There are a lot of resources on the internet for making a podcast. But, how do you actually put all these tools together to get a finished podcast episode. In this guest post, producer Michael Falero walks through the nitty gritty of the process of creating Backyard Cambridge, a local news podcast he created in collaboration with Glow's co-founder.
Note from Michael: Different podcasts use different production methods. This is ours. Backyard Cambridge is a narrative show, so talk show style podcasts often use a different process. This guide also assumes you've already set up your podcast website and hosting.
Step 1: Develop the podcast season’s focus and episode topics
First we have to develop the focus of our podcast season. This means creating an entire season overview, one that also gets into the questions of what each episode will focus on. What ideas do we want to tackle, and in which episodes will we cover those ideas? For Backyard Cambridge, which is a podcast about the politics of Cambridge, MA, we’re always trying to find the best stories that represent the city’s political currents.
Google Docs is a great place to hash out these ideas. Our producer, reporter and editor will often exchange comments and make edits as we finalize our overview.
The series overview is often where our pie-in-the-sky ideas about what we want to accomplish in a season collides with hit reality. We have to decide what questions we prefer to answer, how many topics we can get to, and who we can find and interview to fill each episode.
Generally we will create between 2 and 5 versions of the series overview, naming them “Backyard Cambridge Season 3 Overview V1” and so on.
Step 2: Write a detailed storyboard for each episode
Once our team has come to a final series overview, we essentially do the same thing at the episode level. That means making a “storyboard” for each episode. What questions will we ask at the top? What points should we try to hit in Act 1, 2, and 3? What existing tape will we use, or who do we need to interview for the episode? What active sounds can we include to craft the sound of the episode, like from public events or outside venues?
In this phase, our reporter conducts research of previous reporting on the topic and contacts experts in the Cambridge area who can provide ideas of whom to contact or provide introductions for us.
Here’s what our episode storyboards usually look like in Google Docs:
As you can tell, these documents are short, rarely more than two pages long. Like the season overview, the episode storyboard serves as a focusing document for the reporter when she writes the first script.
Step 2.5: Start keeping track of your production cycle
We always say that organization is key, and it can even make your podcast better. We didn’t create a dedicated production tracking spreadsheet for Backyard Cambridge Season 3, and at some points in the process it would’ve really helped. We like using a spreadsheet like Gimlet producer Eric Mennel’s Google spreadsheet. This information can include your interviewee contacts, a tracker for your interview tape, the season’s production schedule, research of found news articles and online information, as well as other items.
Here’s how the Mennel spreadsheet lays out that information:
Step 3: Conduct your interviews, transcribe them, pull quotes
Once we’ve got an idea of our episode, our reporter starts setting up interviews with experts and attending public events relevant to our topic. She uses field recording equipment (you can find an affordable recorder and microphone using Transom.org’s Field Gear Guide). The reporter writes up her list of questions beforehand and researches her guests before sitting down with them.
Transcribing interviews is the step that nobody wants to do, yet is essential for narrative podcasts. Transcription shows us what tape we have to work with. It’s a step that pays dividends long after we’ve done it. Instead of only finding good audio moments based on our memory of the interview, we can consult our transcript(s) to find the quote we need. Here’s one our reporter made from her interview with a campaign spokesman:
Free tools for manual transcription like OTranscribe or paid auto-transcribing services like Trint and Descript make this time-consuming step less difficult. We generally will use OTranscribe for Backyard Cambridge. Using a Tape Tracker spreadsheet like the one mentioned in Step 2.5 can help you keep track of which interviews you have and haven’t transcribed.
Note: Don’t underestimate how much time transcription can take. A word-for-word transcript can take between 4x and 6x the audio file’s length to complete. So, a 30-minute interview could take you up to three hours depending on its complexity. Still, transcribing your audio is essential.
Once you’ve finished and transcribed your interviews, you can pull your quotes or “selects” that you want to use in your scripts. If you’re using the Mennel spreadsheet above, you can see the Tape Tracker tab includes fields for logging your tape and then noting which selects you’re likely going to use from each interview. With your selects chosen and your research complete, you can begin to write your script.
Step 4: Write the first episode script
At the script-writing phase, our reporter builds a narrative using everything she’s created so far: the episode storyboard, the interviewee transcripts, background research, and her selects. Generally, she will use the selects that work best with the ideas she’s trying to communicate, and then write into and out of those quotes with narration. Here’s what our scripts usually look like:
As you can see, we try to spend about equal time narrating and using interview tape. This ensures that our reporter isn’t narrating long, dense explanations, and that listeners are hearing a diverse set of voices that keep their attention. In the completed version 1 of the script, our producer and editor will add comments until it’s a full, functioning story, with an introduction, two or three acts, and a conclusion. Note that we’re still using Google Docs to collaborate on writing and editing the scripts, and we’ll continue to share files via Google Drive folders, like interview audio and transcripts.
Step 5: Make a bad first audio draft
In creating Backyard Cambridge, we’ve learned that the fastest way to make a great podcast episode is first to make a bad podcast episode. That’s because it’s always better to get to your audio editor as soon as you have a workable story script. The story will in all likelihood be pretty bad. It won’t sound right. But hearing your story will help you find the flaws in the narrative, which you can then use to improve your written script.
For our audio production we use REAPER, an inexpensive and lightweight digital audio workstation (DAW).
Our Backyard Cambridge drafts tend to look like this:
The big takeaway from this screenshot is that we have four basic audio groups: music (our intro music, any music during the episode, and credits music), ambient sound (also known as “ambi” — in this episode it was a campaign phone banking session), our reporter’s narration, and interviewees (also known as actualities or “acts”, each on their own track). Our producer creates a rough cut using the script as the guide, and then renders everything to an MP3 file for the team.
Step 6: Listen to the audio draft, edit the script, rinse and repeat
Once we have an audio file of the first version of the story, the team listens to it and make notes about what does and doesn’t work in the narrative. We try to think about what ideas are unclear, which interviewees we’re hearing too much or not enough from, and what sounds or music the story needs to make it more compelling.
We then create a new version of the script - script V2 - and edit it together according to that feedback. We then repeat Steps 4 and 5 - create a script, make the audio, listen to the story - until we’re satisfied with the final script and audio cut.
Here’s what the final version of the above episode looked like:
Step 8: Add music and FX, and normalize the podcast audio
Once we have the final draft of our story and our audio matches the script, we add any last music elements, like music beds for narration. Finally, we add FX to “mix” the sound. At a minimum, we add EQ to each track (to shape the sound so it sounds more pleasant to the ear), as well as a compressor (which catches louder sounds so they don’t overwhelm the listener). If you want an overview on how to EQ, we like producer Rob Williams’ video and related chart on the effects some simple EQing can have.
Some more advanced FX, also known as “plugins”, come from companies like iZotope and Waves. iZotope’s RX 7 Essentials gives some great tools for cleaning audio, making background sounds or breathing noises less harsh, and overall making your audio sound clearer. We started off with EQ and compression, and then moved onto these advanced plugins over time. We talk more about plugins like these in our tools guide.
One of the most important steps in our podcast production cycle is at the end, when we “normalize” our podcast audio. What this means is we use a tool to make sure every voice and music track on the final audio file has a roughly equal apparent loudness. This is called a “loudness standard”. The basic thing to understand is that when a podcast is normalized, listeners don’t have to adjust their phone or car volume when they hear different speakers on a podcast. Instead, everyone sounds equally loud. We normalize our podcast to -18 LUFS, per Transom’s guidelines.
To normalize our podcast, we leave it to the experts at Auphonic. Auphonic is a freemium site that normalizes audio files and provides other audio processing services. We have an Auphonic account and upload our final audio using a “production”, with settings for a final mono MP3 file set to -18 LUFS. Auphonic’s processing takes about five minutes for our 20-25 minute episodes. (Note: you can also purchase a Loudness meter from Waves or iZotope if you’d like to adjust your podcast’s Loudness yourself.)
Step 9: Upload to the podcast host
When the podcast audio is normalized, we’re ready to post it. We upload our podcast to Squarespace, which is our podcast host.
We add show notes using Squarespace’s blog feature, making sure to include links to everything we’ve mentioned in the episode, as well as credits to organizations we talked to and music we used:
Step 10: Market the podcast
When the episode is live, we do our best to market the new episode across social media, particularly Twitter and Facebook, but also Cambridge-specific Facebook Groups and local sites like NextDoor. We also contact interest groups we spoke to in each episode and mention that the episode they’re in has been published, and provide them with a Twitter link so they can share it with their followers.