Swapping Adsense ads for fast repeat visitors

When Adsense won't credit 2nd or 3rd click, use diff't ads?

[billegal]

Here's what I'm concerned with (thanks to that tracking script):

User 1 (based on IP) clicks 3-4 ads on my site. Adsense gives credit for the first one or two...maybe. The numbers are not sensibly matching up for me. I'm not thinking "fraud" because it makes complete sense for one of these visitors to click more than one ad because these visitors are clearly looking for something to buy and the ads match up nearly perfectly for the four spots.

Is anyone else considering swapping the Adsense code for something else on those return visitors? If I'm not getting credit, why give free impressions/clicks? It surely benefits the advertisers for these visits.

Does anyone have any ideas about how those multiple clicks on different ads are discounted?

[richmondsteve]

That's a good idea you posed. Before the click tracking software I have experimented with disabling AdSense ads on a per-user/visitor basis after X page views or mixing in affiliate ads after X page views. I couldn't accurately measure results, but it seemed logical. It didn't dawn on me that with the click tracker I can now automatically trigger affiliate ads instead of AdSense ads (and vice versa) based on what a user has clicked. Excellent, excellent idea!

[billegal]

Thanks, although I think the general idea is still fairly raw. In order to refine it, does anyone have any input into when to swap Adsense ads out? Is it one click per IP per 2 hours?

I'm guessing that formula is a little bit simplistic. I'd think there'd be some way to get this pinned down in order maximize revenue.

If Google would do this for us, the equivalent of what they do now for non-paying ads, I'd stop thinking about this. I mean if they know they're not going to pay for anything that a visitor clicks, why should Google get to show ads?

[jomaxx]

I'm not convinced you can analyze clicks this accurately. The scripts I've seen published can generate false positives and can also miss clicks (from Netscape and Opera at least; probably other scenarios as well). Add in unknown time delays in reporting data, and there's no way you can balance numbers down to the last click unless you are serving an extremely small number of impressions.

Anyway, from my testing, if the user simply returns via the back arrow then the page is served up from the browser cache, AdSense and all. If you want to serve up something different, maybe some kind of cache-busting code will do the trick.

[adfree]

On top how would Google sell this to their top paying advertisers? Where do you draw the line and explain to them: under these and those circumstances you will be placed 5th or your ad will not be shown alltogether...

Tricky...

[billegal]

Here's how Google has to sell it to their customers: In the absense of fraud, if the ad gets clicked, I should get paid. That's PPC.

I don't know what people consider a small number of clicks. There'll all important to me but I can't reveal any numbers. Suffice it to say that 25% were not credited. If I can figure out which 25%, Google's ads won't show up to those visitors.

As for the false or missing clicks, I'd like to learn more about this issue so I can eliminate them.

Any chance Google will tell us when subsequent impressions will never generate click revenue or at least put in the alternate ads then?

[richmondsteve]

I'm not looking at this so much from the standpoint that Google might not credit subsequent clicks (though that is a good idea) as in my particular case at some point users are less likely to click an ad. My site is a very focused niche site. 95% of the pages with AdSense enabled rotate between the same 6-10 advertisers each day promoting a pool of 2-3 types of services. On a per visitor basis users are shown mostly AdSense ads for the first X page views, then mostly affiliate ads after that. It might not be a valid assumption, but I assumed that on average a user who has seen the same AdSense ads for X pages is less likely to click them on subsequent page views. With the click tracker I'm considering switching to affiliate ads after a user has:

1. Both clicked on an AdSense ad and moved to a different page on the site

or

2. Viewed X pages without clicking an AdSense ad.

I realize the limitations of the click tracker JS, but for my site a rule like this may work well. The goal is to hopefully increase earnings per visitor by tailoring the ads shown to user behavior. Does this make sense?

[yump]

Isn't another key part of this to know on average how many times an advert has to be present to be noticed and also how many times it has to be present to be clicked by someone who's interested. We know from our own research that many visitors don't even notice basic site changes for several months.

[loanuniverse]

richmond: You should also take into consideration the possibility that there are more than 6 advertisers in adsense for your site. For example, I have found some UK based advertisers that are being shown to "I am guessing UK visitors". There could also be other advertisers that might pop-in and you have not noticed. Lets not even get into the next time that Google decides to start messing with their targetting algo..... or how many pageviews you get per visitor.

If you get an average of 5 or less, I would not bother with any changes.