Re: UKNM: ho ho ... oh.

[Paul Cook]

We looked into overcoming this problem when we were developing one of our

products and it was not a simple as it seems.


If someone is determined enough they can get by most systems. It is quite

easy to fake all request information and the only method we came up with

that wouldn't break a few rules was to automatically check the requesting

IP number. This would need to be done by the server and not the page

(someone browsing off line would find their modem dialing up again) which

adds to the problem.


Basically click throughs are a meaningless measurement for ad performance,

they also tell you nothing about how long someone stays at the site or

anything else that is of value. The Millward Brown study showed they have

no relevance to recall.


What we need is something else.... then again I would say that!


Paul


PS About 8 months ago 2 ex employees of a big publisher who produce a

travel site came back from the pub and thought it would be amusing to do

the exact same thing for Thompson Holidays. When Yell twigged what was

going on they found that the script was running on this (not so hard to

guess) publisher's site! 


PPS The ad servers don't attempt to deal with this and neither, to my

knowledge, do the auditors.


----------------------------------------------------------------------------

Paul Cook				The Makers of:

Managing Director			Plannet for Buyers - Knowlege is Power

Red Eye International		NetOutcome - What's a click through worth?


http://www.redeye.com		AdReporter - Full Reports, half the price



At 02:42 PM 09/04/98 +0100, Richard Houston wrote:

>Ben,

>We know that this type of spoofing is not new and that it can be avoided.

>However, the point is that even a comparatively respectable property such

>as Yell *wasn't* guarding against it and was properly duped.

>

>>For those who have not read the article the gist is that you can write a

>>program to simulate ad impressions or click throughs. If this is new to

>>anyone I would be surprised although personally I would never have tried

>to >demonstrate it. > >Actually this is not new. I have had discussions

>about this problem for the >past three years (if not longer). One thing


>worth noting is that this is not >a problem with Porn sites where each IP

>address is counted only once a day. > >True its a problem if you are not

>aware of it but there are simple methods >of avoiding it. > >yours, > >Ben

>

>-----Original Message----- >From: Richard Houston <richardwebmedia [dot] com>

>

>To: uk-netmarketingchinwag [dot] com <uk-netmarketingchinwag [dot] com> >Date: 08

>April 1998 17:23 >Subject: UKNM: ho ho ... oh. > > >>Good stuff: >> >>

>>>http://www.internet-magazine.com/news/apr98/08a.htm

>Richard Houston

>------------------------------------------------------------> Webmedia

>Group, 21 Noel St, London, UK, W1V 3RD Land: 0171 494 3177, Fax:  0171 434

>1304, Mob:  0961 11 60 59

>