Ruthless Optimization, Part 3: How we use Slack, Zapier, and Dropbox to automate and optimize Zengineering Podcast production.
A lot of our production workflow is automated. This is a breakdown of how that works, and the time is saves us.
“Ruthless Optimization” is a series of posts about how we produce Zengineering Podcast without wasting a bunch of time. Check out Part 1: Thoughts on a Workflow, or Part 2: Tools and Tech to get caught up. If you really don’t have the time, at least take a look at Part 2, that’s the breakdown on the all the tools mentioned in this part. Otherwise, strap in, this is where it gets good.
Let the Robots Work
As I’ve eluded to in the previous sections, I’ve carefully chosen these tools in order to assure we at Zengineering Podcast blow as little of our lives and creative energy as possible on mechanical tasks that robots can do for us. There are a number of things in anyone’s workflow that are purely administrative like, moving files around, uploading, publishing, and making sure everyone knows there’s a new episode. I’ve deliberately built out our workflow using tools that we’re able to link together to keep this simple.
Zapier.com
There’s a bit of hackery, but largely this is all orchestrated with Zapier. Zapier is a trigger service. IFTTT is another. These are services that do nothing but provide you the ability to link different internet tools together, and set up “if this happens, then make that happen”, triggers to do one thing or another in a fairly non-technical way. It doesn’t require coding, but it is still basically scripting out a series of interactions. You’ll start to get this as I go on. Let’s continue with the workflow, I’ll run through each phase.
Pre-Production
Pre-production is mostly run in Slack and via email. When we have an idea for an episode, or we book a guest, we start a new channel with the prefix “exxx_.” (That “xxx” is for the eventual episode number (like 042), not anything more fun.) There isn’t much automation in here, but it’s worth running through because this channel will be used in a number of triggers.
In that channel we drop links on the topic, have spurts of conversation, chat with the community a bit if anyone dips in. Over time the “Zengineering” of the conversation rises to the top, or we get a date locked on a guest and we decide we’re good to do an episode. Often nothing comes of it, and eventually the channel is archived.
(Want to see this in action? Drop us a buck on Patreon, you’ll get access to most of this.)
Production
We pick a topic, or get things going with our guest, and then record. The technical details for that are in Part 2 on podcasting tools. As we’re recording we use the Slack channel to jot down notes, and find any links we’re stoked enough about in the course of recording that we want them in the show notes. (Episode 38 is a great one for show notes. Lots of YouTube links to old pop music.) If we’re on our game enough, which only happens sometimes, we’ll drop a pull quote here right after it happens.
Post-Production
Okay. Here’s where the automation kicks in!
Autopost
We use Zencastr to run a little post production magic, and as a post Zencastr fallback Auphonic (which is actually also how Zencastr does it, via API.) This evens out the mic levels relative to a certain dB target, and relative to one another. Then it mixes them together and runs some other filters to clean that mix up. It can tell who’s speaking on a track and favor that on in the mix, which cleans up some interruption and background noises, stuff like that. After this it outputs a single mp3 file. This is the one we edit.
Estimated time saved: 2 to 5 hrs on mix, match, processing and output.
Editing
No real shortcuts here, some episodes don’t take too long to edit, some take longer. It usual depends on topic complexity, and other factors like weird locations. Here templates are your friend. An edit starts with a template file, which is just an empty work file, saved with a bunch of settings dialed in. Intros, outros, levels to make that match the output track from Zencastr are all dialed in. Drop the file in and you’re good to get editing. It’s nothing big, but it saves a half hour of dragging, dropping, and number setting.
As the edit happens we start a Slack post with a dirty summary of what’s ending up in the edit topic-wise, and we catch whatever quotes we catch and drop them in the Slack channel.
Once the edit is finished the final file is exported.
Estimated time saved: 1 to 2 hrs on re-listen for quotes, .5 hr on edit prep.
Packaging
In Slack we refine the descriptions, come up with a final title, pick a pull quote. Those are finalized when we add them to Slack messages including a particular ID tag.
In Photoshop we use similar templates to make graphics to go with each episode. This mostly entails changing the episode number and dropping a new photo in the background. These templates are built for one click export to the right files sizes and resolutions needed for every publishing platform we use. The files are exported, named based on the title of the episode, and placed in particular files on our Dropbox.
At this point all of the media is ready to go, and it’s all sitting in the right place.
Estimated time saved: 1 to 3 hrs on image and graphics prep
Publishing
Files are manually uploaded to Patreon for our backers — Thank you guys! — and set to automatically change from “patron only” to “public” after 7 days. Patreon spits out an RSS feed that we use with Zapier to trigger a script to generate the Medium post that will become our show post and notes.
This trigger gets the title, description, and pull quote from the Slack channel, and the images from the Dropbox. Then the trigger adds a bunch of other links and stuff to the mix and creates the show notes post via a bit of hack. (This is probably the most complicated part, Zapier is running a chunk of Javascript that we hacked together.)
Files and graphics are uploaded to Libsyn, which runs our public RSS feed. It’s set to publish a week after Patreon. This RSS feed is the one that delivers the podcast to the rest of the world, and triggers a bunch of other fun stuff. Libsyn automatically publishes to YouTube and Soundcloud, if that’s where you happen to listen to us.
Estimated time saved: 1 hr on show notes compilation, 1 hr on YouTube/Soundcloud publishing.
Promotion
Here comes the automation monster. Promotion.
In the week between Patreon and public release we clean up the show notes, put in any cool links from the episode or that we dropped into the Slack channel.
Zapier watches the RSS feed and triggers once the public episode is released to all the podcast apps around the world. Zapier lets us know via Slack that it’s triggered, and then sets a 24 hr delay.
In that 24 hours we add the embed to the show notes, publish the show notes, and put the final link to the show notes into the Slack channel.
After the 24 hour delay Zapier fires up again, it pings the Slack channel to grab the title, description, the pull quote, and the show notes link. It pings Dropbox to grab the packaging media, and it sends all of this to a bunch of social media posts in Buffer. All these are set to different specs on different schedules. Twitter, Facebook, Instagram, and LinkedIn all get an initial post letting our followers know that there’s a new episode, then there’s a weekend post with a different image, and an ICYMI post the week after.
This queues us up with a baseline set of consistent reminders. It’s nothing fancy, but it makes sure that anyone who likes to follow us via any given channel won’t miss an episode. We use our personal social accounts to promote and interact with the episode posts, but that’s personal, not automated (mostly).
Estimated time saved: 10–12 hrs on post generation, scheduling and publishing.
And that’s it! Well most of it. Conservatively, we’re saving 15–30 hours in a given week by having the robots handle this stuff. That’s between 1 and 4 full workdays. That’s nothing to scoff at. If you want to go deeper on the methodology behind this kind of optimization check out this episode of the podcast:
Thanks for reading, if you enjoyed this be sure to check out the podcast, give me a clap, and follow for tech nerd stuff. — Adam
