How to Practise Your Startup Pitch Out Loud

Why rereading the deck does not work

Reading slides silently rehearses recognition - you get better at recognising your own content. It does not rehearse retrieval under pressure, which is the actual task in the room, and the two are close to unrelated as skills.

It also hides length. A pitch that reads in forty seconds routinely takes seventy-five to deliver, because reading skips the pauses, the reformulations and the moment where you decide whether to include the second metric. Founders discover this in the meeting.

And it produces no signal you can act on. You finish the reread feeling either fine or vaguely anxious, neither of which tells you what to change.

The four numbers to check

Length. Time the sixty-second version. If it runs past seventy-five seconds you do not have a sixty-second pitch, and the fix is cutting words rather than talking faster.

Pace, measured in each half separately. The signature of an under-rehearsed pitch is 140 in the first half and 175 in the second. The average looks acceptable and the delivery does not.

Filler rate. Fillers cluster before numbers, which in a pitch means they cluster on your traction slide - precisely where you least want hesitation. Check where yours fall rather than only how many there are.

Pause placement. There should be a full stop after your one-line description of what the company does, and another before your ask. Both are moments the listener needs to process, and founders routinely run straight through them.

Rehearse the structure, not the words

Memorised pitches fail in a specific way: the first interruption destroys them, and investor conversations consist of interruptions. Once the sequence breaks, a founder who memorised words has nothing to fall back on.

Rehearse instead as five beats you can deliver in any order: what it does, who has the problem, why now, what the traction is, what you are asking for. Practise each beat as a standalone twenty-second answer, because that is how they will actually be requested.

Then practise the joins. A large part of what reads as polish is the transition between beats, and transitions are the first thing to disappear when you rehearse the beats in isolation.

Keep the wording loose deliberately. A pitch delivered in slightly different words each time sounds like someone who knows their company; a pitch delivered identically sounds like a recitation, and recitation flattens pitch range and eliminates pauses.

The one-line description is the whole game

If the listener cannot repeat what your company does after one sentence, nothing later in the pitch is being evaluated on its merits - they are still working on the first question while you deliver the rest.

Test it directly. Say the sentence to someone outside your sector and ask them to say it back. Not whether they understood it, which people say out of politeness, but what they think it is. The gap between the two is your real problem statement.

Delivery matters here more than anywhere else in the pitch. This sentence needs a full stop after it. Founders almost universally run it into the next sentence, which removes the moment the listener needed to actually absorb it.

A week of rehearsal

Days 1 to 2: record the full sixty-second pitch daily without correction. Note length, pace by half, filler rate and where the fillers land. You are looking for the pattern, not a good take.

Days 3 to 4: cut. Almost every pitch at this stage is over length, and cutting words is the only fix that survives contact with nerves. Target the number of words your slot actually holds - roughly 140 for sixty seconds at a sustainable pace.

Day 5: beats in isolation. Twenty seconds each, in random order, unrehearsed sequence.

Days 6 to 7: interruptions. Have someone break in with a question partway through, and practise answering it and returning. This is the rehearsal that most closely resembles the actual meeting, and the one founders skip.