Sound like yourself, only clearer: building an on-air personality that isn't fake

Every host gets the same two notes at some point in their career, usually from different people, usually in the same month:

“Be yourself.”

“You need more energy.”

Follow both literally and you get the thing everybody in this business can spot in four seconds: a person doing an impression of a radio host. Too big, too smooth, too pleased about the weather.

The two notes aren’t actually in conflict. They’re just badly worded. What they mean is: be yourself, and be legible.

Why the fake voice happens

Nobody decides to sound fake. It arrives on its own, from three places.

The microphone. A live mic makes people perform. You can hear the exact moment it opens in most inexperienced reads, because the pitch rises and the pace picks up.

The absence of a face. In conversation you get feedback constantly — a nod, a laugh, a look. On air there’s nothing, so people compensate by pushing harder, and pushing harder is what “fake” sounds like.

Imitation. Everyone starts by copying somebody. That’s fine and normal, and it’s supposed to be a phase. The problem is that a borrowed personality is brittle — it works until something unexpected happens, and then there’s nobody underneath it to handle the moment.

“Be yourself” is incomplete advice

You at a dinner table isn’t broadcastable. There are digressions, false starts, inside references and long stretches where nothing happens. Radio needs a version of that with the dead weight removed.

The useful formulation:

You, turned up about fifteen percent, with the boring parts cut out.

Not a character. Not a persona with a different name. The same person, edited.

Fifteen percent, not two hundred. The most common overcorrection is a host who was told to have more energy and came back as a carnival barker. Energy isn’t volume — it’s investment. Somebody genuinely interested in what they’re saying reads as high energy at conversational volume. Somebody shouting about something they don’t care about reads as noise.

What actually sounds weak

Hosts worry about sounding weak and then almost always misdiagnose it. It’s rarely volume. It’s these:

Hedging. “I don’t know, I just thought it was kind of interesting, maybe.” You’ve pre-apologized for the content, so nobody has any reason to invest in it.

Asking permission. “Is it okay if I tell you a story?” It’s your show. Tell the story.

Over-explaining the setup. Three sentences of context before the thing happens isn’t politeness, it’s fear that the story isn’t good enough on its own. Trust it or cut it.

Laughing at your own line first. It reads as please find this funny. Say it flat and let it land. If it doesn’t land, move.

Trailing off. The sentence that fades out instead of ending. This one is almost universal and almost invisible until you hear yourself do it.

None of these are diction problems, and it’s worth being sure which you’ve: if the words themselves are arriving mangled, that’s a warm-up problem, not a personality one.

Apologising for existing. “Sorry, I’m rambling.” “Anyway, nobody cares.” Once, self-deprecation is charming. As a habit it teaches the audience that you’re not worth listening to, and they’ll believe you.

Fixing those six things does more for perceived strength than any amount of volume, and none of them require you to be a different person.

What actually sounds strong

Specificity. “Weird weather out there” is nothing. “It was 41 degrees when I left the house and it’s 68 now” is a person who was outside. Detail is the raw material of personality — generic statements have no author.

A point of view, cheaply held. Not politics; preferences. The correct way to load a dishwasher. That the drive-through line at that place isn’t worth it. Small, defensible, and something a listener can disagree with while grinning.

Ending sentences. Land it, then stop. Silence isn’t an emergency. The willingness to pause is the single most audible marker of on-air confidence, and it costs nothing.

Being wrong in public, briefly. “I said that wrong yesterday, here’s the right version.” Two sentences, no grovelling. It builds more credibility than being right does, and it’s the thing no automated voice can do.

Consistency. The same person, every day, in the same chair. Recurring bits, recurring opinions, an ongoing saga. Personality is largely just accumulated familiarity — which is why the temptation to reinvent every quarter is the enemy of ever having one.

The one-listener rule

Don’t talk to “everybody out there in radio land”. Nobody has ever been in radio land.

Picture one person. Ideally a real one — a friend who fits your format, in their car, on the way to work. Talk to them. Say “you”, not “everyone”.

This single change fixes the announcer voice faster than any other adjustment, because it’s impossible to be pompous at one specific person you know. It also fixes pacing, word choice and the impulse to over-explain, all at once.

Three drills

1. Tell it to a friend first.

Before you open the mic, say the story out loud to an imaginary friend, in kitchen language. Notice which words you used. Those are the words that go on the air — not the ones you’d have written down.

Written language and spoken language are different, and reading written language aloud is a large fraction of what “fake” means.

2. Record yourself and listen at the wrong time.

Not immediately, when you’re still defensive. The next day, in the car, cold.

Listen for one thing at a time — this week trailing sentences, next week hedging. Trying to fix everything at once produces a host who’s thinking about being a host instead of being one.

3. The line test.

Take any line you’re about to say and ask: would a person say this?

“Coming up, we’ve got a great show for you.” No person has ever said that in real life. “Stay put, because the thing that happened to me at the DMV is worse than you think” is a sentence a human being would say.

Where the line is on personal material

Personality needs real material, and real material comes from your life. But there’s a difference between a host who’s a person and a host who’s a therapy session.

The rule that works: share the thing you’ve finished processing, not the thing you’re still in.

The argument from last year is a story. The argument from this morning is a liability. Something you’ve already made peace with, you can shape, land and end. Something raw runs the show instead of you running it, and the audience can hear that they’re being used.

The same applies to other people. Your family, your co-workers and your callers didn’t sign up to be content. Ask, or change the details, or leave it out.

If you’ve a co-host

Personality is much easier in a pair and much easier to get wrong.

Divide the roles. Somebody starts the story, somebody reacts. Somebody is the optimist. Two people doing the same job is why a lot of pairs sound like noise.

Let each other finish. Talking over a partner is the most common reason a show sounds chaotic rather than energetic.

React honestly. The fake laugh is more damaging in a pair than solo, because there are two people to disbelieve.

The thing to remember

An audience isn’t choosing between you and a better version of you. They’re choosing between you and the silence, the podcast, the playlist, and — with each year — a synthetic voice that will read the liner more evenly than either of us.

Nobody in that list can be a specific person from a specific town who was outside this morning and has an opinion about the dishwasher. That’s the whole of the argument in what an AI host can’t do at 6:40 on a Tuesday, at greater length.

Be that, on purpose, every day. That’s the whole job.