People ask how the Roadman Cycling Podcast gets made. Two episodes a week, 1,400+ interviews in the archive, over a million monthly listens, and guests who don't usually say yes to podcasts. The honest answer is that the show is a small media operation now, not a microphone on a desk.
This is the walkthrough. The gear, the booking system, the prep, the editorial calls, and how one recording becomes two products — an audio episode and a YouTube cut.
If you're a cyclist thinking about starting a podcast, or a coach who wants to understand what goes into the podcast before pitching yourself as a guest, this is the piece to read.
The gear in the studio
The Dublin studio runs a Shure SM7B into a Cloudlifter CL-1, then into a Rodecaster Pro II. The SM7B is the industry standard for a reason — it rejects room noise, handles loud vocals without clipping, and sounds broadcast-ready with minimal processing. The Cloudlifter adds 25dB of clean gain so the Rodecaster doesn't have to work its preamps hard.
Monitoring runs through Beyerdynamic DT 770 Pro headphones at 80 ohms. Closed-back is non-negotiable for interview work — any bleed from open-back cans ends up in the guest's audio via your mic.
Video capture uses two Sony ZV-E10 cameras: one head-on, one at a 30-degree angle for cutaways. Both feed into an Elgato Cam Link 4K and get synced in post. Lighting is two Aputure Amaran 100d panels with softboxes, which is overkill for audio-first work but matters once the video goes to YouTube.
For remote guests, everything is captured in Riverside.fm at 48kHz WAV with local recording on the guest's machine. That local capture is the single most important piece of the chain. Broadband drops, Zoom compresses, Teams mangles consonants. Riverside records locally and uploads in the background, so a guest in Boulder at altitude with flaky wifi still sounds clean.
Total studio build sits around €4,500. The podcast ran for 18 months on a €400 setup before any of this was justified. Upgrade when the audience demands it, not before.
The guest-booking workflow
Guests are booked 6–10 weeks out. The pipeline lives in Notion with four columns: pitched, confirmed, scheduled, recorded. At any given time there are 30–40 names in flight.
Roughly 60% of bookings come from direct outreach — a short email with three specific questions we'd want to ask, a link to a recent episode in their field, and two date options. The specific questions matter. Generic "would love to have you on the show" pitches get ignored by people who get 20 of those a week.
Another 25% come from guest referrals. After every recording, the last off-air question is: "Who should we talk to next?" Dan Lorang's referral network alone has produced six episodes. The remaining 15% are inbound through press & media kit enquiries, usually tied to a book launch, a team announcement, or a product.
Scheduling uses SavvyCal with a 90-minute block, built-in buffer, and automatic timezone handling. The buffer matters — interviews with a World Tour coach that stop dead at the 60-minute mark because the next calendar block hit are the ones that leave the best material on the floor.
How we prep for a podcast interview
Every guest gets 2–3 hours of preparation. That's non-negotiable. An unprepared host produces an interview the guest has already given ten times.
The prep sheet has four sections. One: a biographical summary and the two or three things this person is genuinely known for. Two: a list of questions they've already answered in public, so we don't waste time there. Three: 15–20 original prompts, organised into three thematic blocks. Four: two or three contrarian angles — places where their published view conflicts with someone else's, or where the data has moved since they last spoke.
Listening to two or three of their previous podcast appearances is the highest-leverage prep. It tells you their verbal tics, the stories they lean on, and — crucially — the questions that made them think rather than recite. You want to engineer more of those.
The question list is a map, not a script. A good interview derails inside the first ten minutes and follows the guest's energy. The prep is what lets you derail confidently, because you know where every thread eventually reconnects to something useful.
What we cut and why
A 75-minute recording usually ships as a 62–68 minute episode. The cuts are mostly three things: false starts, tangents that went nowhere, and moments where the guest asked to revise an answer off-air.
We don't cut disagreement. We don't cut awkward silences if the silence is doing work. We don't cut the guest correcting the host — those are often the most valuable 30 seconds in the episode. A podcast that edits out every moment of friction ends up sounding like a press release.
What we do cut ruthlessly: filler preamble before the guest has settled in, dead air from tech issues, and any segment where we realised mid-recording we were on the wrong track. Better a tight 60 than a bloated 90.
The YouTube vs audio split
The same recording produces two different products. Audio listeners are usually training, commuting, or walking the dog — they want depth and will sit through a 90-minute conversation. YouTube viewers scrub, skip, and bail inside 30 seconds if the thumbnail and opening hook don't hold.
The audio episode ships with full show notes, chapter markers, and timestamped links. It goes to Apple, Spotify, and the main RSS feed within 48 hours of recording. The YouTube version gets a different opening — a 15-second cold open using the single best moment from the interview, pulled forward before the intro.
Three to five vertical clips of 45–90 seconds each go to Instagram, TikTok, and YouTube Shorts. Those clips are responsible for roughly 40% of new listener acquisition, based on the referral data in Spotify for Podcasters. The flywheel only works because every episode is built to produce clips, not retrofitted afterwards.
If you want to pitch yourself as a guest, or learn more about Anthony Walsh and how the show got here, the media kit has everything. If you want to start your own cycling podcast, buy a €100 dynamic mic, record ten episodes before you touch any other gear, and only then start worrying about the Shure SM7B.
