Edge Ideas

The day that text-based paid links were combined with algorithmic search results was one of the most important days in the history of the Internet. But what was even more important is what came in the days following that brilliant synthesis: nothing.

In those early days, the wise men and women of the Internet ad world waited expectantly for what would surely be phase two: graphical ads next to search results. Maybe they’d be mini-banners. And then maybe they’d be animated.

We waited and waited. And it never happened. Search ads remained text ads.

It turns out that text-based paid links work just fine without images and animation. In fact, they work better than any ads anyone has ever seen on the Internet.

* * *

Text-based paid links work so well not despite their lack of banner stylings — but because of them. By focusing on a snippet of plain text and a single link, search engines stumbled onto two principles that are essential for successful advertising today.

The first principle is that the advertising must match the purpose and substance of the content. It must directly satisfy the reason you are consuming the content and be as close as possible to the nature of that content. Movie trailers do this. And glossy fashion magazine ads do this. And of course paid search links do this perfectly.

The second principle, which is especially relevant to digital spaces like the Internet, is that the cost to make the ads must be low enough that advertisers can freely experiment with creative. This enables poorly performing ads to be killed and replaced without regret, inspiring a kind of Darwinian creativity of search engine marketing (or “SEM”) creative that pushes the most valuable ads to the top of the heap.

Once the success of paid search became obvious, Google developed a version of the service called AdSense, which matches text-based paid links to content instead of search results. For many sites, particularly those in commercial categories, this AdSense unit became a valuable part of their monetization approach. If you were on a page explaining how to select a cell phone, the AdSense links to genuinely great deals on handsets or services were a consistent and valuable part of your experience.

But for news-focused sites, text-based paid links never achieved the same financial significance. If you are reading an article about how cell phones may give you brain cancer (they don’t, by the way), the paid links to cell phone advertisers are at best irrelevant and at worst annoying. Given the choice, news sites always prefer to sell bigger and better banners.

The reason is that, when you’re reading news, you aren’t looking for links to deals or companies. You’re looking for recent and interesting updates or opinions from highly relevant sources. (You’re also not looking for bigger and better banners, but advertisers have convinced themselves that at least you can’t help notice them.) The elegant platform for monetizing news reading would charge marketers to provide recent and interesting updates and opinions on the news. And, of course, creating these updates would be inexpensive enough that marketers could quickly learn and be rewarded for better product.

* * *

About two months ago, at their Chirp conference, Twitter COO Dick Costolo announced his company’s new “monetization platform.” He didn’t explain it this way, but Promoted Tweets are the first plausible attempt to build an advertising system for news on the Internet.

First, Promoted Tweets is tweets, just like paid search is links. They will be created as tweets, have the same functionality as tweets, be presented in a stream with tweets (though clearly disclosed as paid), and exist for the same purpose as tweets: to notify of news or opinion.

Second, creating tweets and seeing whether they take off — through a engagement metric called “resonance,” which Twitter will provide — costs very little. People with backgrounds in public relations and communications (i.e., people who understand what makes news interesting) will quickly learn what kind of updates work best on what kinds of stories. Just like SEM, these marketers will battle it out to win the top spot, with the benefit going to the user, who will be shown the demonstrably best content.

From a personal perspective, I am certain these tweet-ads will be highly successful. My beloved UberTwitter already incorporates a foreshadowing of the new Promoted Tweets system through partnership with ad networks 140 Proof and One Riot. Despite the fact that their limited inventory prevents meaningful targeting to my interests or to the rest of the tweets in the stream, the facts that these ads are newsy and right in the stream compel me to read them. In just a few months, I’ve clicked on far more of them than I have clicked on banners in the past ten years.

* * *

This is an opportunity for news-centric sites, but there are two challenges.

First, this business model will not succeed if long-form articles are the primary unit of news consumption. For an ad to match the purpose and substance of a site based on articles, the ad would need to be an article. A good article, ideally. This is being tried, by the way, on more adventurous news sites now, like Alleyinsider’s Sponsored Posts. But articles are expensive to create, slow to test, and embarrassing to revise. As with text-based paid links in search, the ad chunk needs to be smaller, cheaper, and amenable to iteration.

Can news organizations adopt a smaller unit of content? A new news atom much smaller than the article? I think there are strong reasons to do so, especially with evidence of soaring mobile (i.e., small screen) news consumption. And the micro-chunk and the article aren’t mutually exclusive. One can build to and reinforce the other. But this is a question of the core user experience, and traditional news organizations are not likely to demote the article to a secondary status.

Second, the ad needs to resemble the content and be consumed in the flow of the user experience. It must be clearly disclosed as paid, and it can be adjacent to rather than mixed in with the editorial content. But, again, as with text-based paid links in search, the eye must pass over the ad in the natural course of its journey through the user interface.

Can news organizations give a useful purpose to their advertisers, as Twitter’s Promoted Tweets will propose? The data will show that users prefer tested and successful ads that are relevant and demonstrably readable to ads that are unrelated and ignorable.

Another important day in the business history of the Internet is almost here. Twitter is about to roll out the first serious attempt at an ad system specifically designed for news consumption. Ironically, perhaps, it won’t work well for most existing news content or news organizations. Can they adapt? Maybe. But if they don’t, new news organizations will emerge to seize the revenue opportunity.