Your podcast recorded 6,750 words this week and YouTube got none of them

Open the transcript of your last episode and count. At a normal speaking pace of about 150 words per minute, a 45 minute episode holds around 6,750 words. A 60 minute one holds 9,000. That is not raw notes. It is a finished argument with an opening, examples, objections answered and a conclusion, because you already said it out loud to an audience.

Now measure a YouTube video in the same unit. A 12 minute narrated video is about 1,800 spoken words. Divide 6,750 by 1,800 and you get 3.75. One episode is three to four complete videos of material, and not thin ones. Each of them can take one segment of the episode and give it the full treatment that the segment never got, because in the episode it was four minutes between two other topics.

Podcasters usually get this backwards. They try to turn one episode into one video, which means a 45 minute upload with a retention graph that falls off a cliff at minute two. The episode was built for someone doing the dishes. A YouTube video is built for someone deciding in eight seconds whether to keep watching. Same words, different shape, and the shape is what the split fixes.

There is a second reason the split matters more than the upload. Search does not read audio. Every one of those 6,750 words is currently invisible to the only discovery engine that sends new listeners to a show for free. Three or four titled, described, subtitled videos put the same argument in front of a search query. The same principle applied to written articles is the closest cousin of this one.

• 150 words per minute: a 45 minute episode is about 6,750 words.
• A 12 minute narrated video is about 1,800 words.
• One episode is three to four full videos, not one long upload.
• A 60 minute episode is 9,000 words, or five videos.

There are two monetisation doors in front of a podcast right now and they count completely different things. Neither one counts audio downloads, which is why a show with a healthy download chart can still earn nothing from either.

Door one is YouTube. From 1 February 2027, a channel joining fresh needs 1,000 subscribers plus 8,000 qualified public watch hours in 365 days, or 1,000 subscribers plus 20 million qualified Shorts views in 90 days. Both numbers doubled from the old bar. Turned into planning units, 8,000 hours is 480,000 minutes, and a 12 minute video watched at 40 percent gives 4.8 minutes per view, so the door is roughly 100,000 views.

Door two is the Spotify Partner Program, and its requirements are published: at least 3 published episodes, at least 2,000 consumption hours on Spotify in the last 30 days, at least 1,000 audience count on Spotify in the last 30 days, hosting on Spotify for Creators, and a legal address in one of the eligible markets, which include the United States, the United Kingdom, Canada, Australia, New Zealand, Ireland, France, Germany, Austria, Belgium, Denmark, Finland, Iceland, Liechtenstein, Luxembourg, Monaco, Netherlands, Sweden and Switzerland. Keeping it requires publishing an episode and earning at least US$ 10 every six months.

Note what the Spotify video payout is not: a published rate. Spotify states that premium video revenue is calculated by a proprietary formula that it may adjust, based on factors including total qualifying consumption time and the markets the Premium users are registered in. So plan on the eligibility numbers, which are public and testable, and treat the payout as upside rather than a forecast. The full arithmetic of the 8,000 hour door is the one you can actually model.

Every podcaster has already decided to do video at some point, and most of them stopped at the same wall. Podcast editors in the United States charge roughly US$ 40 to US$ 120 an hour, or US$ 150 to US$ 500 per episode for a standard interview format, and a 60 minute interview commonly takes two to three hours to edit properly. Narrative shows with sound design run from US$ 500 to well past US$ 2,000 an episode.

Multiply by a weekly schedule. At US$ 300 an episode, 52 episodes is US$ 15,600 a year, and that is for the audio plus a straight video cut, not for three or four separate videos per episode with their own hooks, their own titles and their own visuals. Doing the split properly with a freelancer multiplies the bill by three or four, which is how a US$ 15,600 line item quietly becomes a US$ 50,000 one.

Doing it yourself replaces the money with time and the time is worse. A 45 minute episode cut into four videos, each with a hook, b roll, subtitles and a thumbnail, is a full working day. Add it to recording, guest booking, editing the audio and publishing, and video becomes the task that slides to next week every week for a year.

That is the real reason the archive stays audio only. It is not a lack of interest and it is not a lack of material, because the material is already recorded. It is that the marginal cost of the fourth video is the same as the first, and nothing about doing it more often makes it cheaper when a person is doing it by hand.

• US$ 40 to US$ 120 an hour is the standard freelance rate.
• US$ 150 to US$ 500 per episode for a standard interview format.
• Two to three hours of editing for a 60 minute interview.
• 52 episodes a year at US$ 300 is US$ 15,600, before splitting into separate videos.

You take the transcript or the show notes of the episode, pick the segment worth its own video, and hand the text over as the brief. FalconVid runs the research, writes the script to be heard rather than read, narrates it with ultra realistic premium voices, edits scene by scene, adds karaoke subtitles, renders in 4K and produces the 9:16 cut for Shorts and Reels. The specialists work in parallel and a finished video arrives in up to 30 minutes.

Channel DNA is what keeps four videos from an episode looking like one show instead of four experiments. The voice, the visual identity, the intro, the subtitle style and the format stay locked across every video, which is exactly the consistency a podcast brand needs when the same argument appears in four places. If the show has two hosts, two voice narration is available and each speaker keeps a distinct voice.

If you want a face on screen, an ai avatar can present with the same face and the same voice across the whole catalogue, and it is charged by the seconds the face is actually visible, so most shows use it only in the hook. The Studio handles the rest: you watch the first cut and shorten the intro, swap a piece of media or change the music without regenerating the whole video.

The cost side is where it stops being a comparison. On today's ruler a 12 minute video is 1,731 credits in economy, US$ 5.42 at US$ 0.003133 per credit. The four videos that come out of one 45 minute episode are 6,924 credits, US$ 21.69. A full year of weekly episodes turned into four videos each is 208 videos, 360,048 credits, about US$ 1,128, against the US$ 15,600 that the audio edit alone costs at freelance rates.

Quality mode is a lever, not a setting you pick once. Economy at 1,731 credits is right for the wide catalogue where you are still learning which segments travel. Balanced at 4,624 and premium at 16,158 belong to the two or three videos that already pull views and deserve the better engine. Nobody runs a whole month in one mode, you mix them, and the credit cost is shown before you generate anything.

Be clear about what you are getting, because the distinction decides whether you are happy with the result. FalconVid does not film your recording session, it does not edit footage of you and a guest talking, and it does not sync a camera angle to your existing audio. What it produces is a new video built from your words: narrated, illustrated, subtitled and cut for YouTube.

For most shows that is the better artefact anyway. A video of two people talking for 45 minutes is a poor YouTube video and a good podcast, and putting it on YouTube unchanged is why so many podcast channels have thousands of uploads and no watch time. A tight 12 minute video on one idea from that conversation is a good YouTube video, and it points people back to the full episode.

There is one case where this genuinely does not fit: if the appeal of your show is the chemistry between two specific people on camera, the interview footage is the product and no generated version replaces it. Film that. Use this for the argument segments, the explainers, the topic breakdowns and the listener questions, which is where most of the searchable value in a podcast archive actually sits.

The other honest limit is that a transcript is a brief, not a script. Spoken conversation contains tangents, interruptions and half finished sentences that read badly when narrated. The machine rewrites it into something built to be heard as narration, which means the finished video says what you said, not word for word how you said it.

The maths above assumed you start from this week. Most shows are not starting from this week. A podcast two years into a weekly schedule has around 104 episodes, which at 6,750 words each is roughly 700,000 words of finished argument, and at three to four videos per episode it is 300 to 400 videos of material that already exists and cost nothing new to create.

Nobody works through an archive like that by hand, which is why it sits there. With production running in parallel, 2 to 50 pipelines at once depending on the plan, the archive becomes a queue rather than a project. You approve a calendar, and the back catalogue publishes itself in the background while the new episodes keep coming.

This is also the cheapest way to answer the 8,000 hour question. 100,000 views is the door, and 300 archive videos aimed at 300 different search queries is a far more plausible route there than 30 uploads of full episodes. Every one of them is content you already made, converted into a format the search engine can actually read.

The same approach applied to course material works on lesson recordings and workshop transcripts for the same reason: the expensive part, deciding what to say, was already done.

Here is the move that has no equivalent in the manual version. Narration covers 63 languages, and duplicating a project into another language charges only the difference rather than a full new production. An episode segment that performed in English becomes a Spanish and a Portuguese video without rewriting or re recording anything.

For a podcast this is unusually valuable, because the audio show is locked to the language it was recorded in. Your English episode cannot be listened to in Spanish, but the video built from it can be watched in Spanish, and it carries the same argument, the same examples and the same conclusion to an audience the podcast feed will never reach.

The channel side scales with the plan: 1 channel on Starter, 5 on Pro, 10 on Business, 25 on Agency and 50 on Scale, each with its own calendar, identity and language. A show that wants an English channel, a Spanish channel and a Portuguese channel is a Pro account, not three separate operations.

None of the creation features are locked by plan. Every plan gets the engines, the avatars, the Studio, the subtitles, the 4K, the Shorts and publishing to YouTube, Instagram, TikTok, Rumble and Facebook. What moves with the plan is volume, channels and simultaneous generations, plus the dedicated server and the Senior AI Analyst from Pro upwards, a fixed person with name, face and voice who reads your account and writes every two days with the move to make.

The manual ceiling is easy to calculate and it is low. One episode into four properly cut videos is a full working day, so a weekly show that also wants video is asking one person for one extra day a week forever, or roughly US$ 50,000 a year in freelance time to buy that day back. In practice most shows do neither, publish the occasional full episode to YouTube, and conclude that video does not work for podcasts.

The automated ceiling is a different scale of number. A year of weekly episodes becomes 208 videos for about US$ 1,128 in economy, the two year archive becomes 300 to 400 videos in the same queue, and the whole thing duplicates into other languages for the difference. The bottleneck stops being production and becomes the only thing that was ever really yours: deciding which segments deserve their own video.

That is the honest comparison. Not a machine that makes better arguments than you, because it does not have your guests, your reporting or your opinion. A machine that stops the arguments you already made from staying inside an audio file where nothing can find them.

Start on the free plan, four videos of up to three minutes every 30 days with no card, and put one episode segment through it before deciding anything. Everything on the FalconVid home page and the text to video page works the same way on the free plan as on Scale. If your archive is more reference than conversation, the book summary channel format is the closest neighbour to what you already have.