Best strategy for developing negative site list?

Since I'm limited to 25 I want to make the most of it

[limitup]

Since we're limited to blocking our content ads from showing on up to 25 sites, I want to make sure we are blocking the 25 sites that are sending us the most junk clicks. Thing is, once you block a site you obviously get no clicks from that site, so as time goes by how do you determine which, if any, new sites to block? It doesn't really seem possible unless I'm missing something.

A simple example in case I didn't explain myself correctly. Say I block the 25 sites that are sending me the most bogus clicks right now. A week later I find another site that is sending me tons of worthless clicks, and so I'm thinking of adding it to my 25 in place of another one. But it seems like I really have no way of knowing which sites are or would send me the most bogus clicks, since once I start blocking a site I don't get any more clicks from them. It's like a catch-22.

Why doesn't Google just let us block 250, 2500, or better yet, unlimited sites? Do they really care about us advertisers? Are they worried we will block too many sites and their revenues will take a big hit? I mean, they have to know that most people would like to block 100s of sites, not only 25. So what's the point? Grr.

[FromRocky]

I don't know where you get a limit of 25 sites. I believe this limit if it's is applied for each ad campaign.

[limitup]

What does it matter if it's per campaign or not - since you have no idea which sites are showing your ads for which keywords it's not like you could separate keywords into different campaigns and make use of the 25 "per campaign" limit to block the specific sites you wanted to ...

[stuartmcdonald]

Good question. I just blocked the first 25 I came across that looked dubious -- for the moment I've given up on the tool and am waiting for Google to increase the limit -- with 25 they're just wasting my time.

What would be REALLY useful would be a set of stats that broke up the content network into traffic by publisher site, something like:

www.example1.com impressions 2,345 clicks 233
www.example2.com impressions 543 clicks 5
www.example3.com impressions 10 clicks 0
www.example4.com impressions 12,987 clicks 0
www.example5.com impressions 32,888 clicks 2
www.example5.com impressions 7 clicks 6

So then you could look at those stats, put your tea leaves away, and make an informed decision about where you want your adverts to show.

[corbing]

So then you could look at those stats, put your tea leaves away, and make an informed decision about where you want your adverts to show.

I'm not sure how those stats would help you make an informed decision. Looking at your sample list, which one would you ban?

Without looking at what happened after they arrived at your site, those numbers mean nothing to me. You need to know things like how long did they stay on your site, how many pages did they view, and did they "convert" (which could be a sale, a lead, a newsletter signup, a click on one of your banner ads, etc).

[limitup]

I wouldn't say that's necessarily true. Yes it would be ideal to be able to track conversions all the way back to an individual content site (which you can if you really want to), but it's not absolutely necessary.

You could simply ban sites that have nothing to do with your subject matter, assuming that visitors from these sites will be worth significantly less to you. Or maybe I'm the only one seeing their ads being displayed on completely unrelated sites? ...

One of my sites is targeted only towards retired persons. Looking through the content sites sending me a lot of traffic I have already found 3 job sites, 5 or 10 "get rich quick" MLM scheme type sites, etc. These are useless to me ...

[Steve6]

This has been covered in another thread [webmasterworld.com], but I believe it applies here. It seems the biggest issue is that clicks from content sites are simply not as valuable (in terms of conversions) as clicks from searches. In one of my groups, the conversion rate from clicks was 13.79%, while the conversion rate from content was 3.36%. The cost per conversion was about 4 times as much. So, as suggested in the previous thread, I made a new campaign, and copied the best performing keywords from my first campaign. Then I un-checked the "Content" option in the first campaign, and set the max CPC for the second (Content) campaign to $0.05. Now at least the junk clicks don't cost me as much.

[inasisi]

One relevant metric that you could use for analyzing which site to ban, would be the number of clicks per day from these junk sites. So you have a list of all your junk sites and you sort it by clicks per day. You ban the top 25 sites from this list. If a new site gets you more clicks per day than the 25th, ban it and "unban" the 25th site. This is assuming the clicks are around the same cost. If you are able to track the clicks down to their cost, then a better metric would be cost per day.

All this is unnecessary complications due to this really tiny limit of 25 sites. Instead of doing all these calculations, I would be better off honing my ads or targeting more keywords.

[limitup]

Thanks for all the replies. I bit the bullet and had my techie create a custom tracking system so we can track conversions all the way back to individual content sites. We'll just track everything and ban the ones with the lowest conversions.

[limitup]

... and what we're finding out is that although there are a handful of sites that send a lot of traffic, the majority of clicks are spreadout over tons of very questionable sites. 100s of them, sending 1 or 2 clicks each per day. So until we can block hundreds and hundreds of sites I just don't see how to make these content campaigns work. Google just doesn't do a good enough job screening these sites.

I think the new strategy will be to use this new tracking system to identify the sites we DO want to advertise on, and then when their new CPM site targeting system comes online "in a few weeks" we'll know which sites to target.

[ownerrim]

"What would be REALLY useful would be a set of stats that broke up the content network into traffic by publisher site, something like:

www.example1.com impressions 2,345 clicks 233
www.example2.com impressions 543 clicks 5
www.example3.com impressions 10 clicks 0
www.example4.com impressions 12,987 clicks 0
www.example5.com impressions 32,888 clicks 2
www.example5.com impressions 7 clicks 6 "

Wouldn't this be opening the door for advertisers (who'd like to cut out ALL of the crap sites that send worthless clicks) to directly contact those sites that send the most relevant and productive traffic? Wouldn't really be in google's best interest's to do this would it? It could cut them out as the middleman.

[stuartmcdonald]

True perhaps, but personally I'd be using it to cull the rubbish -- one of the main reasons I use adsense is because it means I don't have to worry about dealing with the advertisers directly.

[limitup]

Wouldn't this be opening the door for advertisers (who'd like to cut out ALL of the crap sites that send worthless clicks) to directly contact those sites that send the most relevant and productive traffic?

You can do this now if you really want to. It only took my techie about an hour to make the necessary upgrade to our in-house tracking system ...

Check referrers, parse out the content domain, set a cookie with the domain or an ID that references the domain, log to DB, then check for the cookie on your conversion page. Whip up a simple script to analyze the results. Done.

Now we know exactly how many clicks we're getting from each content site, and the exact conversion ratio for each. If Google doesn't start providing us with REAL and useful tools to manage our content ads we WILL be going to the sites directly. This 25 negative site limit type of thing is complete BS. I think they are just scared of how much revenue they'll lose when advertisers block 80% of the content sites they are getting clicks from ...

[stuartmcdonald]

limitup, of the sites you're seeing clicks coming from, at ths early stage, what percent do you think you'll be blocking if you could? 80% realy?

[limitup]

Too early to tell as we just set it up. While we're getting a lot of clicks from a small handful of sites, we're also seeing tons of sites sending us 1-5 clicks per day. For now we will just ban the 25 sites that are sending us the most junk clicks per day, and hopefully G will let us ban more soon. I can definitely see us blocking at least half of the sites, or more. This is just what we're seeing though, it's going to depend a lot on your industry, your keywords, etc.