How to create a podcast [Bonus Issue]
Substack recently had a comment thread for publishers using the platform for podcasts. I tried to be helpful and ended up writing far more than I originally intended. I thought I would duplicate it here for reference purposes. If you are not interested in how to produce your own podcast then you should not waste your time reading further. Back to regular scheduled programming next week!
Some thoughts for people thinking about doing podcasts
Quality matters. A lot. Adding another newsletter is usually possible for most people - it might replace time spent on Twitter or Facebook. But adding another podcast comes out of "audio listening time" - and that usually means it needs to replace an existing podcast they are enjoying. So you need to be better than the next best alternative (maybe a lot better as people are generally risk adverse and will judge you quickly vs a feed they have enjoyed in the past)
Sound Quality. Quality is about content, but it is also about sound quality and sound editing. Getting to NPR level is difficulty and/or expensive. But getting much better than Anchor is not that hard if you invest a little. If you aren't going to make a small investment in doing this right, I'm not sure it is worth your time to bother doing it at all
Investment. What does investment look like? Here is where I have landed with a lot of trial and error and speaking to many people in the podcasting space (Hi Adam Davidson! Hi Ben Shapiro!)
Get a good mic. I use Blue Yeti. It was out of stock for a long time, but it looks like it is now available for delivery tomorrow.
Settings. For settings you want it on high volume, medium gain (experiment!), and the pattern that looks like a little heart. Then you speak very close to the front of the mic. The heart pattern means the mic only picks up sound from directly in front of it. With that setting you are usually safe from street noise and kids downstairs. It might not be as good as a full podcasting studio, but its pretty close and much cheaper!
Cost $118 right now on Amazon
Troubleshooting. If things don't sound great (which they did not at first when we did Super Serious) the most likely issue is the cable. Try to fix that before buying a new computer (my partner did it in the opposite order...)
Recording Software. Use recording software. DON'T use Zoom's included feature. I use AudioHijack and like it a lot (~$40 for software one-time - not a subscription!)
Templates. You will want to create 1-2 "recording templates". One for when you record on your own, and one for when you record over Zoom
Single Stream. Record on your own is pretty straightforward. USe a subset of the interview set-up.
Interviews. Ideally you have the person on the other end record their audio locally and then send you the file. I have found that this is not worth the effort (especially for one-time interviews. But even when I am recording with my partner on Super Serious)
Template. Here is what the template looks like (two 'paths', with both joining up in the final step into one file):
APPLICATION (Zoom) [playback gaps filled with silence] -> Peak/RSM -> Output device (airpods) -> Recorder (WAV) -> Recorder (MP3)
INPUT DEVICE (Blue Yeti) -> Peak/RSM -> Recorder (WAV) -> Recorder (MP3)
This will result in 3 files - two large WAV files - one for each individual, and one combined MP3
When I work with an audio cleanup company I send them all three. They work with the WAV files and just use the MP3 for reference. They send me back a final MP3 file built from the WAV files
When I edit myself (which I generally do now), I mostly just use the MP3. Only when something has gone wrong with the recording do I use the WAV files (maybe 2-5% of the time)
Troubleshooting. Sometimes the audio does not record, but you should be able to see that visually. Test before you start, and if it is not working, I have found the answer is usually (always!) "update the software" or "re-boot the computer"
Cleaning up the audio with professionals
I had my assistant create a short list of professional companies that clean the audio. I then spoke with and tested four different companies at different price points and effort levels.
Effort Level 1. The lowest level cleans up the sound, fixes the levels, etc.
Effort Level 2. All of #1 plus they will listen to the entire show and take out the ums, ahs, you knows, coughs, and long unnecessary pauses. They will also follow any directions you put directly into the audio (i.e., “Audio engineers: Please remove that last question and answer. I am going to re-ask it now.”)
Effort Level 3. All of #2 plus they will put in background music, etc. I found this was both not worth the $$s, and was actually a negative. NPR may be able to do this right but for under $300 per episode the quality was such that many of my listeners complained about it - the background music messed up listening at 1.5x speed for example.
I recommend Level 2. For most episodes for most speakers this shortens episodes by 10-20% (!), without removing any content, and having the remaining content sound a LOT better. The ability to give instructions during the show (even if it is just, “Please take out the part where my kid interrupts us and I need to explain where the peanut butter is”), makes the entire process seem way more professional with guests and puts everyone at ease (If you make the guest sound smarter they will be more likely to share it!)
If interest I can pass along my short list of who I tested, but there are lots out there and prices and capabilities have likely changed in the last 18 months. Cost for this should be just under or around $100 per episode for "level 2" (and include transcription)
Cleaning up on your own. For Super Serious I clean all the audio on my own. I've gotten a lot better over time. If you listen to the early episodes you will notice a HUGE difference from the latest episodes (I think). It's enough of a different that one of these days I will go back and re-edit the early episodes.
For personal editing I use Audacity (but again, lots of options). I have a "project" that has my intro and out-tro music, then I upload the MP3 audio file. You can see visually where you need to adjust levels. For the ums and ahs, I listen to the entire episode "zoomed in" and every time I hear an um, or a repeated word or "see" a long pause, I just cut it out. Then I listen to it again and if it sounds right. If it does I move on and keep listening (if not, lots of use of command-z). For a 12 minute episode I usually takes me ~20 minutes to edit these days. There are a few judgement calls. I should likely be outsourcing this, but there is value in my hearing the full episode again after it was recorded (it's like feedback on my performance), and doing it this way is more mentally stimulating for me than just listening to a full edited version by a professional
4. Transcripts: I do transcripts for Marketing BS, but we do not for Super Serious. There are two ways to do transcripts for two different purposes:
a. AI transcripts. There is software that will do this for you. It's pretty good! I would say about 90% accurate. If you just need something for SEO purposes and you are upfront that it is AI 90% accurate, this is a cheap way to do it (and something we are considering for Super Serious). It is also very cheap
b. Professional transcription. You can pay someone to listen to your entire show and manually transcribe it. This will be someone from a low wage country, and you should be about to transcribe an hour long show for $50 or so (if I remember correctly - don't hold me to it). This will get you a fully edited transcription in whatever format you agree to (i.e., I had them put in the speakers name in bold with a colon every time they speak). This is what you need if you are expecting people to read your episodes (which a LOT of people do. I think my Marketing BS podcasts are something like 70% read, 30% listen).
c. Hybrid. You can use AI to transcribe the show, and then a professional to listen and fix the 10% errors. This is not worth doing. The cost is going to be the same (or higher!) than the pro version.
5. Indexing your podcast: You will want to submit your podcast to the major podcast players. Its a bit of work, but once you get it on Spotify, Apple and 1-2 others, the rest will just do the work without you having to do anything. I believe Substack even has a primer on how to do this (just google it)
I think those are the major points.
When I started the Podcasting tools on Substack were pretty terrible. I only used it because I already knew how to do it from my newsletter experience. But now the analytics are MUCH better!
tl/dr: Invest in quality. Your subscriber's time is valuable, don't waste it with low quality audio or low quality content. Better to not do it at all rather than create something that lowers the quality of your newsletter brand.
Keep it Simple,