Imagine being able to limit how often your ad is shown to someone with the same ip address within a certain time frame.
What kind of limits if any would you like to be able to put in place when showing your ads? Would it help against click / impression fraud?
Imagine being able to limit how often your ad is shown to someone with the same ip address within a certain time frame.
One important thing to keep in mind while having conversations along these lines: it is possible for many people (dozens? hundreds?) to share an IP address. So if you were to block an IP address after one visit, you'd potentially miss out on a lot other customers visiting your site.
AWA
One important thing to keep in mind while having conversations along these lines: it is possible for many people (dozens? hundreds?) to share an IP address.
Old AOL clients, try thousands!
limiting access to the same ip address within a quick time span might work, the problem is you would only be able to stop them after they had clicked your add and you already incurred the cost.
At risk of this becoming a click fraud debate... anything from the same ip address within the space of 10/15 mins.... could be click fraud, this is just long enough to not arouse suspicion, and would probably beat most anti click fraud systems
Short answer, there isn’t much you can do, unless you are getting 100's of clicks in sec/mins... I wouldn’t stop any traffic!
That’s my 10cents
Alan
[edited by: Kings_on_steeds at 4:29 pm (utc) on Nov. 2, 2006]
At the moment I'm getting two to three clicks from the same ip within about a 2 min time frame. Just the odd time, nothing too major. I'm not saying they are bad clicks, although after clicking on my ad they never go further than the landing page. They could be genuine customers which is why it's hard to know whether I would want those extra clicks from (possibly) the same person or not.
It's easy to drop a cookie on a system [time/ip hash/unique id] to eliminate click fraud from the same ip (university, AOL, etc)...
Would you then block user agents who disable cookies? Keep in mind that even though the user is blocked, the click is still processed by the ad server, unless it is using the same technology (blocking user agents who disable cookies).
FWIW, fraudsters have ways around this, e.g. by using botnets they can generate clicks and/or impressions from a topologically widespread set of computers. Or they can just use human clickers.
You can see what Google thinks about those quick clicks, run a keyword report in Google and include the column invalid clicks.... will show you how many where stopped, wont it wont show you is the 100's it missed. lol.
Alan
The only "solution" is for "G" to go CPA - but this is another story....
If a user's browser is not accepting cookies, the system can still determine click fraud from the following: Session ID's (as applied to each unique IE window) & User Agent. Admittedly, this is not a "solution."
I figured everybody was doing it and after all it’s your money. They want to run a totally automated system and outsource most of the work to cheap labor from what I interpreted with their first response. Its up to you and I to keep them on the “straight and narrow” even if it yields nothing at times.
If the advertiser could have the control to limit how many times an ad is shown to any one person within a specified time it might help control this a bit more. I would probably set a limit of three times per 5 minutes or something like that, I'm not sure though.
If there is a way of determining if a search is made by the same person it doesn't have to be 100% reliable, if it was over 90% I'm sure it would still be well worth using for some at least.
If the advertiser could have the control to limit how many times an ad is shown to any one person within a specified time it might help control this a bit more. I would probably set a limit of three times per 5 minutes or something like that, I'm not sure though.If there is a way of determining if a search is made by the same person it doesn't have to be 100% reliable, if it was over 90% I'm sure it would still be well worth using for some at least.
Because of the variability of means of access to a site, it's difficult to put estimates on how likely it is that one person made the same set of accesses. (Especially when we're talking about viewing and/or clicking an ad, as opposed to a more detailed sequence of actions such as logging in to a site.)
That said, a limit on how many times an ad is shown over some time period and/or number of impressions or clicks might serve as a useful cost control tool for advertisers, click fraud aside. (To a certain extent, smartpricing does this.)
