Does this become less true the more niche the website is? People who read my math blog should be served ads for technical things (math books, online courses, etc), regardless of recent search history.

In reality, I have a math blog with ads for Cap'n Crunch cereal and strange new yogurts which seem very distracting and poorly-targeted when compared to, say, ads for O'Reilly math/CS books or a MOOC on probability and statistics.

>I have a math blog with ads for Cap'n Crunch cereal and strange new yogurts

Seems like to me the algorithm gods decided that your site was better suited to advertise to parents of toddlers.

> remove sources of income that people find relevant in exchange for ads that some higher up in marketing thinks relates to its users.

Yes, that's a move toward integrity and away from endangerment and excess income.

If "some higher up in marketing" doesn't understand their users, the problem is they're out-of-touch and that should be corrected, as opposed to endangering user's privacy.

Currently, my tiny site operates this way, with hand-chosen Amazon ads and no google Adwords.

>excess income

I don't think this a problem most small publishers have.

They need to realize that eventually the market will drive their users elsewhere. Content is king and there is more content elsewhere.

If you provide a worse user experience with the same content as another, then income will dwindle altogether.

Innovate, move forward.

If you have an interest in the 3rd party web, you should have an interest in a functional ad technology industry. The market might just drive small publishers out of business. Or drive them and their customers to walled-gardens, so we'll be reading their content on Facebook, being served Facebook ads.

And I'm not a fan of going back to AOL. But the ad tech industry and the "small publishing" industry is lazy and greedy. And I won't support them if they treat me like a product. If they don't I will support them.

I'd like to see Netflix for written content and I'd pay for it.

There are lots of automatic ways to make money like this, even more so on the cost side of a business. They are always going to look great on a spreadsheet but what you're missing is the opportunity cost of all the people that are turned off by the experience. They will fade away in a manner that's harder to measure than CTR.

I suppose that the decrease in daily / weekly / monthly audience is easily detectable.

Probably it's not very realistic to hand-pick individual ads, unless you run a very niche site. Probably it's still realistic to hand-pick the automatic ad streams you'd agree to show. I suspect the major ad networks allow you to be quite specific.

Podcasts do this very thing and it's lucrative enough for the lot of them to stay afloat while providing great content. The only thing I'd say sucks about it is the price for curated ads are probably prohibitively high compared to highly targeted, algorithmically picked ones.

> Intent-driven ads are far more relevant to a site's audience than hand-picked, static ones. They have much higher CTR and conversion rates, when done right.

Not in my experience. My hand-picked ads are alway at a higher CTR than programmatic. Intent-driven programmatic gives me a 2x CTR over random static ads. Hand tuned gives me about 10x more CTR.

You may be able to optimize down a set of hand-picked ads programmatically, but you still need that manual curation in the process.

More important is losing your audience to a bad user experience via programmatic ads when they visit your site.

Also, I'm pretty sure Vogue or the advertisers use DoubleClick as just the ad hosting network for analytic purposes. They are likely sold on a display basis in Vogue.

Good point. A combination of handpicked optimization and programmatic, intent-driven placement is likely optimal.