Your post also made me think of the genius Sony ads for their HD televisions. I’ve watched those repeatedly and shown them to many friends.

Effectively the market rate for impression is near enough zero to be considered so is it not.

A conclusion here is that ad models are failing in all sorts of ways, and some of those ways are to do with the fact that mainstream advertisers are not savvy when it comes to matters technological. You need a maths geek, an ethnography geek, a tech geek, and and geek to advertiser geek to talk to the advertisers in a way they can understand and those roles are by no means mainstream – outside of O’Reilly I suspect.

That ads that work are the adwords ads that appear on Google. Google gets all the above and more importantly employs a whole bunch of people to do things like look at search trends for predictive reasons – they ain’t doing that for kicks – it’s so they can build it into their pricing models for adsense bids. I believe that same approach will work for Newspapers et al, but you gotta believe that will work and you have to make the jump that says we will follow the data religiously (so to speak). Advertising has become an engineering exercise, not about the play of font upon image or the quality of light falling upon the car as it swoops down the mountain road. that method does not work at all now that we have crossed the rubicon in the way we consume our information. So you got to get smarter about it – Honda’s live advert with the sky dive (done here in the UK last year), Another Uk supermarket trying a live 3 min how cookery lesson. Adverts people will not FF through. Ads as useful content.

I bet no newspaper would ever go to an advertiser and say – we aren’t running your ad – it just didn’t perform well enough for that space on our paper, you need to go away and do it better.

Google does that all the time with adsense. Theirs is a different way methinks.

Domain expertise is important for advertising people to have to some degree. The best advertisers understand this.

I think one of the major issues I’ve found (at least in science publishing) is that the people buying ads are not scientists, their background is in advertising. They don’t have much of a grasp about the ways of scientists, how literature is read, or even what scientists do with their days. When we tell a potential reagent manufacturer that we publish protocols that use their reagents, very often the first question asked is, “what’s a protocol?” Our papers are directly relevant to selling their products, but it’s often an uphill climb to get them to understand why.

There’s also a huge difference between putting an ad for Coke on the 7th page of People Magazine and an ad for a centrifuge on the 7th page of Cell. Readers are going to flip through People and look at each page. Most Cell readers are going to the website and downloading the pdf of individual papers they want to read. Most will never even look at the html version of a published paper, let alone the print version. Those who do still get the print aren’t idly skimming through the pages for branding opportunities.

Good points. I had some references to the scarcity of print in a first draft of my reply, but took them out because I actually think for advertising a print journal was a proxy for an audience. And I think reaching the right audience with advertising is still valuable since audience (and attention) scarcity hasn’t changed much. That’s why I think for advertising the main issue I see is a mindset that’s inappropriately about transactions (advertising as direct marketing). If online advertisers had completely embraced advertising online, they’d be happy to put 50K impressions in front of the right specialists or domain experts, and they’d measure differently — more on brand recall, creative effectiveness, market share, mind share. Now, they measure click-throughs. In fact, a person could argue that online advertising is much more valuable than print because it does everything print advertising does AND it provides direct mail-like transactional opportunities.

For Free…

Here’s the (possibly) genius bit. To do this, you have to sign an agreement that allows the Guardian to run adverts next to it’s content on your site. It’s Adsense for their content, but done backwards as it were so that the Guardian controls the ads AND the Content. This is another case of Scot Karp’s link economy by the way.

Kent – your point about the print ad prices also needs to take on board the scarcity aspect of the print real estate – in print you can’t just go and find another paper with the reach/prestige/target demographic/whatever of the NYT. Not a problem on the web. Also as David mentioned – how many creative types are mathematicians?

One thing that was implicit in Tim’s talk was the fact that he has a team of folks doing detailed analysis of the various business models he runs and the experiments. Those guys really do fail fast cheaply and often, and they do it pretty successfully (at least to the outside world). There are some skill sets we really need to get hold of and analysis of that thing we call online is definitely one of them – it’s more than visitor numbers and clicks, it’s deep mining for user behaviour patterns and all sorts.