Finding an audience

In 1981 Roland released a small silver box called the TB-303 Bass Line. It was built to stand in for a bass player: program a line, press play, and practise or perform over it. The people it was made for compared it with a real bass guitar, and it lost. It sold poorly, Roland stopped making it after a few years, and second-hand ones turned up cheap.

A few years later, producers in Chicago did something the box was never sold for. They let a pattern run and turned the filter knobs while it played. The bass line started to squelch, whine and talk. Phuture's "Acid Tracks", released in 1987, was built around that sound, and a whole genre, acid house, took its name from it.

The synth below starts close to what the 303 was sold as: a round, polite bass note.

Before you turn anything, a prediction. If Roland had played this box to a hundred guitarists in 1981 and asked "do you like it?", what would they have said? And would their answers have told anyone about Chicago?

Same box, two audiences

The instrument didn't change between 1981 and 1987. What changed was the thing it was compared with.

A guitarist who wanted a backing bass held the 303 up against a friend who plays bass, or against recording the bass part themselves. Next to those, a buzzy little box with a fiddly step programmer was a poor substitute. A producer making tracks for DJs held it up against the drum machines and sequencers already on the table, and nothing else there could make a bass line move like that while it played. Same sound, opposite verdicts.

Roland's TR-808 drum machine has a similar story. Its drums didn't sound like a real kit, which disappointed people who wanted realism. Then that booming kick and snappy clap became the sound of electro and hip hop: Afrika Bambaataa's "Planet Rock" and Marvin Gaye's "Sexual Healing", both from 1982, are built on it.

Your instrument is always judged against what someone does now. Finding an audience means finding the people for whom your instrument beats their current way of making that sound, and checking that it really does.

Why "it sounds great" tells you so little

Here are two comments a tester might give you. They're made up, but you'll hear both:

• "That sounds great!"
• "I tried it on yesterday's bass part, but went back to my old plug-in because I couldn't get the slide to land on the beat."

Which one tells you more about whether that person will use your instrument next week?

The second, by a long way. It names a session (yesterday's bass part), a job (a sliding line), an obstacle (timing of the slide) and the thing you're competing with (their old plug-in). You can go and fix that. The first comment may be completely sincere, but it can't tell you whether they'd ever open the instrument again.

Praise answers a different question from use. It helps to keep four moments apart, because each one tells you something new:

Friends and waitlist sign-ups mostly live in the first lane. They're lovely and they're useful for reaching people later, but they don't tell you about the other three. Don't let "everyone loved it" turn into "we have customers" in your head.

Find the person in the session

"Producers" is too broad to tell you which obstacle matters. Someone sketching a song on a laptop at midnight and someone preparing a hardware live set are both producers, and they'd want different things from the same bass synth.

Try writing one sentence with three parts: for this person, in this session, it helps them do this.

• The 303 as sold: "For a guitarist practising alone, a bass line to play along with."
• The 303 as loved: "For a producer making a track for the club, a bass line they can reshape live while the pattern runs."

Both sentences are clear, and they describe two different instruments. The first wants realism and easy programming. The second wants big filter knobs, resonance that screams, and a sound that changes every bar. When you write your own sentence, you're choosing which instrument you're finishing.

Add what the session demands: a laptop or a hardware rig, a quick setup or a long one, whether the sound has to be saved and recalled next week. A beautiful sound that needs a different machine, or loses its settings, still loses to the old way.

A narrow first audience doesn't stop you reaching others later. It gives your first test a clear comparison, which makes the result easier to read.

Ask about the last time, not the future

People are bad at predicting what they'll use, and good at remembering what they did. So start with a recent occasion: "Tell me about the last time you made a bass line you were happy with."

Then follow the story before you show them anything. What did they use? Where did the time go: finding a preset, programming the notes, fixing the tone? Did the part make it into the finished track, or did they give up and use something else?

That story gives you a baseline. Your instrument has to improve on it in a way they'd notice: a better sound, less time, more control, or a new musical idea they couldn't reach before. It doesn't compete with an empty screen.

Feature requests are worth hearing, but treat them as clues, not proof. "I'd use it if it had a sequencer" is interesting. "Last week I spent an hour programming slides by hand" is a problem. Ask which recent session the feature would have changed.

Watch someone play it

Give a willing tester one task that looks like their recent session ("make a two-bar bass line for this drum loop"), then stay quiet. Watch the path: getting a first sound, shaping it, saving it. Notice where they hesitate, and what they reach for when they're stuck.

The hardest rule: don't help. If you take over the tricky step, you've tested yourself. An obstacle they hit is exactly the information you came for.

Afterwards, ask how they'd normally have done it, and which way they'd choose next time. Sometimes the slower way wins because it gives them control they value. That's worth knowing too.

Three people won't tell you how big your market is. They will show you whether the same obstacle keeps appearing, and that's enough to decide the next change.

Where it breaks: testing in the wrong room

Friends are kind. Other builders love clever techniques they'd never use in their own music. Someone outside your audience may dislike a choice you made on purpose for someone else. None of them is giving you bad feedback; the mistake is treating every reaction as if it came from the same person in the same session. Write down who tried it and what they were making.

Make negative comments specific. "It took too long" could mean loading, finding a sound, editing or recovering a lost setting. Ask which, and you have something to fix. Do the same with praise: "the soft notes left room under the vocal" tells you a strength to protect.

Now the 303 lesson in reverse. Sometimes the person outside your intended audience is the real one. If a techno producer takes your gentle ballad piano, pushes it into distortion and keeps coming back to it, don't file that under "wrong user". Ask what they were making. You may be looking at your Chicago.

And sometimes the honest result is smaller: narrow the promise, fix the first minute, or stop that particular idea. A test is worth running when its result could change what you do next.

Make it yours

Write your one sentence: this person, this session, this job. Under it, write what they use today and three things you'd need to see to call a test a success, such as "made a part without help, saved it, and used it in their track". Then run the worksheet above with three people who match the sentence, and finish with one decision: fix a named obstacle, narrow the audience, or test the next open question.

If you'd like to feel the 303's two lives in your own hands first, open Acid Bass, find the setting you'd play to a guitarist and the one you'd play to a DJ, and save both.

Quick check

1. A tester says "I'd definitely use it if it had a sequencer." What do you ask next? "When did you last need one, and what did you use?" A feature request tells you what they imagine. A recent session tells you whether the problem is real and how often it comes up. If they can't remember a time, the sequencer may be a nice idea rather than a reason to switch.

2. A friend loved your demo. A stranger from your audience made a bass part with it and called it "fine". Whose reaction tells you more about next week? The stranger's. They're in the session you're designing for, and they finished a real task, which is a later lane than liking a demo. The friend's reaction still means something: the sound caught an ear. It answers an earlier question.

3. You built a soft keys instrument for singer-songwriters. A drum and bass producer is using it, heavily distorted, every day. Is that a mistake to ignore? Not necessarily. That's how the 303 found its audience. Ask what they're making and what they used before. If the instrument beats their old way, you may have found a second audience, or a better first one. It's still worth checking that your original audience gets what they need.

Next steps

• Use it: Identity and story. Some testers will care because of where a sound came from, not only what it does. When does a story help them play, and when is it a costume?
• Sideways: Finishing and references. You judge a mix against a reference track, not against silence. An instrument is judged the same way, against the tester's current workaround.
• Use it: Interface and first use. Fix the obstacle your task tests kept finding in the first minute.