Forum Moderators: martinibuster
The writer (Nate Elliott) thinks that Google will fall because the new competitor will be ready to combat this year. And then...
What kind of publishers will sign in with an ad network which pay less? Cheaters? Pills/S*x/Gambling/Credit-reporting/Car-insurance sites? Where will be published the ads?
And the best of all: the article is under a 728x90 Adsense leaderboard!
I think he is suggesting that while possible, such segmentation is very unlikely.
That's exactly it. Of course it is possible but to do a manual segmentation defeats the whole purpose of the system that was meant to be automated.
And we also know that as soon as Google were to automate it then the fetish site would find a way to optimize to be a information, community and commercial site :-)
JAG
I think he is suggesting that while possible, such segmentation is very unlikely.
That's hard to say. With search, Froogle is a step in that direction. And targeting ads by keyword alone is a halfway measure at best. If Google doesn't offer targeting by content type and/or site, a competitor almost certainly will.
As for automation, giving advertisers more choices wouldn't require a sacrifice of automation at Google's end. In fact, it would decrease Google's need for human "quality police," because market forces would determine which publishers were included or excluded from advertisers' campaigns.
If Google doesn't offer targeting by content type and/or site, a competitor almost certainly will.
I don't know what else to say to that other than Google bought Sprinks who did content type and categories and buried it. Shut it down. Replaced it with the current scheme. That should give a clue as to what path Google is on but I'm not sure why you don't want to believe it?
The automation part is in setting up the categories and ensuring sites are properly categorized in an on going manner. That would most definitely require quite a bit of hands on especially since the technology Google uses is page specific. How do you handle that manually? How do you handle sites that have all possible categories because they are so large? Even Overture has admitted that they have to be selective in the sites they allow to participate because of the overhead.
JAG
Historically, advertisers have placed different valuations on different types of content--e.g., high CPMs for editorial or search pages on "money" topics and rockbottom CPMs on chatroom pages. So there's certainly a precedent for differentiating between pages that Google can identify as being editorial pages, forums, blogs, e-commerce pages, or whatever.
As for giving advertisers include/exclude filters similar to the blocking filter that publishers have, that wouldn't require any Overture-style editors or any human intervention whatever on Google's part. Without such filters, Google is delivering what amounts to a run-of-network audience. As I've said before, keyword targeting isn't everything: targeting by audience is equally important if Google and advertisers want to fully exploit the potential of contextual advertising on niche sites.
Leaving the cyberworld..... contextual advertising has been around for a century+ in magazines and newspapers. You sell more flowers and gain more mind share on the obit pages. EuropeFV likely generates better quality leads due to the quality of his work. Such leads may convert better in time - not on the click through, but a month later when the deliberate, well informed shopper finally makes a buy.
Depends, in part, on your target behavior. Do you want to boost your credibility by having your ad appear on a quality site? Do you want to gain mind share? How long and complex is the purchase cycle for your product/service? T.M.V., unless as an advertiser your spend is about "click now and buy....dammit". More control over targeting your demographic would help.
Lot's of variables. Lot's of reasons why contextual ads will be around forever. Magazines and websites are not that different. Targetting and restricting ad appearances will improve to improve the ROI. Advertisers will see to that by voting with their dollars - placing them with the best targeting systems.
Now, define best ;-)
T.M.V. Too Many Variables for the inadequate analytic tools to solve the riddle at this time.
DMOZ
You're comparing apples and oranges. I'm not talking about topic categories like "computers" or "travel," I'm talking about content categories like "information," "community," and "commerce."
Apples or Oranges doesn't matter so much because it's the process we were talking about and my point was that Google has already ruled on what they think about categories, however defined.
As I've said before, keyword targeting isn't everything: targeting by audience is equally important if Google and advertisers want to fully exploit the potential of contextual advertising on niche sites.
Yep. I've said that since day one myself :-)
Now, define best ;-)
A system that delivers the right advertisements based on many factors and provides the highest payout to the content site and achieves the advertisers campaign objective. Simple, isn't it :-)
JAG
...it's the process we were talking about and my point was that Google has already ruled on what they think about categories, however defined.
Yes, I remember that you mentioned the closure of Sprinks as evidence that Google doesn't like categorization. Unless I'm badly mistaken, Google's acquisition of Sprinks had more to do with gaining distribution on About.com than with Sprinks per se. Google wanted advertising impressions on About's niche sites, About.com and its financially strapped corporate parent needed money, and the demise of Sprinks (which had already seen better days) was inevitable under the circumstances. There have been a lot of changes and cutbacks at About.com since the Primedia takeover, and from Primedia's point of view, guaranteed revenues from Google must have looked like too good a deal to pass up.
In any case, the demise of Sprinks isn't proof of anything about Google's plans for AdSense (other than serving serving ads on About.com).
Unless I'm badly mistaken, Google's acquisition of Sprinks had more to do with gaining distribution on About.com than with Sprinks per se. Google wanted advertising impressions on About's niche sites, About.com and its financially strapped corporate parent needed money, and the demise of Sprinks (which had already seen better days) was inevitable under the circumstances.
Which would indicate that their advertising efforts, using what they were using, wasn't paying off. I guess it could have been because of lower paying advertisers but why not just replace the Sprinks advertisers with the AdWords advertiser and keep the model if it was a good one?
Google also bought and then buried Applied Semantics and their algo.
I suppose even though it walks like a duck, looks like a duck and sounds like a duck it could still be cow :-)
JAG
Which would indicate that their advertising efforts, using what they were using, wasn't paying off. I guess it could have been because of lower paying advertisers but why not just replace the Sprinks advertisers with the AdWords advertiser and keep the model if it was a good one?
Whoever said the Sprinks model was a good one? (Well, the people at Sprinks, I suppose, but they aren't posting in this thread.)
Whoever said the Sprinks model was a good one?
Sorry about that. I thought you kind of did because they used the category model and I thought you suggested someone use it.
JAG
Sorry about that. I thought you kind of did because they used the category model and I thought you suggested someone use it.
No, I wasn't suggesting anything like the Sprinks model.
BTW, I thought Webwork's post was excellent (and I agree with it completely).
Most of the campaigns I've worked on over the last 8 years or so were designed for immediate sales or action with the exception being for very large companies who could afford to wait while the branding effect took place.
Those companies had budgets in the 100k+ per month range. Almost all campaigns less than about 50k per month were designed for immediate action.
Now for the ones who did have a cycle we used custom software that we wrote in house to track it. It's not that there were too many variables but rather when to apply what set of rules to the campaign objective. All campaigns boiled down to just about 5 objectives and were quite easy to figure out what was going to work and where.
JAG
