What makes you read something? And more importantly, what makes you read the next story on a site?
Perhaps you arrived at a story through search. Or because someone pointed out an interesting piece to you, maybe on Facebook or Twitter. Or you saw an intriguing article on a site you bookmark. OK – but once you’re done, what makes you stay on that site and read something else?
Recommendations. Whether machine- or human-generated, mechanisms for surfacing related content are critical to keep users on site and exploring what’s there. The world is awash in recommendation engines these days: Think Amazon’s and Netflix’s systems for suggesting books or movies you might like. These engines work to a greater and lesser degree, but this post isn’t about how well they work or don’t work. It’s more about how those results are displayed to us.
To a great extent, most recommendations are displayed in a more-or-less linear way – in other words, as a list. Should they be?
Obviously we’re very used to lists, and it’s a great way of quickly navigating a lot of information. But recommendations, by their very nature, aren’t always linear – because the reasons we come to story aren’t linear. Perhaps you’re reading a story about a controversy in the local school board over test scores – are you there because you’re a parent? A teacher? A student? A researcher? That has potentially a huge bearing on what you consider to be a related story.
Now, to some extent a good recommendation engine, in knowing a bit about your profile and your browsing history, might figure out whether you’re a parent worried about getting your child through the next hurdle of tests or a researcher wanting to understand the national debate over testing – and then offer you only suggestions that fit your likely needs.
But there are potentially other ways to let you navigate a world of related content. For example, we might build a simple two-dimensional grid could offer up stories based on how they relate to the issues at hand – for example, pieces that are more national rather than local in nature; more about higher education about elementary school, and so on. The (simple) graphic I came up with shows how a reader could hunt down stories in the quadrant they’re interested in – the top-right one being about national stories on higher education. Perhaps we could build toggles so readers can change the axes to other dimensions of information as well.
Obviously this is easier said than done – for one thing, there’s the problem of coding stories so that we can plot them on such maps. And then there’s finding a way to scale this over the broad universe of stories a news organization might cover. But for a focused site, perhaps that’s doable. Or over a focused area of coverage.
In some ways this is simply a variant on one visual representation we’re already very familiar with – the timeline. But we could tweak that as well, for example, to have stories about national issues higher on the chart and ones about local issues lower on the chart. In the example here, there’s a cluster of stories about national issues a while back, and a spate of local stories more recently.
It’s true this form of navigation may take some getting used to. When the NYT used something similar – nicely – to map out comments on Osama bin Laden’s death, not everyone got it. But it was – and hopefully will be – a smart way to show sentiment and aggregated commentary in the future.
Other sites have experimented with other forms of display. We’re used to the idea of navigating maps, of course, and there have been various attempts to build news maps that show clusters of similar stories. WhoRunsHK was another attempt to allow a more visual exploration of information, just as silobreaker is. And an interesting site, Recorded Future – and more on them in the future – has a timeline that extends into the future as well as into the past.
In a world awash with content, a big part of what information organizations have to do is find ways of helping people better find what they need. We’ve been very focused on the creation of content, and on the tools that allow us to gather, analyze and tell stories. Plus aggregate information. But we should be working just as hard in helping people find their way around the universe of content around them.