Part IV — Sell direct

Selling access, not audio

Design an offer someone will actually pay for, given that the recording itself is already free to them everywhere.

Start from what is already free

Your listener can hear your record right now, in full, at no cost, on a service they already pay for. Every direct sale in this Part has to survive that. If your offer is "the same audio, but you pay for it", their refusal is not apathy — it is arithmetic.

One framing in the source material is worth keeping: streaming did not take your audience, it absorbed the cheap part of it. Everyone who would once have taken the music for nothing now takes it for a subscription that never reaches you in a meaningful amount. What is left is a smaller group who were always willing to pay you specifically. They do not want more music. They want something the stream does not contain.

So: the stream is the free sample; the product is the thing that has a price. The record is the proof that you are worth buying from. It is not the thing being bought.

What people are actually buying

The most useful model in this library names three things, and insists they stack in order.

Access — how much of your world the thing actually contains. Not "exclusive" as a word on a button: what does the buyer now know, see, hear or get into that a streaming listener does not? The test is three questions asked before you make it. How much insight into you does this contain? Is that insight worth the price? On a scale of one to ten, how good is the experience? None of those are about materials or margin. A product that fails them is one anyone could have made.

Acknowledgement — being recognised as a specific person. Replying to comments from the very beginning. Writing to the email list instead of merely collecting it. Naming individuals. The claim worth taking seriously here: acknowledgement degrades silently as you grow. Early on you are seeking recognition; past some point you are expected to give it, and nothing tells you when you crossed over. Build the habit before you are big enough to need it.

Achievement — a status only some fans hold, visible to the others. Limited runs, a members-only tier, a meet-and-greet, a credit by name. The point is not the perk inside the tier. The point is being the kind of person who is in it.

The order is load-bearing. Achievement is only wanted by people already given access and acknowledgement. Sell exclusivity to strangers and you have priced a rank in a hierarchy nobody has been admitted to. It reads as a paywall, because that is what it is.

The same three words appear elsewhere as a test you can run over any product before you build it: does it acknowledge the specific culture you built with your fans, does it grant access deeper than the stream, does it confer an achievement. A plain digital download scores zero on all three.

The other half of the argument, from a different angle: nobody buys an object, they buy the state they expect to be in afterwards. So the design question is not "what can I make" but "what are people already doing while this plays, and what would help them do it better". Answer that and you have a product brief. Skip it and you will conclude, as most artists do, that the answer is to release more music.

The menu

Two of the sources give a product typology, and they overlap enough to treat as one menu:

Three worked shapes are offered for one release: a packaged album gathering the lyrics, credits, thank-yous and extras you currently scatter across social media into one bought object; a ticketed listening session in a room with a good system, the package handed over at the door so people can read the credits while the record plays; and a paid teardown of how it was made. Three income streams off one release, from material you are already producing.

One warning attached to that: giving the components away as free posts does not lose them, it spends them. Nobody buys an assembled version of what they already scrolled past.

⚠️ Two numbers here do not survive. The video proposing these three is titled around making "10 times more", and the speaker retracts it on camera mid-argument: "well not 10, maybe double or triple." The same video puts a $50 minimum on the package with no source. Take the structure; leave both figures.

Why the offer has to exist before you buy traffic

One source lays out the full sequence: product, free samples, word of mouth, packaging, market, storefront, promotion, offer, sale, upsell. The reason to write it out is that streaming lets you skip nine of those steps. You can release for years and never build the ability to sell anything — and you will not notice until the day someone has to buy a ticket rather than press play.

That is why the offer comes before the spend. An ad that lands a curious stranger on a page with nothing to buy has paid to disappoint them. The reach is real; it just has nowhere to go.

Pricing is a method, not a number

Here is the discipline, and it is the durable part: know the price of the unit and the cost of producing it, and every goal becomes a division problem.

Pick a number you want. Divide by the price of one unit. That tells you how many transactions you need. Ask whether that many is plausible given the people who currently know you exist. If not, you have two levers and only two — charge more, or make the thing worth more. The chain from audience to money runs through several multipliers (how many see it, how many of those buy), each a separate number with a separate remedy.

The arithmetic is often most valuable when it tells you to reject the route. One source prices a million dollars of streaming income, finds it absurd, then prices the same million as a far smaller number of direct sales. The rejection is the finding.

⚠️ Do not carry any of the inputs. The per-stream rates in this library disagree with each other by an order of magnitude. A production cost is spoken one way and implied another in the same passage. A royalty haircut defers its derivation to a paid course. One headline promises "$100 a day" and the video never does the arithmetic — no price, no unit count, no conversion rate. Another promises a "2000%" boost never said aloud anywhere in its own body. A third states a segment model as 100% / 5% / 1%, then works an example computing to 80% / 19.5% / 0.5%. The method survives a wrong input. A quoted number does not.

The cannibalisation rule

If you build tiers, one rule governs them: never give a higher tier's experience to a lower tier's ticket.

The moment the person who paid least gets what the person who paid most bought, the higher tier has no content left. The cost falls on your best customers, who conclude their extra spend bought them nothing anyone else did not get. Having no top tier at all does the same thing more quietly: it tells your most committed people there is nothing further to be committed to.

The concert ladder is the clean illustration — nosebleeds, mezzanine, floor, front row, group meet-and-greet, personal meet-and-greet. Each rung gates something the rung below cannot reach. (The prices attached to the last two in the source are the speaker's illustration for a hypothetical tour, not market rates.) Design each tier with an answer to "what does this deny to the tier below", and start at a size you can actually fulfil.

Two live examples from this network's first cohort

Mental Stamina has three release tiers fully specified — standard, deluxe, and a collector edition capped at 100 — wired end to end through cart, checkout and download provisioning. The design work this chapter describes is done. The tiers gate different things. The collector edition is genuinely limited. And no storefront UI renders any of it, so no customer can reach any of it. That is this Part's exact failure mode: the argument is finished and the door is missing. The pipeline runs product, sample, promotion, offer — and stops at the step that moves money.

Mozay Calloway's store is deliberately closed until a payment provider is wired, and the catalogue stays browsable meanwhile. That is the correct version of not-yet-selling: nothing is advertised that cannot be delivered, the free route stays open, and the decision actually outstanding is the one being worked. The difference between the two cases is not effort. One of them is honest about where it is in the sequence.

What to do this week

1. Run the three-part test on whatever you sell or plan to sell. One sentence each: what it

acknowledges, what access it grants beyond the stream, what achievement it confers. Anything scoring zero on all three does not ship as the flagship.

1. Answer the use-case question in writing. What are people doing while your music plays? One

paragraph. That is the brief for your first non-audio product.

1. Do the division once. Take a number you need this year, divide by a price you would charge,

look at the transaction count. Decide whether to change the price or the product, and write down which and why.

1. Audit any tiers against the cannibalisation rule. For each, name the one thing it denies to

the tier below. If you cannot name it, that tier is decoration.