When The Washington Post’s redesigned homepage debuted yesterday in a cleaner, sparer, more quickly loading version of the original, it was the final step in an almost two year effort to change the look, feel and user experience of the digital version of the Post. It also signified the completion of a migration to the Post’s publishing system, Arc, and to PageBuilder, its layout engine, both built in-house.
The changes came incrementally, beginning with a very modest redesign of the author pages in mid-2013, the first test for PageBuilder. When that was completed, principal architect Greg Franczyk and his team realized that what they were creating was “a really powerful platform.” They moved on to the article pages at the beginning of 2014.
There were two major reasons for focusing on article pages next. For one, the article pages were how most people interacted with the Washington Post. “The article pages have had upwards of 90 percent of our traffic depending on the day we’re having,” said Franczyk, “and most of that is driven by search and social. The amount of social traffic that we’re getting directly to article pages has increased dramatically.” That effect is compounded by mobile users who comprise 60 percent of WaPo’s audience, according to Comscore statistics cited by Franczyk.
The structure of article pages was another draw. “Article pages are easily templatized, said Franczyk. “They don’t require a lot of production work inside the tool. We wanted to do this in a way that didn’t really require a lot of changes for the user all at once.”
The user experience was important for Franczyk and his team. So important, in fact, that PageBuilder from the start was a collaboration between WaPo’s engineers and its newsroom. “We would not have been able to build the product we have today if it wasn’t for the newsroom,” said Franczyk. “We always build cross functional teams. Having those points of view and then working to come up with a really great product when looking at what our editorial staff was trying to achieve was crucial to our success.”
One of the features Franczyk said his team would not have focused on without the newsroom’s help was contextual editing. With this feature, an author or editor can click into a headline, make changes, and see exactly how the edits would look to readers, without the need for a back and forth ordeal with a preview page. “We would not have prioritized that as a feature if we didn’t see how much our users benefitted,” said Franczyk.
Franczyk said the response of the newsroom to the new system has been excitement mixed with “cautious optimism as we work through making sure that the platform is stable, featureful, and will work for them.” It is a process that won’t have an end point.
“We’re in the process of doing almost a perpetual redesign,” said Franczyk. “What you see here is only the first big step in a long and perpetual process in making sure we’re constantly iterating and constantly building this thing that can compete against others.”
Now that the migration to Arc is complete, Franczyk and his team can focus on doing “really radical things in a much faster way.” It will be a phase of experimentation dubbed “rethink,” a laboratory for ideas that will seek to expand the current capabilities of the system.
One of the ideas Franczyk gave as an example is having editors and producers introduce a number of options into the system at once–say, different headlines for one story or multiple story selections for one particular spot–and have the system decide what will work best for different situations. As with the redesign, engineers and newsroom staff will work together to figure it all out.