Demos, pricing and launch

Two versions of the same guitar phrase. Listen to both and pick the one you prefer before the names are revealed.

The only difference is 2 dB of level. If you picked the louder one, you're in good company: on a quick listen, a little more level tends to sound fuller, brighter and closer, and our ears read that as "better". Mastering engineers have argued about this for decades; it's the loudness war in miniature.

Now think about your demo. A listener on a product page can't match levels. They press play, hear "before" and "after", and believe whichever one wins. If your "after" is louder, they may be hearing the volume knob, not your instrument.

Here's the honest version: a drive on the same guitar, Dry against Drive, with a reading of how much louder the drive is. Lower Amount until the reading is near 0 dB, so loudness can't win the argument.

A demo is a promise you can hear

Think how fast you recognise a 303 line or an 808 kick: a second or two. Players judge a new instrument almost as fast. Your demo has about ten seconds to show its character, and those ten seconds are a promise: this is what you'll hear when you play it.

So give each demo one question. Does this drive make a quiet guitar part sit forward? Can this synth go from gentle to biting without changing presets? Does this bass stay clear under a busy kick? Then choose a phrase that shows the answer, and keep it the same before and after.

If the "after" version also changes the notes, adds drums and uses three other effects, you've made a production demo. Those are fine and often lovely, but they answer a different question. Say which one the listener is hearing, so nobody has to reverse-engineer your arrangement to work out what came from the product.

Match level by ear and by meter

Equal peaks don't mean equal loudness. A heavily compressed phrase can peak at the same level as an uncompressed one and still sound much louder, because its body is fuller. So when you set up a before and after, listen to the middle of the phrase, not only the first hit, and use a loudness meter if you have one (the level and loudness guide explains which readings mean what).

A matching tool helps, but it doesn't make a demo fair on its own. The phrase, the settings and the playback all have to be the same too. A matched comparison of two different takes is still a comparison of two different takes.

Let subtle stay subtle

Some instruments are good because their effect is controlled. A ducker made to gently clear room for a vocal shouldn't be demoed pumping like a festival anthem; a buyer who wanted the gentle version will feel misled, and a buyer who wanted the pump will be disappointed in their own mix.

If both uses are real, make two demos and name them: "everyday" and "pushed". Keep the music engaging (a memorable phrase, a clear change, one sentence on what to listen for) and let the listener judge the starting point and the extremes separately.

Price from what you know

Before choosing a price or how big to launch, write down what it costs to get the instrument to someone. Some costs happen once: making the instrument, the artwork, testing it on the systems you support. Some grow with every sale or every user: payment fees, hosting, storage, support time, and for instruments that generate audio on a server, the computing for each job.

Then label every number with where it came from:

• Observed: you measured it, in a situation like the real one.
• Agreed: an actual arrangement gives you this number.
• Assumed: a guess you plan to test. Not a fact about the market.

Here's a purely made-up example to see how the arithmetic moves. Suppose making the instrument costs 300 units, it sells for 20, and each sale costs 5 in fees and delivery. Every sale then contributes 15, so 20 sales cover the 300. Now change the per-sale cost to 10: you need 30 sales. The point isn't the answer; it's seeing which assumption moves the answer most. That's the one to test first.

Test the expensive guesses before you build on them. If installation is fiddly, watch someone install it alone before deciding support will be cheap. If a collaborator gets a share, use the agreed terms, not a placeholder. For taxes and payment-provider conditions, find the real documents that apply to you, and leave a number marked unresolved until you have it.

A smaller first offer

Moog's early synthesizers were large modular systems patched together with cables, mostly at home in studios. The Minimoog, from 1970, took a fixed selection of those parts, wired them together inside a portable keyboard, and needed no patch cables. It did less, and it went on stage with players who would never have toured a modular system.

A first launch can work the same way. A smaller bank, one supported setup, a limited test with a handful of people: these let you learn without promising work you haven't priced.

Go where the music happens

A channel is how someone first meets and gets your instrument: your own page, a demo in a community where people make that kind of music, a partner, a store. Start with where your intended players already talk about the job your instrument does. Then decide what you're asking them to do: listen, try one task, save a result or buy.

Attention isn't the same as use. A thousand plays of a demo video and three people who made a track with the instrument answer different questions (the finding an audience guide calls these lanes). Write down which one this launch is for.

And keep the facts straight in public. A store listing isn't a distribution deal, and an interested partner isn't a collaboration until you both agree. For a first launch, one clear route is much easier to learn from than four routes with different demos.

Decide when to stop

Before you launch, write one question the launch will answer. For example: "Can ten invited players install it, make a short part and reopen their saved sound without help?" That's a question about use, not a sales forecast.

Add how long it runs, what you'll spend, who it's for and what you'll watch. Then decide in advance what result means "try again", "fix this first" or "pause this idea". Keep it small enough that a problem has a name: a missing file, a misleading demo, a setting that doesn't come back.

Where it breaks

The demo is better than the instrument. Extra processing, a session player and a big mix make the demo shine, then the buyer's own phrase sounds plain. List every outside effect in the demo's description.

The price is built on hope. A sheet full of assumed numbers looks precise. Circle the assumed ones and test the biggest first.

The launch is louder than the learning. A big campaign with nothing measured afterwards leaves you with impressions and no next step. A small launch with one question leaves you with an answer.

Make it yours

Make one ten-second demo of your own instrument: the phrase dry, the same phrase through the instrument at matched level, then in a short piece of music. Write under it the settings and any outside effects. Fill in the cost sheet above with your real lines, and write your one launch question with its stopping point. Play the demo to one person from your intended audience and ask them what the instrument does. If their answer includes a promise you didn't mean to make, change the demo.

To practise the habit on something small first, open Drive, put any phrase through it, and see whether a friend can still pick the processed version once the levels match.

Quick check

1. Your "after" demo wins every time you play it to friends. It's also about 3 dB louder. What do you know? Not much yet. A few dB of level can win a comparison by itself, as the first player on this page showed. Match the levels and play it again; whatever still wins is your instrument.

2. Your cost sheet says you'll break even at 40 sales. Is that a sales forecast? No. It's arithmetic on your assumptions. It tells you which numbers matter most, so you know what to measure next. Whether 40 people buy it is a separate question your launch test can start to answer.

3. Your demo video got 5,000 views and two people made a track with the instrument. Which number tells you more about whether the instrument works? The two tracks. Views say the demo caught attention; finished tracks say someone used the instrument for its job. Both are worth knowing, but only one tells you the instrument survives contact with a stranger.

Next steps

• Use it: Rights and maintenance. Someone bought it. What do you owe them a year from now, and are you allowed to sell every sound inside it?
• Go deeper: Limiting and loudness. Why louder wins a first listen, and what a fair comparison needs.
• Sideways: Listening and comparing. The same blind, matched listening you used here works on your own mixes.