Piloting digital-first audio is one of the “four pillars” of Project CITRUS. We’ve already written about our first foray in this area — a weather microcast with meteorologist Dave Epstein. But for the last few months, we’ve been working diligently on another microcast experiment that grew out of the coronavirus pandemic.
It’s called Coronavirus, Briefly. And we’ve learned a great deal from making it, every Monday through Friday (and counting) since the end of March.
The concept behind “CVB,” as we’ve come to call it, is far from novel. So-called “pop-up” news products — primarily email newsletters and podcasts — are starting to become commonplace. (They’ve also become something of an industry punching bag.)
Let’s go back to fall 2019 for a moment. President Trump was in the thick of impeachment — and seemingly every major newsroom in America wanted you to subscribe to their podcast about it. It was (mostly) well-intentioned information overload. But it was also a trend that caught our attention.
Good journalism helps people make sense of complicated subjects. And in this age of on-demand, personalized digital news, it’s understandable for news outlets to target spaces like someone’s inbox or the podcast app of their choice and cultivate a daily habit.
These topic-driven, limited-run podcasts and newsletters don’t fit for just any story. Rather, they emerge when we are collectively — as a country, or a local community — immersed in a simmering crisis, trying to comprehend events that feel both existential and without a clear end.
Like a global pandemic.
Getting The Ball Rolling
The idea that eventually became Coronavirus, Briefly started with an early March meeting at WBUR. It was among the last any of us had before we all cleared out of the office. Even then we were witnessing a spike in demand for consistent, trustworthy news from local sources about this once-in-a-century story.
The pitch was: This pandemic is immense in scope, and it’s going to be with us for a long time. So let’s leverage all of the great reporting that our newsroom is going to produce in the weeks and months ahead and experiment with a local, short-form audio product for digital platforms — RSS, the web, voice assistants, etc.
Taking The Temperature Of The Existing Coronavirus Podcast Scene
Before we got the go-ahead to start production on Coronavirus, Briefly, we did a quick competitive analysis to see what COVID-related podcasts were already out there and what, if anything, we could add to the landscape.
We were able to come up with these main takeaways after surveying a number of the top “COVID-casts” across the journalism industry:
- Due to the rapidly updating nature of the story, most of these podcasts were releasing episodes on a daily basis.
- While formats varied greatly, most of the podcasts required very little production (i.e. we saw a mix of: host two-ways, Q&A sessions, headline reads, etc.)
- The favored (favored? By whom? Do you mean most popular?) episode length hovered around 15-20 minutes. We found nothing less than 5 minutes.
- The majority of the podcasts were produced by large networks outside of the public media-sphere.
- At the time of our survey, we could find only one podcast focused primarily on local coverage: KNKX’s (Seattle metro) highly produced, deep-dive weekly podcast “Transmission.”
The way to stand out was obvious: create a locally focused, 3-minute briefing — a coronavirus “microcast.”
Since doing this competitive analysis, dozens of other COVID-related podcasts have hit — and then left — the scene, including several dedicated to local news, such as “Coronavirus Chronicle” from the Houston Chronicle, “Coronavirus in Florida” from the Tampa Bay Times and “Coronavirus in California” from the LA Times (the latter two focusing on a statewide angle). It’s worth noting all three of these podcasts are produced by print newsrooms, while public radio producers have been surprisingly absent from the local coronavirus podcast market.
The Production Process
From the outset, we knew production had to be as light and easily repeatable as possible. We couldn’t burden the newsroom with another project (let alone a daily one) just as the pandemic news tsunami was cresting. We had to own it.
After a brief discussion about daily release time, we decided the evening would be the most workable for our (now completely distributed) team to pull off. The story was also changing so quickly that any podcast released in the morning — but created the night before — ran the risk of being outdated or, worse, no longer accurate.
An early evening release, on the other hand, afforded us a chance to survey the day’s coronavirus news up to that point, and then script, record and release something far more fresh. We pictured the typical listener as someone busy all day with work, child care and/or homeschooling and doesn’t have much free time, and no commute to catch up on the news. But they still want to know the top three coronavirus stories for Greater Boston. So they listen to CVB while exercising or prepping dinner.
The WBUR newsroom’s reporting kept CVB’s lights on. That, coupled with our team’s varied skill sets — and some luck — meant we could operate in an entirely self-sufficient manner.
- Franziska had on-air experience, plus extensive audio production experience — and her own Zoom H5 audio recorder and Audio-Technica AT8035 shotgun microphone, plus a well-sound-proofed closet for recording.
- Dave had the tech know-how to fast-track CVB into our audio workflow, and create the “superplayer” (see below).
- Jack had digital production, writing and mixing chops.
- Our editor — and former On Point executive producer — Karen Shiffman had an extra 20 minutes a day. Karen has decades of experience and her input often made the script more succinct and conversational. She also provided robust fact-checking.
- Jack and Franziska were quarantining with a graphic designer and multi-instrumentalist, respectively. That helped us quickly create bespoke show branding and scoring.
A Rigid Format, But Not Without Tweaks
We recorded four pilots in mid-March which really helped us get the juices flowing.
Initially we experimented with an inverted-pyramid script with an international —> national —> local approach, telling one story from each of those vantage points, in descending order. That was quickly scrapped in favor of leveraging our greatest strength: fantastic local news coverage. It quickly became an obvious choice: Keep the whole script focused on Greater Boston. (Just before CVB launched in full, NPR released its own daily coronavirus podcast. So our decision proved prudent.)
We knew the bulk of each episode would depend on the newsroom’s daily output of stories, which is determined by processes outside our control. But we also knew we could control a constant intro and outro. So we set about fine-tuning those script elements.
Here is the intro for our very first episode:
And the outro version 1.0, which always summarized the state’s daily pandemic data:
Intro and outro completed, we gravitated toward a three-story approach: a “lede” that was generally the day’s top local coronavirus story thread, plus two more supporting stories. Sometimes, we even tried to lighten the mood with a more heartwarming, human-interest angle:
This formula helped us bring a reliable sense of order to the flood of content coming in each afternoon. Once we winnowed down a pool of stories to choose from, all we had to do was script around them and decide on the strongest editorial flow. We sought to find balance between officials like Gov. Charlie Baker — whose daily news conferences weren’t always the liveliest — and more compelling, first-person stories that allowed us to hear from people experiencing the pandemic (health workers, business owners, grocery clerks and family members of older people, for example).
Audio recorded on the phone or via video call has been the norm during the pandemic, with so many interviews happening remotely. Overall, tape quality had a lot of (understandable) variance as reporters and hosts got accustomed to recording from home. But we would occasionally luck out with in-person recordings.
Numbers In, Numbers Out
Over time, we condensed the intro’s “quick disclaimer” to include just the time. Later, we decided, for several reasons, to scrap the data summary in the outro altogether:
- Numbers and data points — let alone that many in succession — don’t translate well to audio generally. It’s a lot to ask the listener to keep up with.
- The state health department burned us a few times by significantly delaying its daily, late-afternoon data release, which usually comes at around 4 p.m., creating production delays and headaches. On a handful of occasions, the state also walked back some numbers and reporting standards retroactively. Those asterisks became a lot to explain in just a few seconds.
- Perhaps most importantly, we were presenting critical health data absent any useful context. Numbers like these are better suited to, say, a web-based data visualization.
Incorporating Spots And Plugging Into The Newsroom
To repurpose, or not to repurpose? As CVB gradually found its footing, we debated the use of broadcast-first audio and its merits in the microcast.
With one host, it’s simple enough to cover a lot of ground without leaning on other reporting sources: write a script that includes three to four key stories, tack on an intro and outro, record it and call it a day. And while that approach certainly works, it lacks texture. Just a host talking, on its own, doesn’t sound as compelling or urgent as quickly “throwing to tape” (i.e., introducing a quick story from a reporter) or, better yet, three different reporter voices, plus their tape of even more voices.
Having decided to repurpose already-broadcast stories, we quickly zeroed in on newscast spots. WBUR produces a handful of these roughly 40-second kernels of content on a daily basis. They wind up nestled in hourly radio newscasts throughout the day. They’re short, snappy and tightly focused — in other words, just what we were looking for.
Leaning into spots played to our local-news strengths, simplified production and helped envelop listeners in the BUR “warm blanket.” And since Franziska and Jack were helping out the newsroom during the surge, they had a bird’s-eye view of all the coronavirus news we were producing. That helped us find the best, most-relevant content — and find it quickly.
One content gambit that didn’t quite work out was creating CVB-first (and possibly CVB-only) content, like a reporter spot or two-way. Our goal was to both produce original audio and create an opportunity for someone to get “on the air” who only rarely appears on the radio. The goal was similar with a custom two-way: Ask around the newsroom to see if there’s an important storyline that could make it to air, but hasn’t. Then see if the reporter on that beat wants to scratch the itch. As it turns out, getting on the air was not as hard as we thought (those hourly newscasts require original reporting!) In addition, if reporters have put the work into a spot, they want it to reach as many people as possible.
For a brief period, we explored potentially outsourcing the script to another department. We had demonstrated feasibility and come up with a workflow. But we were like the proverbial dog that actually caught the car. We dearly wanted to push this beyond the experimental phase, to create a consistent, quality newsroom product. But once we did, it was suddenly ours, with all the accompanying responsibilities.
Knock Those SOQs Off
Once we decided to curate broadcast audio, we never looked back — except to address one thing: the reporter sign-off at the end of a spot, aka the “SOQ.”
Because we named each reporter in the scripted host throw, the sign-off sounded redundant. So we started lopping them off, saving ourselves a few more precious seconds. It marked another departure from the traditional newscast format, which names the reporter in both the throw and sign-off.
This was a small tweak, but one we’ve grown to appreciate (h/t to Karen for flagging it in the first place).
Some Internal Pushback
“How is this different from just listening to the newscast?”
Numerous people asked us this question. And it’s a good one — although a bit anachronistic.
First off, CVB, as we said, was largely designed for people who are busy juggling work, child care, homeschool and so on. They want to stay informed about the pandemic, but they don’t have time to listen to the radio — let alone wait around for a newscast. And newscasts can be incremental and overlapping (e.g., while covering a day-long developing story). Plus, they contain other types of coverage, like weather and sports (in Boston, there is always sports news, even if there are no sports being played). So CVB offered three distinct advantages:
- It’s truly on-demand.
- It’s a curated, coronavirus-only distillation of the newscast.
- It’s short — perhaps blissfully so, for some listeners. (We do the “doomscrolling,” so you don’t have to.)
Unfortunately, some of our best reporting comes in 4- or 5-minute features (the kind you hear on Morning Edition) rather than succinct 40-second spots. What’s a microcast producer to do?
Of course, incorporating a 4-minute feature would completely break our 3-minute format. And it wouldn’t serve our goal of informing listeners on a range of pandemic storylines, not just one. But the more you listen to these long stories, the more you start to discern their different layers. There are small “scenes” within the larger piece. Each scene serves the overarching narrative — but on its own, when set up with the proper context, a single scene can frequently stand alone.
So we started experimenting with lifting 40- to 50-second sections out of these longer pieces. Given their re-engineered nature, we dubbed them “frankenspots.” Here’s an example, crafted out of reporter Adrian Ma’s restaurant reopening story:
“I heard a long pause on that take — was that a cutout? It was probably a cutout.”
“Shoot, my laptop just died. We’re going to have to do this on my phone.”
“Hold on — really loud truck outside.”
Such were the sort of Slack-call quirks we dealt with while recording each episode, on the phone, from three different locations. (At this point, we all pine for a proper studio, where giving your host real-time feedback is as simple as pushing the talkback buttons on the board.)
Creating CVB from three separate remote locations was far from ideal. If you, dear reader, have also been ad-libbing audio production for the past four months, none of the hurdles we encountered should be surprising: dogs, noisy vehicles, impromptu fireworks, bad Wi-Fi connections.
However, the process afforded us a chance to figure out a workflow from scratch, which was both fun and, at times, frustrating. We created contingency plans (“Can only two of us still create an episode?”) and, more important, actually tested them out. This allowed us to occasionally give each other a day off, or quickly switch roles when, say, the roofers at Franziska’s building decided to put in an especially long day.
Now, more than five months in, we’re in the groove. Here’s how recording usually works:
- In the morning, one of us shares a Google Doc in our team Slack channel. It’s our working script for that day. By 4:30 p.m. or so, a finalized/edited version is set and we’re ready to record.
- We collect audio (spots, segments, etc.) throughout the day via a remote server that provides access to all the shared drives we can quickly sort through on work machines. Then, those .wavs are placed into a shared Google Drive folder. It’s a simple enough way for us to all listen to the files as well as share what we’re choosing from with our editor.
- When the script is solidified after a couple rounds of edits, Franziska gets set up in “the studio.”
- Dave dials everyone up on a Slack call. When we’re good to go, Franziska mutes Slack on her laptop, and starts recording. Dave and Jack can hear still the tracking, and Franziska is free from any fumbling or noise on their lines. Though as we mentioned above, they can’t provide real-time feedback (e.g., “Hold up. We need a pronouncer for that reporter’s name.”)
- Franziska does a take, gets notes on it and rerecords anything that needs another try. When everyone’s in agreement, we hang up and either Franziska or Jack get cutting.
- The episode is mixed and bounced down to a .wav. That undergoes one final pass before being dropped into our audio processing system. A few minutes and one cache clear later, it’s in RSS.
Distribution And Promotion
As with any CITRUS production, the goal is maximum distribution across any and all channels.
The newsroom rolled out several new coronavirus assets as the crisis was picking up, from a dedicated landing page on our website, to a coronavirus newsletter (which rapidly gained thousands of subscribers), to comprehensive data viz, to livestreamed town hall events. Naturally, CVB dovetailed with these coverage efforts and served as another prong in WBUR’s multiplatform approach to telling the pandemic story.
For example, CVB was plugged in the coronavirus newsletter, and we embedded a custom-built inline player in select coronavirus stories on WBUR.org. As was the case with Morning Conditions, we built a custom “About” page with subscribe links and other ways to listen.
Finally, we also hope to launch a house ad campaign promoting CVB on WBUR.org.
We did not have individual episode pages, or even show notes. Because we were pulling information from existing web articles and radio content to create each script, and given each episode’s short shelf life, it would have been redundant and difficult to justify.
A Boost From Platforms
Spotify’s “Your Daily Drive” and Google’s “Your News Update” both gave Coronavirus, Briefly a huge promotional lift. Our download figures grew tenfold when the microcast got consistent placement on those platforms. (Full disclosure: Project CITRUS is funded in part by the Google News Initiative, and WBUR is a content partner with “Your News Update.”) In both cases, geo-targeting helped serve the microcast to users in the Boston designated market area. (We also wrote a small Python script that pushes each day’s episode to NPR One.)
All this being said, Google and Spotify were just a slice (albeit a large one) of our wider distribution strategy.
WBUR.org has a slick little home-brewed audio player. It can be embedded anywhere into any story, to play any audio — whether it be an 11-minute feature or a 5-second clip of a bird chirping. There are a few shortcomings however: It will only play a single, predetermined track and it has neither show branding nor subscribe links.
So we built a quick abstraction layer to add those features. The new “superplayer” will now display show album art and ways to subscribe, and always play the latest episode or feature from a given show, podcast, vertical or microcast. This allows us to better promote the actual microcast (rather than just a particular piece of sound from it), and ensures that listeners hear only the latest episode (without us having to constantly fiddle with it).
Finally A Flash Briefing
Once our RSS feed was up and running, we submitted a Coronavirus, Briefly flash briefing skill to Amazon for approval. But for weeks, we hit a brick wall.
Amazon initially told us in a very automated-sounding email that “in an effort to ensure accuracy and consistency of information shared with our customers, we are limiting skills about COVID-19,” before finally budging. CVB has now been in the skills store since early June. (We’ve written at length about flash-briefing frustrations, and proposed some solutions.)
Takeaways And Lessons Learned
- Making new audio doesn’t have to be hard. Coronavirus, Briefly was (we’re pretty sure) the quickest launch of a new audio product in WBUR history. What’s more, we were able to pull it off remotely, during a pandemic (aided enormously by the newsroom). If we can do it under some of the worst circumstances any of us have experienced, what else can we collectively achieve?
- Move fast. Really fast. As quick as our launch was, we still dawdled. We started batting around the idea of CVB as early as mid- to late-February. But we didn’t start on the paper prototypes until early March. And it was another three weeks before we narrowed down our approach, got the project greenlit and started in earnest. The news, needless to say, moves much faster than that. Though, hopefully CVB can provide us — and hopefully others — with a template to push out microcasts at a moment’s notice.
- Consider our “new normal.” The pandemic has drastically — perhaps permanently — changed the way people consume audio news. At minimum, it’s hasted disruption that was already underway in radio. Among remote employees who listen to “spoken word audio” while working from home, nearly half (44%) indicated they listen to local “news/current event podcasts,” according to a Nielsen survey — a total higher than those for talk and music radio. The same survey also highlighted the growing penetration of voice assistants and smart speakers. So in addition to tightening up your FM stream on digital platforms, think about on-demand segments’ increasingly important role as a trusty companion during these challenging times.
- “A podcast, or a newscast?” As we’ve mentioned, we encountered this feedback a lot. But it hits on something we’re hoping gains traction: There’s a third way, and it’s called a microcast. While one could certainly call CVB a “podcast” in terms of format and distribution strategy, it leans more “newscast” when it comes to content and sound. Make no mistake: You hear conversational scripting in CVB that you would never hear in a public radio newscast (“We’re recording at 5:05 p.m.,” for example). However, its direct, snappy cadence and use of reporter spots — which are made for newscasts to begin with, lest we forget — point to clear radio similarities.
- Direct insertion > promotion. You can promote a podcast all you want, but, as you can see from the graph above, that can’t compete with a tech giant inserting your audio directly into listeners’ ears. The time we spent cultivating relationships with these big platforms definitely paid off.
- Teamwork makes the dream work. Practically every WBUR staffer with “reporter” in their title appeared in a CVB episode at some point. Many appeared multiple times. Even a few of our colleagues who report, but not regularly — like the Radio Boston team, newscast writers and announcers — were heard in the microcast.
- Second-wave ready (if it comes to that). Even though the pace of pandemic news has slowed in Greater Boston since the hectic days of March and April, we haven’t (yet) mothballed CVB, because we’re still able to reliably produce a worthwhile 3 minutes every day. But when that day comes and we put CVB on the shelf, we have a well-worn process that we can rapidly put back into action.