Back in December, Peter Kafka summed up the most important question with regards to the future of online advertising. Do advertising dollars ultimately end up where people spend their time, he asked, echoing Kleiner Perkins’ Mary Meeker says, or, pace Bernstein Research’s Todd Juenger, is that a “fallacy”?

I’m with Juenger on this one. As he says, “time spent is supply, advertising spend is demand… Just because there is a large and growing supply of Internet inventory doesn’t mean advertisers have a correspondingly large desire to deliver more Internet impressions.” Indeed, as the price of online inventory continues to fall, it seems just as likely that online ad spend will go down (because the ads being bought are getting cheaper) as that it will go up.

According to Meeker, some 67% of all ad dollars are spent either on TV or in print. And according to Juenger, ad spend on TV actually went up, between 2009 and 2012, even as Americans’ attention moved away from TV and towards other screens. That makes sense to me, mainly because of the point I was making back in 2009, drawing the distinction between brand advertising, on the one hand, and direct marketing, on the other. TV is brand advertising; online ads, by contrast, are closer to direct marketing.

When people like Meeker look at ad spend, they’re looking mainly at brand advertising. Brands are valuable things, and billions of dollars are spent every year to keep them that way, mostly on TV and in print. And if you have a big national brand, there’s really only one way to reach a big national audience: you need to buy ads on TV. Doing so is expensive, but it’s necessary, and it works, which explains the huge sums of money which still flow into TV every year.

As Juenger explains, the audience for network TV has been shrinking by 1.8% per year for the past 20 years — but at the same time, the audience for every other TV channel has been “atomized into increasingly tinier fragments”, leaving the networks the only game in town for advertisers wanting scale. The result is that network-TV ads have been increasing in price by 4.9% per year on a per-person-reached basis, resulting in total revenues growing, by 3% a year, in a market which is actually shrinking.

The corollary to the continued success of network TV is the utter irrelevance of online ads. Here’s a handy chart from Nielsen, breaking down the amount of time we spend in front of various screens each month:

TV is still the monster, the elephant: for all the talk of cord-cutting, Americans have clearly voted that, given the choice, they’d much rather have cable TV than broadband internet.

And for web-based publishers, the situation is much, much worse even than this chart makes it look. Consider: the number of websites out there is many orders of magnitude greater than the number of TV channels, which means that even as network TV is winning over small cable channels, small cable channels are still in a much better position than just about any website which isn’t called Facebook or Google or Yahoo. Moreover, if you’re running a news site, you’ll be even more sobered to learn that just 2.7% of the time that people spend on the internet is spent on news sites. You think you’re competing against a lot of other news sites to attract advertisers? You don’t know the half of it. In reality, you’re competing against the other 97.3% of websites, and they are competing against TV. It’s a fight you can’t hope to win, especially since non-news websites are so much better at delivering people primed to buy stuff (search) or delivering large numbers of people in narrowly-targeted demographics (Facebook).

The key concept at the heart of Juenger’s fallacy — the thing which Meeker doesn’t seem to understand — is the fact that internet advertising in no way substitutes for TV or print advertising, no matter how often digital ad-sales people bring out their metrics of comparative CPMs.

Felix Salmon is an Audit contributor. He's also the finance blogger for Reuters; this post can also be found at Reuters.com.