Site downloader triggering clicks?

Can a site downloader generate cliks

[frox]

I was having a record day (50% clicks more than any other day), but I got suspicipous and started looking through the logs.

There are several thousands page views from a very small number of IPs. I am starting to think that somebody was using a downloader to copy all my site and this somehow triggers clicks.....

I wonder if this is possible or it's just my paranoia.

Does anybody have experience in this, i.e. a user that uses a web downloader and generates (INVALID) cliks.

I any case, I have cut the ads on the site and written Adsense support. As soon as possible, I will implement a check that only serves ads X times to each IP per day.... I will lose some "real" clicks, but better to lose a fraction of the clicks than to use all the account!

[Hobbs]

Google usually compensates for this, but it never hurts to inform them as you did to be on the safe side. Blocking the IP of the attacker will do you good too but what happens when you are sleeping or not monitoring stats? You just trust Google to do its job.
As for your question on clicks, yes a site downloader will generate clicks if it is setup to follow outside links on your site, most downloaders set themselves to restrict the followed links to the downloaded site though.

[frox]

A small counter-measure I took.

1) If the user Agent contains any of the following:
BOT
HTTRACK
PYTHON
COPIER
CRAWL
SPIDER
ARCHIV
SLURP
MIVA
WGET
do not display ads

2) If the user has seen more than 30 pages in the last 5 minutes, do not display ads.

This is not to avert intentional click attacks, just to avoid that web downloaders like HTTrack trigger invalid clicks.

I will experiment a little, and then put the ads back on my site..

[ganderla]

<edit>
wrong thread.
<edit>

[jomaxx]

Never happened to me even though I get hit by a variety of rogue spiders every single day and watch my log files like a hawk. It would be possible to specifically write a spider that can do this, but in general a spider will never even execute the Javascript necessary to generate an impression, much less actually click on the ads.

If you have the information in your logs, why not just take a close look at the IP addresses and the browser IDs involved? In most cases if there is a pattern it will stand out like a sore thumb.

[frox]

There are 3-4 IP alternating with WebCopier downloading thousands of pages.

But as I do not have an Adsense click tracker I am not 100% sure it's the WebCopier pageviews that trigger clicks....

[jomaxx]

Webcopier is very popular. I've used it and I'm reasonably sure it will never generate false clicks. (Or at least that software product won't. In reality any spider could theoretically identify itself as "Webcopier".)

Besides, if it did then Google would have figured this out on day 1. It would certainly be on their first line of defense against erroneous clicks.

[frox]

Yes, just out of curiosity I tried Webcopier and saw no problem there....

Anyway, I am now putting ads up again...

Cross fingers!

[tebrino]

Just a note, most site downloaders fake their user agent header.

[jetteroheller]

1) If the user Agent contains any of the following:
BOT
HTTRACK
PYTHON
COPIER
CRAWL
SPIDER
ARCHIV
SLURP
MIVA
WGET
do not display ads

And what if Googlebot and Mediabot wonders where are the ads?

Could this not create an alert, that the pages are presented different to Googlebot than to a visitor?

[frox]

Tebrino:

Just a note, most site downloaders fake their user agent header.

Yes, most site downloaders CAN fake their user agent but yet they come by default with the proper description and in my logs I see them so.

As a secondary measure (for the ones that change their user agent) I have enabled the 30 pages/5 minutes check. My site is a dictionary-type like site, where average pageviews/unique ratio is about 2. So I guess there will be not a large loss of income.

[frox]

And what if Googlebot and Mediabot wonders where are the ads?

Yes, jetteroheller, I thoght that too. While technically this is a form of cloaking, the contents are not changed in any way, so I suppose it's a "good" form of cloaking.

I will ask Adsense support about this, anyway...