Data Solves the Questions Couldn’t Answer

More than 85 million visitors per month sounds like an outstanding website statistic. That is, unless that number has barely changed in five years. That was the situation faced when CEO Neil Vogel joined in April 2013. His mission was to reignite the site’s engagement and growth to keep pace with the thriving Internet.

With more than 3.5 million long-form articles written by 1,000 Experts, is one of the “largest source[s] for Expert content on the Internet.” The site’s content comprises a broad range of service-oriented topics—from bicycle repair to tax planning to travel advice. So, whether people are looking to make, learn, fix, or diagnose something, Vogel says, aims to be the answer key to the Internet’s swarm of queries.

“Very simply, we help people,” he notes, adding that the majority of the site’s traffic is intent- and search-driven. “People aren’t coming to About like they’re going to BuzzFeed. When you’re going to BuzzFeed, you’re browsing an article or looking to be entertained…. When you come to About, you’re looking to do something very, very specific.”

But wasn’t always the people’s choice for finding answers to their burning questions. In fact,’s traffic had been fairly consistent for years.

“It’s been years of going sideways,” Vogel says. “And [with] the Internet, when you’re going sideways, that’s not a good way to be going because the Internet is growing. Think of how many more places you can be online than you could even five years ago.”

There were two main factors that contributed to’s flat traffic: First, the website’s design was extremely outdated and resembled a search engine results page. Second, the company didn’t know how people engaged with its content. “We had lines of code from the ’90s in our code base,” Vogel says, “and there hadn’t been any thoughtful design changes since like 2007 or something—it’s 2014.”

Because of these hurdles, couldn’t expose people to additional content that could answer follow-up questions (and, ultimately, drive more page views). If a person searches “how to make barbecue sauce,” for example, they likely have related questions, such as “how to make barbecue ribs” or “how to grill?”, Vogel explains. “[We need to] help anticipate that next thing even before maybe the user thought of it,” he says.

This outdated code and site design led to other challenges, as well. For instance, because has so many different audiences—ranging from foodies to techies—the website struggled with putting advertisers in front of specific markets.

As a result, Vogel decided to give the site an overhaul when he joined the company. He wanted to use the website’s data to better understand the interconnections between its content, and, in turn, leverage those insights to drive traffic, get readers to engage with related content, and place advertisements in a more relevant context.

A data-powered restoration

The online renovation had three main components: create a content plan, establish a data strategy, and build out the right technology and products to accomplish the former two. 

In terms of content, there was a plethora of quality articles. The company just had to determine how to put them in front of the right audiences. To do this, started by organizing its content into more clearly defined verticals, like “food,” “health,” and “travel.” The company also hired more than 100 experts to write for the verticals. For instance, knowing that its food-related content often competes with articles from, recruited a member of the Condé Nast brand to head its food vertical, Vogel says.

In addition to making editorial hires, brought on a handful of data scientists, ranging from coding experts to linguists, to co-create and execute the site’s data strategy. had a wealth of data it wasn’t using. “We were like hoarders,” Vogel says. Plus, the way stored its data was counterproductive. Jonathan Roberts,’s director of data science, says the company had a filing system full of Excel spreadsheets with baseline data dating back to 2000.

“[It’s] like you’ve got this Ferrari and you’re trading it in for a scooter,” Roberts says. “You can do so much with this [data].”

According to Roberts,’s new data strategy centers around one simple question: What makes people click? By tracking a combination of behavioral data and content metrics (number of words in a link, language, tags), can identify commonalities among the site’s most-clicked-on articles. In a way, the strategy is cyclic: Analyze large quantities of data to develop high-level insights; apply those insights to make it easier for users to ask questions; track and analyze clicks as users find more content; develop more actionable insight; apply them; and so on.

To achieve these goals relies on technology, such as its own natural language processing and predictive analytics tools. In terms of natural language processing, ranks keywords in terms of importance to use them as cues for which related content to recommend. For example, the words “barbecue” and “steak” might rank highly for an article about barbecue sauces. The ranking is based on how these keywords were used in previous searches and where they were used within article text.

Sometimes this initial step is all needs to draw connections between articles, Roberts says. Other times the situation can be more complicated, such as if a word has more than one meaning. For instance, the word “stock” may refer to the financial performance of a company, a form of medieval torture, or a cooking ingredient. Therefore, also has to take into account the context of the keywords to ensure that it’s not linking an article about investing to an article about the Middle Ages or one with a recipe for gravy.

Predictive analytics come into play for activities such as better forecasting rises and falls in traffic and recirculating popular content. Knowing that content about family crafts is popular around Halloween, for instance, can help content managers promote Halloween-related content in October.

“You want to make sure that you’re devoting your resources [to] the places where there is something interesting and surprising happening, rather than going to something and saying, ‘Oh, we hadn’t realized it was Easter today,’ which is why that content is trending,” Roberts says. also abandoned its Excel spreadsheets and now has a “clean” and “modern” database structure internally, Roberts says.

Relevance is about context

Not only does produce content, but it also provides a place for advertisers to contextualize their messaging. “Brands care what they’re sitting next to,” Vogel says. “You have to make a beautiful place.”

Using the company’s proprietary tools, analyzes all of the sections of its site to identify the types of content that would be the most relevant for an advertiser. Again, has to consider both the text and the context of the text when deciding where to place an ad. Brian Colbert,’s chief revenue officer, cites as an example that while the term “wine” may bring up a list of grape varieties in’s food channel, the word may bring up wineries and honeymoon content in the travel channel. Once the right context has been determined, inserts the ad within the content.

Recently, it inserted a video from fast-food chain McDonald’s about its McRib sandwich, a barbecued pork item, in the middle of an article about grills, barbecues, and smokers.

“Native is really a chance for us to embed the advertising in the body of our content in a way that’s organic and feels, in some ways, very additive to the content for the user,” Colbert says.

To measure the success of the ads, tracks engagement metrics, such as click-throughs and time spent with an ad.

Colbert says that the company ran native ads before the site redesign, but the site’s new layout and organization makes the ads “easier on the eye.” In fact, on average,’s advertising campaigns now drive a 47% lift in awareness for advertisers.

Measurable results—that’s what it’s all about

After officially launching its redesigned site this past September, saw a more than 10% increase in time on site among users. Roberts says the redesign has also helped the company adopt a more data-centric culture.

“It generated this feeling inside of the company that the data isn’t scary,” he says. “It’s something that can drive a lot of value, and it’s something that can be explored easily.”

But isn’t finished with its renovation just yet. According to Colbert, the company would like more people to visit the site without necessarily relying on searches to get them there. “We’re always going to rank high in search because of our high-quality content,” Colbert says. “We’ve seen rapid growth in social and in our in-house emails. We’re constantly looking for new audiences.”

He also admits that has focused more on desktop visits than on mobile ones—and that focus needs to shift because 35% of its traffic comes from mobile. So, although is a mobile-responsive site, Colbert says that the company will be enhancing elements of its mobile experience, such as mobile navigation, in the first quarter of 2015.

Using data to continuously optimize? That’s what success is all about.

Related Posts