Every pitch competition hands its judges a sheet. Team, market, product, traction, clarity, and sometimes a box for "investability" that nobody can define. You would think the winner is the founder with the highest total. Often the sheet is finished afterwards, to match a decision the table has already made.
The sheet and the table
A judging session runs like this. Eight, twelve, sometimes twenty pitches in a row. Notes in the margin that get shorter as the afternoon goes on. Then the judges sit down together and somebody asks: who stood out? From that moment memory matters more than the sheet. The founders who get discussed are the ones a judge can describe without looking down. "The one with the cold chain sensors." "The founder who opened with the hospital bill." If nobody can describe you in one sentence, you are not in that conversation, whatever you scored for market size.
What survives a day of pitches
Three things survive, almost every time.
One word. Not a tagline, a word. The thing the judges call you when you are not in the room. If you do not choose it, they will, and it will be your weakest detail.
One number. Founders bring five numbers to show they are serious. Judges remember one, and it is rarely the one you would have picked. Choose it for them: the number that proves the problem is real, or that customers already pay.
One ask. A panel trusts a founder who knows what they want: a pilot, an introduction, a first check. An ask that is one sentence says you know where you are going. A paragraph says you are still deciding.
What to cut
Most of what founders add to a competition pitch is there to defend against a question nobody has asked yet. The competitor slide with fourteen logos. The roadmap to 2030. The second use case. Each one is reasonable, and together they bury the word, the number and the ask. Cut anything the judges will not repeat at the table. If they want it, they will ask in the questions, and then it is an answer instead of a slide.
The first ten seconds
By the afternoon judges are tired, and they decide quickly whether to look up from the sheet. Your first ten seconds are not the place to introduce yourself or thank the organizers. Start with the problem, a person, or a number that should not be true. The name of your company can wait until they want to know it.
The body does half the work
Judges watch more than they listen. A founder who stands still, looks at the panel and lets a sentence land reads as someone who will survive a board meeting. A founder who rocks, rushes and reads from the screen reads as someone who might not, whatever the slide says. That part is craft, not talent, and it can be practiced.
Practice it out loud, to people
A pitch that has been written well and said out loud twice is a pitch that comes apart on stage. Reading it on the train is not practice. Give it standing, with a timer, to someone who will tell you afterwards what they remember. If they repeat your word, your number and your ask, you are ready. If they repeat something else, that is what the judges will remember too.