What actually drives a podcast hosting bill
Podcast hosting is priced on transfer, and transfer is file size multiplied by downloads. Both halves are worth understanding separately, because creators consistently misjudge the second one.
File size is completely determined by bitrate and length. A constant-bitrate MP3 has no compression surprises: 128 kbps for 45 minutes is 43.2 MB and always will be, whether the episode is two people talking or a symphony. The formula is bitrate in kilobits per second, times seconds, divided by 8 to get kilobytes, divided by 1,000 to get the decimal megabytes hosts bill in.
Download count is where the estimates go wrong. Almost every creator thinks in downloads per new episode, because that is the number in the dashboard headline and the number a host asks for in a plan comparison. But every episode you have ever published is still in the feed, and still being downloaded — by new subscribers working backwards, by search traffic, and by apps that automatically fetch a back catalogue when someone subscribes. Each old episode contributes little. Sixty of them do not.
The calculator adds both streams because the ratio between them shifts continuously. In month one there is no catalogue. In year three the catalogue is 150 episodes and it may be moving more data than every new release combined. A host plan chosen on release-day numbers will be outgrown by a show whose audience has not grown at all.
Choosing a bitrate, and what it costs you
Bitrate is the only input that trades quality against money, and it is linear in both directions: doubling the bitrate exactly doubles the file, the bandwidth and the bill.
For speech, 64 kbps mono is the working standard and is genuinely transparent for voice on the phone speakers and earbuds where most listening happens. Podcast directories accept a wide range; Apple's specification asks for AAC or MP3 files and recommends stereo only where the content warrants it, precisely because mono halves the data for talk content with no perceptible loss.
128 kbps stereo is the right choice when music matters — a music show, heavy sound design, or a narrative piece with a score. It is also the safe default that most creators pick without thinking, which means a large share of speech podcasts are paying double for stereo nobody can hear the benefit of.
Above 192 kbps you are into diminishing returns for any spoken-word content, and above 320 kbps you are outside what directories require and beyond audible benefit for almost anything. The calculator warns at both ends, because going too low is also a real error: below about 48 kbps, speech starts to sound obviously compressed on the small speakers where podcasts are consumed.
Enter the total stream bitrate, not the per-channel figure. Encoders differ in how they present this, and getting it wrong is a clean factor-of-two error in every number on this page. If your encoder shows 64 kbps per channel in stereo, enter 128.
Worked example: a weekly 45-minute show with 60 old episodes
128 kbps stereo, four episodes a month at 2,500 downloads each, a 60-episode back catalogue averaging 40 downloads a month each, bandwidth at $0.10 per GB and storage at $0.023 per GB-month.
- Episode size. 45 min × 60 = 2,700 seconds. 128 kbps × 2,700 = 345,600 kilobits. ÷ 8 = 43,200 kilobytes. ÷ 1,000 = 43.2 MB.
- Downloads of new episodes. 4 × 2,500 = 10,000.
- Downloads of the catalogue. 60 × 40 = 2,400.
- Total downloads. 10,000 + 2,400 = 12,400.
- Monthly bandwidth. 12,400 × 43.2 MB = 535,680 MB ÷ 1,000 = 535.68 GB.
- Catalogue share. 2,400 ÷ 12,400 = 19.4% of transfer, from episodes you published months ago.
- Annual storage growth. 4 × 12 × 43.2 MB = 2,073.6 MB = 2.07 GB a year.
- Monthly cost. Bandwidth 535.68 × $0.10 = $53.57. Archive of 64 episodes × 43.2 MB = 2.7648 GB × $0.023 = $0.06. Total $53.63.
Two things fall out of that. First, storage is essentially free and bandwidth is essentially the whole bill: $0.06 against $53.57, a ratio of about 900 to 1. Keeping every episode online forever costs almost nothing; serving them costs everything. Second, halving the bitrate to 64 kbps mono halves the episode to 21.6 MB and the monthly bill to $26.79 — a saving of $321 a year for a change most listeners on a phone speaker will never detect.
Choosing a plan from these numbers
Compare the monthly bandwidth figure against your host's allowance, not against your download count. Hosts express limits in three incompatible ways — gigabytes of transfer, upload hours per month, or downloads per episode — and only the first is directly comparable to what you have just computed. When a host quotes upload hours, it is limiting storage rather than transfer, which the arithmetic above shows is the cheap half.
Watch the catalogue share over time rather than at a point. It rises mechanically with every episode you publish, so a show that is 20% catalogue today will be substantially more in two years even if nothing else changes. A plan chosen with no headroom for that will be outgrown by success you have already had.
The bandwidth-per-1,000-downloads figure is the cleanest number for comparing hosts, because it is independent of your audience size. It is numerically equal to your episode size in MB: a 43.2 MB episode uses 43.2 GB per thousand downloads. Quote that at a host and their answer tells you immediately whether their plan tiers make sense for your show length.
Finally, do not confuse downloads with listeners. A download is an HTTP request for the file, and the IAB's podcast measurement guidelines exist precisely because raw request counts overstate humans — automatic app prefetching, partial range requests and repeated fetches all inflate the number. For a bandwidth calculation that is fine, because every request costs you the same either way. For an audience claim to a sponsor, it is not.
Episode file size by bitrate and length
| Bitrate | 30 min | 45 min | 60 min | 90 min |
|---|---|---|---|---|
| 64 kbps (mono speech) | 14.4 MB | 21.6 MB | 28.8 MB | 43.2 MB |
| 96 kbps | 21.6 MB | 32.4 MB | 43.2 MB | 64.8 MB |
| 128 kbps (stereo) | 28.8 MB | 43.2 MB | 57.6 MB | 86.4 MB |
| 192 kbps | 43.2 MB | 64.8 MB | 86.4 MB | 129.6 MB |
| 256 kbps | 57.6 MB | 86.4 MB | 115.2 MB | 172.8 MB |
Every value is exactly proportional to both bitrate and length, so a 90-minute episode at 64 kbps is the same size as a 45-minute one at 128 kbps — 43.2 MB. That equivalence is the fastest way to see what a stereo decision is really costing you.
What this calculation leaves out
- Partial downloads and range requests. Apps that stream rather than download often fetch a file in pieces and sometimes abandon it, so real transfer can be either above or below a whole-file count depending on the client mix.
- Cover art and feed requests. Every episode carries embedded artwork, and podcast apps poll your RSS feed frequently. Individually tiny, but a popular show serves an enormous number of feed requests.
- Dynamic ad insertion. Stitching ads into a file at request time changes the size served and can defeat caching, raising transfer above the static figure computed here.
- Variable bitrate encoding. VBR files vary with content, so this constant-bitrate arithmetic is an approximation for them. It remains a good one for speech, which is acoustically uniform.
- Decimal versus binary gigabytes. Hosts bill in decimal GB (1,000 MB) while operating systems report binary GiB (1,024 MiB). The gap is about 7%, which matters when you are close to a plan limit.
- The download spike on release day. Monthly totals hide the fact that a large share of transfer happens in the first 48 hours, which is what determines whether your host's infrastructure copes rather than what it charges.
- Video podcasts. Video bitrates are one to two orders of magnitude higher, and the arithmetic here would need the video figures — see the livestream bitrate calculator for those.
Where hosting cost sits in a podcast budget
For almost every show, hosting is not the expensive part. The worked example above costs about $54 a month at commodity rates, against production time that would cost hundreds of dollars in any honest accounting — the kind the cost per finished minute calculator puts a number on. The reason to compute bandwidth carefully is not to save the $54; it is to avoid choosing a plan whose limits you breach, and to know what a bitrate decision costs before you make it for three years at once.
The revenue side of a podcast is priced on downloads rather than bandwidth, and typically per thousand downloads with a CPM, exactly like newsletter sponsorship is priced on opens. If you are sizing whether the show pays for itself, the newsletter revenue calculator models that CPM structure directly, and the sponsorship rate calculator handles the case where a sponsor is buying a read rather than an impression. Where a podcast has a paid tier, the membership tier break-even calculator and the churn and lifetime value calculator apply unchanged.
One structural point worth knowing: because storage is roughly a thousand times cheaper than the transfer it enables, there is essentially never a financial reason to delete an old episode. If a host encourages you to prune your archive to stay within an upload-hours limit, they are limiting the cheap resource and you should ask what their transfer allowance actually is.
Key terms
- Bitrate
- The data rate of the encoded audio in kilobits per second. Total stream bitrate, not per channel — a 128 kbps stereo file is 128 kbps in total, roughly 64 per channel.
- Download
- An HTTP request that transfers the episode file. It is not the same as a listen: a request may be automatic, partial, or repeated, which is why measurement standards exist for audience claims.
- Back catalogue
- Every previously published episode still available in the feed. Individually low-traffic, collectively a large and permanently growing share of transfer.
- Decimal gigabyte
- 1,000 megabytes, the unit hosts and CDNs bill in. Operating systems generally report binary gibibytes of 1,024 mebibytes, about 7% larger.