's all.

If the hostile competitor is reading the ads himself, and spoofing IP addresses himself, how does he know if the ads have a terrible CTR? He can't. Only Google knows this from having actual data from thousands of AdWords impressions across multiple AdWords advertiser campaigns. So it even stands up to the human driven CCA scenario.

Thanks.
AndroidTech

europeforvisitors

5:33 pm on Feb 29, 2004 (gmt 0)



Interesting idea, but I can see two problems with it:

1) Much of the program's expense would be borne by publishers, because every "bait ad" would mean one less revenue-producing ad.

2) On low- to moderate-traffic sites, the sample size of bait ads would probably be too limited for statistical accuracy unless bait ads were a significant percentage of total ads served.

ByronM

5:39 pm on Feb 29, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month



just what is your market that you have such a concern about competitor clicks on?

I can't imagine unless your good at pissing people off that this is an issue - and if it is, alert google to it rather then trying to solve it yourself.

jomaxx

6:16 pm on Feb 29, 2004 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



IMO this sort of thing belongs on the AdWords side, as part of the overall algorithm used to spot bogus clicks. I'm sure this is far more prevalent between AdWords competitors, because the competition is head-to-head.

By comparison, websites in the same niche are scarcely in competition at all, from an AdSense point of view, because they don't compete for ads or revenue.