AOL's Latest Dumb Business Plan
Write stories based on search terms, mimic universe's worst news site.
During the last few weeks, AOL CEO Tim Armstrong has unveiled a series of plans to save the long-suffering company from its lurch into irrelevance. First, massive layoffs—Armstrong plans to let go of 2,500 employees, a third of the firm's staff. The company is also getting a new "brand identity." The new AOL logo renders the company's name as if it were a pronounceable word—it is now Aol., with the period. Armstrong is apparently undaunted by the fact that this new word sounds like a-hole without the h.
In an interview with the Wall Street Journal this week, Armstrong revealed his latest scheme to return AOL to its former glory: turn the company into the Web's premier destination for "content." Of course, that sounds like what a lot of Web companies—from Yahoo to YouTube—are trying to do, but Armstrong believes he's hit on a more efficient and accurate way to produce stuff that people are interested in. The new AOL will peer into the world's most revealing repository of human desire, our search keywords. And then the company will generate exactly what it thinks we want.
Here's how the plan will work: At the center of AOL's nearly fully automated content assembly line will sit a vast number-crunching operation that will tap into data from search engines and other sites to predict the stories, videos, and photos that people are most likely to click on at any given instant. As I write this on Tuesday, AOL might identify Tiger Woods' rumored mistresses as a hot topic, or perhaps Solange Magnano, the former Miss Argentina who died of complications arising from cosmetic surgery on her butt.
Next, AOL editors will try to satisfy our desires by putting out a call for entries on Seed.com, a not-yet-extant site where freelancers can submit content. Submissions will be automatically scanned for plagiarism, obscenities, and grammar errors and then quickly edited and fact-checked by a human being. And voilà—a matter of minutes after a beauty queen has died or a golfer hits a tree, AOL hopes, it will have blog posts, stories, essays, photos, and videos at the ready.
This might not immediately sound like a bad idea. Old-school journalists are skeptical of producing news by the numbers—under the high-minded assumption that reporters should focus on what's important regardless of whether people are interested in it—but most news organizations today look at Nielsen ratings, Web traffic, and search trends as part of their day-to-day operations. Some sites go even further. The Huffington Post monitors real-time click data to sharpen its headlines and story placement. Meanwhile, writers at Gawker (and at least one New York Times staffer) get bonuses for writing posts that garner a lot of traffic.
The trouble with AOL's plan, then, isn't that it's based on data-mining. Instead, it's what the company will likely do with search data—publish quick, vapid posts that do little to advance any hot story and instead feed readers a collection of factoids gathered from other places. How do we know this will happen? Because AOL's model is strikingly similar to that of Demand Media and Associated Content, two start-ups that also use search data and user contributions to build Web content. Indeed, AOL's Armstrong—who was an advertising executive at Google until earlier this year—is reportedly an investor in Associated Content, whose CEO is also a former Googler.
Associated Content stands as a cautionary tale for anyone looking to do news by the numbers. It is a wasteland of bad writing, uninformed commentary, and the sort of comically dull recitation of the news you'd get from a second grader. Oh, and here's one more interesting thing about Associated Content—because its stories are bulging with hot search terms, it gets more visitors than just about every news site online, including washingtonpost.com.
For a sense of how bad this sort of thing can get, check out the story that Associated Content published on Monday about the Tiger Woods saga. The piece, written by a frequent Associated Content contributor, carries a headline that reads like a spam e-mail subject line, packed with every possible search term related to the Woods story: "TMZ Tiger 'Kobe Special' Quote Made Before Tiger Woods Mistress Pictures Included Gloria Allred." The lead sounds like a bad translation:
Alleged Tiger Woods mistress pictures of Rachel Uchitel might have spawned the TMZ Tiger quote of having to get a "Kobe Special" from Zales. Now that Gloria Allred joined the fray, will it hurt or help Tiger Woods?
The post repeats news originally broken by TMZ, the National Enquirer, People, and several blogs. More important, it uses the phrase "Tiger Woods mistress pictures" at least eight times. Tiger Woods mistress pictures, not coincidentally, was one of the fastest-surging search queries on Monday, according to Google Trends. And Associated Content cleaned up—for much of the day, its story was the first Google result for that query.
To be sure, it's possible that AOL will deliver more-polished stories than the ones now found on AC. I wouldn't be too confident, though. AOL plans to mimic some of the worst features of Associated Content. First, it will pay writers only a small amount up front—less than $100 or so—and instead give them bonuses based on the traffic their posts generate. This will encourage writers to mimic AC's Tiger Woods story, spending little time crafting prose or doing original research and instead packing the story with search terms meant to spirit the link to the top of search rankings. In addition, just like Associated Content, AOL will allow advertisers to take part in the process of creating content. As the WSJ reports, "if its algorithms show consumers are searching for information about the Zhu Zhu Pets robotic hamster, a retailer could pay AOL to sponsor an article about where to find the hot toy."
Will this plan do wonders for AOL's bottom line? It very well might, at least in the short run. If AOL can replicate the success of Associated Content across its network of sites, it will surely see huge gains in traffic and renewed interest from advertisers. But this plan hinges on something that can't be guaranteed for long—a weakness in search engines. By any measure, stories like those found on AC don't deserve top billing in search results. If you search for "Tiger Woods mistress pictures," you should get pictures of Tiger Woods' alleged mistress, not a story that repeats that phrase a dozen times. Google and other search engines constantly battle search engine spam, and over time they're sure to steer people away from sites that rely on such trickery to get visitors. Then what? Associated Content gets 90 percent of its traffic from search engines. Once Google and co. wise up to AC's schemes, its business model is toast.
There is another model for producing news that people want to read: Hire a team of professionals to cover a beat, and pay them for excellent results. As it happens, AOL knows a thing or two about that technique. It already owns the celeb-news outfit TMZ and the gadget blog Engadget, two of the best news sites online. Both excel at covering subjects that people are looking for—their posts always rank highly on search engines.
TMZ, for one, has owned the Tiger Woods story. The gossip site has repeatedly broken news about the car accident and its aftermath—including the "Kobe Special" tidbit that Associated Content grabbed for its "story." (Another sign that Google is broken: As of Tuesday afternoon, a search for "tmz kobe special" lists the Associated Content item above TMZ's original post.) These are the kind of scoops that readers can't resist and that can't be generated by an army of traffic-hungry, search-term-repeating pseudo-journalists.
TMZ realized early on that it couldn't publish enough Tiger Woods items. But the site also reports on stuff that people don't know they want to read about. TMZ didn't get the scoop about Michael Jackson's death, Chris Brown's domestic abuse, or Mel Gibson's anti-Semitic tirade because it saw that people were searching for those topics. Instead, its staffers are constantly developing sources that blanket the entertainment industry, an effort that generates enormous returns—and wins a loyal readership that keeps coming back for more. There are no shortcuts to that kind of reporting and no algorithms that can tell you what makes for a good story. That's the difference between producing content and producing journalism.
Become a fan of Farhad Manjoo on Facebook.
Farhad Manjoo is Slate's technology columnist and the author of True Enough: Learning To Live in a Post-Fact Society. You can email him at firstname.lastname@example.org and follow him on Twitter.