jared (jared@netspacenospamnet.au) wrote:
> It is unclear to me who are the swindlers and how they benefit. Are
> people clicking on competitor's links to drive up costs?
Yes, as a matter of fact they are. There are many types of click
fraud, such as:
* Competitive clicking that drives up a competitor's ad spend, and/or
affects their ranking in the sponsored listings.
* Competitive clicking of one publisher by another trying to get that
publisher kicked out of the content ad network.
* Publishers clicking on their own ads, pocketing the revenue.
* Investors in search engines or ad networks clicking on ads,
increasing ad spend, and thus the profitability of the SE or ad
network in question.
Click fraud can be committed various ways, such as:
* Auto-clicking "bots", possibly distributed via spyware or adware.
* The same, operated through networks of dedicated machines.
* Click rings organized to spread fraduluent traffic across a large
range of IP addresses, user agent types, etc.
* In some cases, sites are designed so that novice web surfers cannot
exit them without clicking on an ad. (This isn't always considered
click fraud, but it contributes to the drain on an advertiser's
budget.)
> Not everyone who clicks will purchase, but that's what advertising is
> about.
True, but these techniques can drain advertisers budgets quite easily.
Some very savvy advertisers realize this, and put caps on spending, or
move spending to places they feel are less vulnerable to fraud (such
as the search engines themselves, rather than content networks). But
most advertisers are unaware of click fraud, or don't feel they can do
much about it. Some liken it to a "tax" on Internet advertising, and
others have large enough budgets that they can absorb some fraudulent
clicks.
Furthermore, it is not possible to determine that a click is
fraudulent in all cases. (Otherwise, the SEs and ad networks would do
this.) Especially when bots or click rings are spread through the
Internet, their traffic shows up just as any other traffic. One
cannot be sure that they are filtering out legit clicks when filtering
out clicks that look a little unusual, e.g. a small rise in accesses
to one site or a few new IP addresses.
Bruce Schneier recently wrote an article for Wired magazine on click
fraud from a more formal security perspective. You can find it at
http://www.schneier.com/blog/archives/2006/07/click_fraud_and.html.
Personally, I am glad this issue is starting to get the exposure it
deserves.
gregbo
Austin, Texas, USA