Forum Moderators: martinibuster
Hurricane hits. Adsense revenues are down. US traffic down!
This type of natural disaster will have a "noticable" effect
on Google Adsense statistics (across the board) right?
In doing so, it presents an interesting opportunity for the
Google engineers. Also - a major threat to fraudsters. With
such widespread fluxuations in Google Adsense averages over
the last few days, Google engineers should be able to find
those websites that do NOT suffer ...
The power of Google lies in "averages". If there is a tiny
downturn across the board, then those that are faking their
traffic and clicks are going to stand out from the crowd.
Thoughts?
Personally, I didn't notice much difference in traffic or earnings until the lead up to the long weekend. Doesn't automatically make me a fraudster, despite the majority of my traffic being US-based.
If there is a tiny
downturn across the board, then those that are faking their
traffic and clicks are going to stand out from the crowd.
I would bet the smart ones have taken things like holiday weekends and world events into account.
A simple example: specifically design (or select from real ads) a VERY low-performing ad (low CTR). Test it on 10.000 sites.
Get the 100 sites where that ad has higher CTR, and give a score of 1 to those sites.
Repeat this with 100 other ads. After this process, if a site has a score greater than N, cancel the account.
i.e. if a site as a regular pattern of clicking VERY low-performing ads as if they were good ads, that site is cheating
[edited by: frox at 9:49 pm (utc) on Sep. 4, 2005]
On a serious note, my heart breaks watching the images from the Gulf coast. Once New Orleans rebuilds, I'm definitely going back.
Interesting theory. I wonder if G spoofed some of the auction ads for that purpose, eg "new and used dead pope"!
Arrrrrgh, the black helicopters!
After reading several posts at WW, I noticed a trend among
a few people indicating that perhaps Katrina was responsible
for a downturn in earnings. It got me thinking about the types
of people that wouldn't be effected by this natural disaster.
Obviously, someone who continues to earn regular amounts of
clicks in a "certain" industry, may be telegraphing fraudulent
activity - if they are not "in-line" with the average statistical
trend changes associated with such a disaster (assuming they
even exist and can be quantified).
Now, before all the stat junkies & Google lovers attack me,
consider the idea of someone in the travel industry - or
someone who attracts visitors for keywords that perhaps
relate to this natural disaster.
Either one of two things would happen. The searches in total
and subsequent clicks would rise across the board, or fall -
depending on the nature of the information. This may seem a
little difficult to understand, because I am having trouble
explaining it ...
But surely, a disaster of this size - coupled with external
events (people watch more TV etc) would have some sort of
noticable effect on particular areas of web traffic. It is
these areas that would offer some sort of trend, unto which
fraudulent activity may be emphasised.
I don't mean to start a flame war here, and I think we've
had enough of those "oh - jee, looks like that makes me a
fraudster" type posts. Not called for. Just interested in
learning your thoughts on this.
but ...
Perhaps particular "keywords" that are associated with those
areas can be understood in terms of trend movement. Lets just
assume, that for instance
"Holidays to New Orleans" got say 10,000 searches a month,
and maintained a fairly standard click rate over the last
6 months. Then suddenly, a HUGE reduction of queries were
noted. Adsense revenue for webmasters goes down ...
and yet, some webmasters maintain high click rates etc. It's
this type of statistical anomally that I am interested in.
This would occur for a lot of keywords.
Doesn't matter if the trend was up (more interest), or down
(less interest in particular terms) - there would be a big
movement, relative to that sub-set of keyword/queries and
from this average movement, fraudulent activity would be
pretty easy to spot.
Lets face it - if there is a downturn (or upturn) in terms
of average impressions/clicks for say the top 500 keywords
relating to New Orleans, and some webmasters fail to show
this trend, then ...
