Business model
Where revenue comes from, what it costs to earn, and how the first customers are found. Read Music business first — the per-stream arithmetic there is what rules out half of the obvious options.
Everything on this page is a hypothesis until phase 4 of the validation plan. Pricing in particular should be tested, not reasoned about.
The five candidate revenue lines
| Line | Who pays | When it can start | Assessment |
|---|---|---|---|
| Artist subscription | The artist, monthly | Phase 4 | Start here. The market already pays for these tools, the amount is predictable, and it does not depend on listener scale |
| Distribution fee | The artist, per release or as a share | Phase 4 | Natural pairing with the subscription. Directly comparable to incumbents, so pricing is constrained by them |
| Marketplace commission | Buyer and seller | Phase 6 | Good margin, but needs liquidity on both sides and drags in regulated payments |
| Listener subscription | The listener | Phase 6+ | Requires a catalogue and scale Bitrate does not have. Do not plan on it |
| B2B / API | Labels, other platforms | Phase 6+ | Only once the platform exists and is worth integrating with |
Do not launch more than one. Two pricing models at once make it impossible to learn which of them works, and multiply the legal surface.
Why artist-paid, and not listener-paid
Listener subscriptions require licensing a catalogue Bitrate cannot license, at a scale it does not have, against companies who did it a decade ago. Artist tools require none of that: the customer has the problem today, already pays someone for a partial solution, and can be reached one at a time.
It is also the only model where the first euro is reachable in phase 4 rather than year three. That single fact is why the whole strategy is artist-first.
Pricing hypothesis
The reference points an artist already knows: distribution runs roughly €20–€30 a year for unlimited releases, or a per-release fee. Anything Bitrate charges is measured against that number whether or not the comparison is fair. Those figures are from general knowledge and were not checked against each provider's current pricing — verify before using them in a pricing decision.
A defensible starting shape:
| Tier | Roughly | For |
|---|---|---|
| Free | €0 | An artist page, and the workspace without distribution. Gets them in |
| Artist | €10–€20 / month | The release workflow, distribution, analytics, AI insights |
| Pro | Higher | Multiple artists, collaborators, advanced automation — a manager or small label |
Three things this shape gets right. The free tier creates artist pages, which are the listener-acquisition channel that costs nothing. The paid tier is priced against the workflow, not against distribution alone, so it is not trapped in a race to €20/year. And a higher tier exists for the customer who has more than one artist, who is both easier to sell and worth more.
Test it rather than deducing it. Phase 1 interviews should establish what artists pay today and for what; phase 4 should test at least two price points.
Unit economics
The vocabulary that has to become second nature — with the caveat that no number here can be filled in yet, because there are no customers. That is precisely why they are worth setting up now: the shape of the model determines which experiments are worth running.
| Metric | What it means here |
|---|---|
| ARPU | Average revenue per artist per month |
| CAC | Fully loaded cost to acquire one paying artist — including founder time |
| LTV | ARPU × gross margin × average lifetime in months |
| Churn | Share of paying artists lost per month. In a per-release product, seasonality is severe |
| Gross margin | Revenue minus the cost of serving it — storage, egress, transcode, inference, distribution partner |
| Payback period | Months of margin to recover CAC. Below 12 is healthy for this kind of product |
| Burn / runway | Monthly spend, and how many months of it remain |
The costs that scale with use
Unlike most SaaS, this product has real marginal cost, and three of the four are unusual:
- Storage — masters plus every rendition, kept indefinitely. Grows monotonically.
- Egress — audio delivery is the largest variable cost at scale, and today there is no CDN.
- Transcode — CPU per upload, three encode passes, and currently encoded twice over.
- AI inference — per artist, per release. Must be a designed constraint, not a discovered bill.
An artist who uploads a large catalogue and never converts costs real money. That is a reason to bound the free tier's storage, not a reason to have no free tier.
The one number to compute early
Cost to serve one artist for one month, at current infrastructure prices. It can be computed today, from the existing stack, without a single customer — and it sets the floor under every pricing conversation that follows.
Go to market
The sequence, and the reason each step is shaped the way it is:
| Stage | How | Why |
|---|---|---|
| First 10 | Direct outreach by the founder. Walk each one through a release by hand | This is concierge validation, not sales. The product design is the output |
| To 50 | Referrals from the first 10, plus continued direct outreach | If the first 10 do not refer, that is the finding — stop and fix it |
| To 100 | Content and case studies built from real results | Requires results, which requires the first 50 |
| To 1,000 | Creator marketing, community, partnerships with studios and schools | Only once activation and retention are known |
| Listeners | Artists bring their own audience via their Bitrate pages | Free, and unavailable to a pure streaming competitor |
The loop to build — and the only one that compounds:
artist releases → content and results → listeners arrive → a visible case study → new artists
Where the first artists actually are. Independent artist communities, production and mixing forums, music schools, small studios, beat marketplaces, and local scenes. Poland and the surrounding region is a legitimate first market rather than a limitation: it is reachable in person, underserved by tools built for the US market, and small enough that ten artists is a visible presence rather than a rounding error.
Funding
The honest position: bootstrap through phase 4. Everything up to the first hundred artists is reachable on a single VPS and founder time, and money raised before there is evidence is the most expensive money available.
Understand the vocabulary before it is needed — cap table, dilution, valuation, SAFE and convertible instruments, pre-seed and seed, the deck, due diligence — because learning it during a negotiation is a bad position. See Founder skills.
The moment external money becomes rational is when there is a repeatable acquisition channel and the unit economics work — that is, when capital buys more of something known to work, rather than funding the search for it.
What would falsify this model
Stated explicitly, so the evidence is recognisable when it appears:
- Artists complete one release and never return → the product is a tool, not a workflow, and the subscription is wrong.
- They will pay for distribution but not for the workflow → Bitrate is a distributor, and competes on price with entrenched incumbents.
- CAC is dominated by founder time and does not fall → there is no scalable channel, and the business is a consultancy.
- Cost to serve exceeds what artists will pay → either the free tier or the infrastructure has to change before anything else does.
Any one of these appearing in phase 4 is worth more than another year of building.