It should be readily apparent that audience size, uber alles, does not yield the best content -- nor does it yield the best audience. I would bet that 10,000 legitimately high-quality readers for a specialized blog post are worth more to advertisers than 1 million readers of a tossed-off Buzzfeed poll. The problem is, advertisers don't know that, and nobody's making an effort to educate them. Everyone's still chasing impression levels and scale, presuming that conversions will always be awful.
The pubs who break this cycle will be the ones who invest in high-quality journalism and high-quality analytics and innovative ad models. It's time for us to completely rethink the journalism ad model. Only then will we change things.
I apologize if I sound like I'm on a soapbox here. It's just that I've spent the better part of this past year wrestling with this problem at various publications, and close to a decade wrestling with it in the TV business before that. The Homogeneous Audience Theory is one of the great, untackled problems in media today. Everyone knows audiences can be segmented. Few people do it. Instead, people assume conversion percentages will always be a tiny fraction, and thus, their job is to multiply that fraction by a huge set of eyeballs. Let's reengineer that equation in favor of the conversion multiplier, by way of building more purposeful and curated audiences.