At the Canadian Broadcasting Corporation (CBC), we write about 400 stories a day. Our lineup editors then publish these stories into about 40 different regional, national, and beat lineups.
Like many publishers, our strategy has been shifting from emphasising reach to emphasising engagement. We’ve been moving from reach-based metrics (like visits, unique visitors, and pageviews) to engagement-based metrics (like time spent on a story, session depth, or session quality).
But with our volume of content publishing, it’s often been difficult to get a clear snapshot of which stories are getting the most reach and engagement somewhere in our ecosystem.
We were often left wondering whether we were overlooking diamonds in the rough.
We wondered about stories that were out there, maybe local stories or smaller stories, that were really engaging our audiences but weren’t reaching a wide audience. Maybe we’d missed them because we didn’t promote them in a lineup, maybe because they hadn’t caught on through social media or didn’t show up in our most popular lists according to visits or pageviews.
Of course, we have all the standard metrics and tools to rank stories. We can easily generate lists according to pageviews or unique visits or time spent. We can scan Chartbeat and lineups to see what’s performing at any given moment. But the problem with these approaches is that they often only give you half the picture. They are great at ranking stories, but they aren’t as effective at showing the engagement in the story itself.
The lists were also biased by our own promotion. Some stories got a high reach just because we put it in a lineup, not because they were inherently popular stories.
There was something missing from our approach. We often felt we were missing and not promoting some of the stories the audience might love. So we developed an internal tool to help solve this problem.
We named the tool CONRAD, a portmanteau of content and radar. The idea was that CONRAD would work like a real radar, scanning the ecosystem for highly engaging stories that the audience reacted to but that we missed or overlooked because they weren’t getting a lot of pageviews or they hadn’t had a chance to get traction in a lineup.
The tool is intended to make it easy to identify stories that are highly engaging but haven’t reached their full potential.
“A local story that has very high engagement but low reach could well be repositioned for a national lineup that could expose it to, and engage, a much larger audience,” said Quin Parker, one of the analysts who worked on CONRAD.
Here’s how it works: CONRAD ingests a bunch of reach and engagement metrics. Then it ranks the stories according to their relative percentile performance. The result is that all the stories end up in a quadrant, scored according to their ranking of both reach and engagement.
“If we can move from a narrow perspective of just reach, to reach plus engagement, we’ll be able to showcase more of our journalist’s work,” Parker added.
“The most striking and effective part of CONRAD is that it gives you both: It gives a more complete picture of a story performance from not just clicks and reach perspective, but also an engagement perspective,” said David Freeman, who runs the home page lineup.
Blending engagement metrics into our rankings also allowed us to control for the impact of our own promotional bias — the reach impact from us placing in a lineup. “We are taking the effect of the home page placement, and promotional bias, out of the equation,” said Jason Kim, the lead manager who built CONRAD.
The introduction of the tool has also led to other benefits.
“Having worked in news at both a network and regional level, until some of these tools (existed), the relationship between the main desk and a beat desk or a regional desk was largely based on a pitch process, where you had to sell your story,” Freeman said.
“Now with a tool like this, we’ve got tangible evidence that shows that this is a story worth paying attention to, so that the relationship between the main desk and the regional desks can evolve. That will allow all the editors at all desks to react more quickly and be more nimble to help engage their audiences.”
“These kinds of systems are something that forward-thinking news organisations need to think about. We need to move away from the pageview paradigm,” Kim said. “By building engagement into how we measure the performance of content, it shifts the conversation from getting stories on the home page to getting engagement on the story itself.”