Back in November, I grappled with the fact that online display ads in general, and banner ads in particular, are clearly not working very well; my suggested alternative was for brand advertisers to embrace the power of the external link. That was one suggestion; there are many, many more. But what they all have in common is that they’re attempts to go beyond the ad, and to leverage the interactive power of the internet.

Over at Tumblr, David Karp is being characteristically vague about what he’s offering to potential advertisers: all we know for the time being is that he “wants brands and marketers to use Tumblr as a way to tell stories that they can’t otherwise tell on other social networks”. Which sounds great, but doesn’t even come close to answering the obvious first question, which is “how?“. I understand that the idea is to sell space on the right hand side of the screen, and that clicking on one of those units will take Tumblr users to the advertiser’s tumblog. But this seems uncomfortably close to the idea that advertisers buy a banner ad and that clicking on that banner ad will take users to the advertiser’s website. The tumblog itself might well tell a story — but then again, so might the advertiser’s website. The difficult thing is getting users to click on things, especially when those things look like — and are clearly labeled as — ads.

Similarly, Facebook’s revenue problems are based on much the same underlying issue: Facebook itself is highly interactive and immersive, but the ads you find there are not. And while there are one or two companies I will follow on Facebook, they’re invariably companies which are run by my friends. Facebook is a great place to keep up with what your friends are up to, but it still hasn’t cracked the nut of working out how to make itself valuable to brand advertisers.

Now, Gawker Media’s Nick Denton has a new idea:

In an internal memo on Thursday, Denton announced the formation of a new sales unit that will focus on helping advertisers and brands take part in the new commenting system

According to the memo, Gawker is creating a new content unit within the sales department that will be headed by Ray Wert, formerly editor of the Gawker-owned automotive blog Jalopnik. This new unit will take over responsibility for all of Gawker’s branded content functions, as well as marketing communications and events — and the purpose of the unit will be to promote the new Gawker discussion platform as a way for marketers and brands to engage with customers in an open forum. Says Denton:

We all know the conventional wisdom: the days of the banner advertisement are numbered. In two years, our primary offering to marketers will be our discussion platform.

Last Friday, Denton gave me, along with a few other New York digital-media types, a preview of his new commenting system; yesterday, I had a pretty geeky conversation with Wert about how he intends to turn it into dollars.

At Gawker, as at most other popular sites, the number of people reading the comments is vastly greater than the number of people writing them. But the way they’re presented, they’re not easy to read, there’s far to many of them, and the signal-to-noise ratio tends to be extremely low.

So Gawker’s new commenting system is based around threads, with the default view being the main, most interesting thread. It’s possible to click through to other threads, and every thread — indeed, every comment — has its own unique URL; what’s more, the person who starts a thread has quite a lot of control over which comments in that thread will get featured.

What that means is that if an advertiser buys a sponsored post — and sponsored posts have been part of Gawker’s menu of offerings for some time now — then once the new commenting system is in place, the advertiser will have a reasonably large degree of control of the conversation that most people see in that post.

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