Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Episode #49: Brian Boyer
139 Plays9 years ago
This week’s episode was recorded live at NPR in Northeast Washington, DC with Supervising Senior Editor of the NPR Visuals team, Brian Boyer. No, I didn’t get to sit in one of the fancy chairs, but I did get to...
The post Episode #49: Brian Boyer appeared first on PolicyViz.
Recommended
Transcript
Introduction to MICA's Online Program
00:00:00
Speaker
This episode of the PolicyViz podcast is brought to you by the Maryland Institute College of Art. MICA's professional graduate program in information visualization trains designers and analysts to translate data into compelling visual narratives, benefit from the resources of a premier College of Art and Design while learning online.
00:00:20
Speaker
Earn your information visualization degree in just 15 months. Expert faculty includes Andy Kirk, John Schwabisch, Marissa Peacock, and Rob Rolleston. Learn more at mica.edu slash MPSInvis.
Meet Brian Boyer at NPR
00:00:46
Speaker
Welcome back to the Policy This Podcast. I'm your host, John Schwabisch. I'm sitting here at NPR in northeast DC. Just barely. Just barely with Brian Boyer. Brian, thanks for coming on the show. Happy to be here. Lovely new building. Newish building. Newish. Couple of years old. Couple of years old. So thanks for the tour around. That was a lot of fun. Maybe we can start by having you introduce yourself and the visuals team here at NPR.
00:01:08
Speaker
Sure. So yeah, I'm Brian Boyer. I'm the visuals editor at NPR. We're a sort of ragtag band of misfits where we are photo editors, photographers, video producers, interaction designers, graphics editors, and hackers. So we do picture editing, we make graphics a website, we do interactive design, we make video for Facebook, sort of all aspects of visual journalism here
The Role of NPR's Visuals Team
00:01:34
Speaker
at NPR.
00:01:34
Speaker
And how does that group fit within all the other groups within NPR? Right. There's the individual desks and shows have will frequently have people with sort of similar talents. They'll be you know, the science desk, for example, has people who make video and then images and things like that.
00:01:51
Speaker
So we're often working with folks on those other desks to help them produce stories. But I'd say the stuff that we're most proud of and maybe known for are when we're working really closely with the desk or one of the shows, like Morning Edition, and working with them to tell a story online. That could be something as simple as a great pairing of graphics and text and a normal layout of a story on our website.
00:02:20
Speaker
or something that's pretty outlandish, sort of breaking the format, telling a story with a completely bespoke interface that you might need. But the idea really being that we want to borrow a phrase, we want to let the form of the piece follow the function. So we're never really starting with a format. And that's really maybe the cool thing about having a team of people with so many different skills is that we can sort of approach the problem with a group of three or four people and figure out how it's going to manifest depending on the story.
Collaboration Across NPR Desks
00:02:48
Speaker
And are you driving those conversations with other desks? Are they coming to you? And when you're pulling that together, how does that work as you try to build the team sort of across the different groups?
00:02:57
Speaker
I mean, yeah, so the majority of our work is work on series or stories that are part of the newsroom's priorities. So we're covering the National Park Service this summer. We're covering the general election, right? And we're working with the various desks on helping them tell the stories. Occasionally, there's a story that we love so much that we pitch back to a desk. There's a handful of pieces we've pitched to the international desk recently.
00:03:25
Speaker
So it can kind of come from any which way. There isn't a strict formula for the exact kind of work we do. A lot of our most successful stuff has been with the books and arts teams, which is maybe a little unexpected, but the audience loves that stuff.
00:03:42
Speaker
So let me back up just a second to ask, how does a visuals team fit within an organization that's identified as audio only? How does that work from sort of management and the identification of the organization? Honestly, it's not always easy.
00:03:58
Speaker
This is National Public Radio. I have such tremendous respect for my peers because they can paint a picture with their voice and with ambient sound. They're experts at their craft. Unfortunately, that doesn't always turn into a good story to look at. And so I'd say there's a decent amount of work that a large portion of work at NPR puts on the web that is sort of radio stories that are sort of translated or produced back for a text-based experience.
Web-First Stories and NPR's Evolution
00:04:27
Speaker
And some of those are great and some of those are okay. But I feel like the work that really works best on the web and the work that our team is most proud of and the most frequently the most satisfying or easy to do is the stuff that is conceived initially for the web. And that frequently depends on the capacity of the various desks and shows. For example, we've done a number of pieces with the education team about school funding that were really truly conceived for the web first and show it.
00:04:54
Speaker
Yeah, they really sing that way. Right. And then that team takes that on the air and does sort of a different angle or a different piece of that? Yeah, again, it depends. I mean, sometimes we'll produce, for example, a lot of the work of the Code Switch team is produced for the web first. This is our race, ethnicity, and culture desk. They'll be writing for the web, and then they'll go on the radio after they've written it.
00:05:17
Speaker
That's probably the opposite of the usual pattern, but it's a pattern that more and more we're moving towards. At least we know that it's a good one. Can we talk a little bit about the analytics that you do? Because I feel like NPR, I'm sure everybody does analytics on their stories.
00:05:34
Speaker
but I feel like NPR is one of the few places that publishes a lot publicly about what they did and what they saw. So can you talk a little bit about A, how you go through that and B, how and why you feel like you should be out there publicly sort of assessing that.
00:05:48
Speaker
There is a pattern that I've seen in the news industry is that teams like ours tend to start doing analytics like this. The interactive news team in the New York Times is a great example. We have the unusual combination of
00:06:05
Speaker
a strong interest in storytelling and a strong interest in knowing of our storytelling work. But also paired with the ability to actually analyze it ourselves. So we're not relying on another team to do analytics for us. We can do our own analytics on the visuals team. And so we've sort of had the ethos of conducting our work experimentally. So frequently we'll publish a story and we don't know
00:06:30
Speaker
if the button should be an arrow or the button should say next. And that's something that instead of sort of arguing about or having sort of intuitive ideas about, we can just test it. And since we're building the interface, you know, soup to nuts, implementing a test like that's not difficult for us, whereas it's much more difficult to run tests like that inside your CMS or inside your app.
Exploring the Carebot Project
00:06:57
Speaker
I mean, we're working on these like self-contained chunks of designing code. So at one, it's easier, a little easier for us and we have the tools. And so, you know, I mean, I think being public about it is more, you know, there's certainly no one telling us to do that. Luckily, no one's telling me not to do that either. So I think we've got institutional support.
00:07:20
Speaker
there, but the thing is that I think we firmly believe and I think it's probably partially because we're public media but also just because we love journalism and this is a mission driven job that we're all in this together. If I can help make somebody else's stories better, either by open sourcing the code we wrote to build a story, talking about how our analytics work,
00:07:45
Speaker
then that makes the world a little bit of a better place. That's kind of what gets me going. One of our team motto is work in public and that's really about not just working in public media but doing our work in the public's interest.
00:08:00
Speaker
But it's also a little bit beyond the simple analytics, like you have the Carebot project, right? That's sort of more assessing how people respond, I guess, sort of, innately to the stories that you tell. Yeah, I mean, Carebot, you know, we've been doing Carebot-like things for a little while, and the idea of the Carebot is started with the question of, okay, so we're like a photo and graphics team at a radio organization. That's our job.
00:08:24
Speaker
What should we be doing? And that led us down the road to think, OK, well, I think our job is to create more empathy in the world, is to make people give a shit. Because our story is, I mean, the thing about NPR is we're not ProPublica. And we're not the Chicago Tribune, right? We're not, you know, Chicago Tribune Impact is throwing a governor into jail.
00:08:45
Speaker
The public impact is getting a bill passed in Congress. We do some accountability journals about our bread and butter is introducing people, introducing you to people you've never met. It's making you care about people or places or stories that you would have never heard unless you spent time with us. We said, okay, so if our mission is to make people care, how do we measure for accomplishing our mission? That's a hard question to answer because page views, the traditional metrics do not really tell you that.
00:09:13
Speaker
Unfortunately, we can't put our audience into an MRI machine and we can't actually analyze their brain, but we can do things. We can just ask them. When they finish a story, we can ask the question, did the story make you care? People click a yes or no question like that. Then we've done a lot more indirect measurements like the number of shares on Facebook divided by the number of pages.
00:09:34
Speaker
So one proportion of people who read the story were moved to tell their friends about it. Or would be doing very granular scroll depth tracking to come up with a better notion of completion rate. How many people finished the story? How far down did they get into it? So all those sort of things ended up becoming a small grant funded project we're calling the Carebot.
00:10:00
Speaker
open source framework that does two things. One, it implements a handful of alternative measurements for success, and we're hoping that will become a much longer list. And then the other thing it does that's really important to me is it's a different way of celebrating that success. So what we care about is partially about the metrics, but it's also really about how do you talk about success on your team.
00:10:23
Speaker
How do you, because like a daily analytics email, I get one of those every day about our website, and it's just a list of stories about celebrities and back pain.
00:10:33
Speaker
And occasionally it's about, there's good journalism there, but it's not really a measurement of journalistic success. And when you get that email every day, even if you believe in your work, even if you believe in your mission, that email can really be a lie down because you'll see your story maybe didn't even get on there. And your story was a really good story. And so the question we implement metrics is, the question you need to ask yourself is, what behavior are we trying to change? What are we trying to get people to do to give them this information?
00:11:03
Speaker
And I want people to be creating more powerful journalism, not necessarily be creating more pages because it's not like we make any money on pages anyway. So the actual implementation of Carebot, there is no dashboard, there is no email newsletter. It's just a chatbot that hangs out in your team's Slack room and says, hey, I've been tracking the story for four hours and people scroll down this far or people completed it at this rate or
00:11:30
Speaker
People spent 60 seconds looking at your graphic. And then we say, hey, good job, teammates. And doing that in a public space, like a Slack room, is better than doing it in a private space, like a dashboard. So we've learned a lot in the last few months doing this project. Interesting. And it seems like you've learned a lot in terms of the types of products that you're putting out there. You were doing a lot of elections work, obviously. There's a lot of elections dashboard, which I'm sure everybody's going to be doing now.
00:11:59
Speaker
then it seems like it changed a little bit is that is that right little more maybe a little more narrative you have photo and photo editors and video now so how have you seen that evolution over the last months or years yeah i mean so we started sort of measuring our work in different ways a year and a half ago years ago and
00:12:16
Speaker
And what we found was rather striking that, you know, for example, we have a series of stories where we call sort of look at this. And look at the stories are usually sequential visual narratives or for lack of a better term, they're slideshows, right?
00:12:32
Speaker
They're very nice, they're lovely slide shows, but it's a story that it's pictures paired with a little bit of text and or a graphic that sort of changes as you click through it and you sort of tap, tap, tap and you're reading the story and it's more like flipping through a picture book. And we've produced stories like this that are 70 or 80 slides long that take you a good deal of time to read and people finish them.
00:12:59
Speaker
Yeah, it's really, really interesting. And so our sort of joke about these stories is once we start to measure them, we realize that this is, we are tricking people into reading long-form journalism with lovely photography and graphics. Like this is, these completion rates can be astounding. And, you know, my favorite example recently is a piece we did about the civil war in Yemen.
00:13:19
Speaker
And it's a difficult, sad story. About 60,000 people saw the story. Like, page view-wise. And that's not, you know, a dog. That's not a terrible, you know, sort of reach. That's not a great reach. We will do half a million page view stories.
00:13:35
Speaker
And so if you looked at it just by those numbers you might be let down but you look at it we use a metric called engaged completion rate. So what that is is of the people who stepped inside how many finished. So a lot when you hit a story a lot of people will show up and just bail immediately.
00:13:53
Speaker
So we throw them out for this metric. And so when they start reading, when they click the begin button, or when they start scrolling down, then we start counting. And we see how many people finished the piece. And so 60,000 people showed up, 50,000 people came in the front door, and of those 50,000, 70% finished reading this story. And it's a story, like I said, about the Civil War in Yemen, about a country you probably can't place on a map, about people that you didn't even know cared about.
00:14:20
Speaker
Did you even know there was a civil war in Yemen? It's terrible. By that measure, we really succeeded. That's what I want to be celebrating on our team. It's interesting how we've started it. I think a lot of people started to shift away from page views, but trying to figure out what is the new metric. What is success? You also talked earlier a little bit about
00:14:42
Speaker
you know, making the world a better place, especially for journalists. You have a new project called Alex.
The Alex Project and Election Data
00:14:47
Speaker
Yeah, Alex. Alex, that's sort of multi news agency, right? Yeah. So, Alex, my teammates, David Eads and former teammate, Jeremy Bowers, who's currently in the New York Times, they're leading it. And the idea is that we need to end the elections arms race to borrow Jeremy's phrase, that everybody at
00:15:10
Speaker
Every major news organization is getting election data from the Associated Press for national races. What we all do is we all end up building software just to consume that API or that feed and put in the database.
00:15:28
Speaker
And the fact that we're all doing the exact same work is absurd. And we've done the exact same work. So this project was inspired by the LA Times actually put out a great Python library for consuming the Associated Press's elections FTP drop, which is what they did for it.
00:15:47
Speaker
That's how AP used to give data as they drop it on an FTP site, like a giant tab-separated file or something like that. When the AP decided to change how they deliver results, now they deliver over an API, somebody needed to build in the library. We said, well, why don't we just do this together? At this point, people from six different news organizations have contributed, and this library, Alex, is powering
00:16:12
Speaker
Depending on how you measure it, dozens of election results websites, the New York Times, MBAR, gally, I'm starting to forget the list. Small places like the Spokesman Review, all of them are clatchy. That's one user organization or 29 depending on your view. It's being used by a good deal of people. I'm just proud that we all didn't have to build our own stuff.
00:16:37
Speaker
Yeah. There was just no point in us each building our own. Now, do you feel though that now everybody's going to have the same day that there's still going to be this? I mean, arms race is a great term. This is the arms race of all the different dashboards. I'll move to the next level. This is the joy of software. I've been in the business for a while.
00:16:54
Speaker
every couple of years or every year something new comes out that adds a layer of abstraction that takes away a chunk of plumbing. And your gut reaction is, oh crap, I'm going to lose my job. I was really good at making that plumbing. Whereas really, it's like you're just taking one step up Maslow's pyramid. You're saying, all right, I'm one step closer to self-actualization. Now I can think about slightly more interesting problems. I can think a little bit more about my audience and a little bit less about plumbing today.
Visual Storytelling in Breaking News
00:17:22
Speaker
So, okay, so let's finish up. Where do you see the visuals desk at NPR growing over the next? I mean, we know where we'll be over the next six months. We'll all be talking about the election for six months. But where, you know, one year, five years down the road, where do you see the desk, especially as it relates to the rest of the organization? Yeah. So, one thing that I've been working on with the team is thinking about how we can
00:17:48
Speaker
improve our workflow to work on shorter-term things, which is a little, I'll explain it. You know, we've got a pretty good process for a six-week project, like the Elections app, for a three or four-week storytelling project that combines pictures and video and graphics.
00:18:07
Speaker
stuff. So we're pretty good at projects and I'd say that our shorter term work has been good but what we haven't done as well is really have picture editors and graphics editors and data folks working really closely for like a two-day project or a one-week project or breaking news situation. There's a story we did recently about the bombing in Brussels that I'm really proud of because it was
00:18:38
Speaker
You know, instead of, you know, what we could have done that morning is started to make a bunch of locator maps of the city of Brussels. But for an American audience, I don't think that adds a whole lot of value. But that was the default behavior. Make a map and there you go, right? And we said, OK, let's not do that default thing. Let's get an editor from the international desk, one of our picture editors, one of our graphics editors, me, you know, a couple of people. Let's sit down for 10 minutes. Let's sketch out a plan for the day.
00:19:08
Speaker
And let's plan on publishing something four hours from now, five hours from now.
00:19:13
Speaker
that really tells the story of what we know and tells it as a visual story, as opposed to having visuals as sort of an accessory to a text-driven story. It's not to say that visual resort is better than words, but people really like visual storytelling, like it works for you on the web and on social and wherever. And so telling a story where photography and graphics are the centerpiece more than sort of the side piece.
00:19:42
Speaker
That can be really difficult to do when you have to work really fast. What we're trying to figure out is what's the breaking news checklist or what's our little playbook for a one-day story, a two-day story, a six-day story, and how do they vary. That's something I think we're in the process of learning right now. Interesting. Well, good luck. Thank you. Sounds great. Lots of great stories and I'll post them all on the show page.
Closing Thoughts and Listener Engagement
00:20:05
Speaker
Brian, thanks for coming on the show. Absolutely. It's been fun.
00:20:07
Speaker
And thanks to everyone for tuning in this week. Please let me know if you have comments or suggestions on the website, on Twitter, via email, and please rate the show on your favorite podcast provider. So until next time, this has been the PolicyViz Podcast. Thanks so much for listening.
00:20:34
Speaker
This episode of the PolicyViz podcast is brought to you by the Maryland Institute College of Art. MICA's professional graduate program in information visualization trains designers and analysts to translate data into compelling visual narratives, benefit from the resources of a premier College of Art and Design while learning online.
00:20:54
Speaker
Earn your information visualization degree in just 15 months. Expert faculty includes Andy Kirk, John Schwabisch, Marissa Peacock, and Rob Rolleston. Learn more at mica.edu slash MPSInvis.