Section 1
The mistake underneath the other mistakes
Almost every founder storytelling failure traces to the same root: the company arrives before the problem. The pitch opens with what we built, the homepage opens with who we are, the sales call opens with a slide about the team. Each of those is a company entering a conversation the listener has not yet agreed is about them. The corrected version is unglamorous. Open with the situation the listener recognises, stated in their vocabulary, including the part that is annoying and specific. Only after they have nodded does anything about you become relevant, because now it has somewhere to attach. Founders resist this because it feels like withholding. It is the opposite. You are earning the right to be listened to, and the cost of skipping that step is a listener who is polite and gone. A version of this applied to mission-led companies is in [Storytelling for Social Impact Entrepreneurs](/blog/storytelling-for-social-impact-entrepreneurs).
Section 2
Biography is not evidence
The second common error is turning the company story into the founder's autobiography. Origin stories are useful in one narrow way: they explain why this team has judgement a competitor lacks. Used for that, twenty seconds is plenty. Used as the main event, it asks a buyer to care about your life, which they do not. Related is the unquantified proof problem. We helped a client improve efficiency is not evidence, it is a sentence shaped like evidence. Either name the number and what it was before, or say plainly that you cannot share it and describe the mechanism instead. Buyers forgive a missing figure. They do not forgive a figure that dissolves when questioned.
Section 3
A defect list you can score
Treat these as defects rather than style preferences, because defects can be counted and fixed. The table below lists each mistake, the symptom that reveals it in a live conversation, and the specific rewrite that removes it.
Section 4
Rewriting one broken asset
Do not attempt a full messaging project. Take the single asset where the failure costs most, usually the first ninety seconds of the sales call, and fix that. Pull three recordings. Mark the exact moment the buyer's attention changes: the question that comes out of nowhere, the pause, the request to circle back to something you said earlier. That moment is where your story broke, and it is almost always the same moment across all three. Rewrite only that passage. Then run it live five times and mark the same moment again. If it has moved later in the call, the fix worked. If it has not moved, your diagnosis was wrong, and you have learned that cheaply. Repeat with the next asset. Sequential repair beats a rebrand, and it produces evidence a team will believe. Parallel thinking about which claims survive scrutiny is in [Common Myths About AI Automation Debunked](/blog/common-myths-about-ai-automation-debunked).
Section 5
The mistakes that only surface later
Some errors do not cost anything for a year. Telling five audiences the same story is one: it works while you have one customer type and fails silently as you add more, because the story now fits none of them well. Never retiring a story is another. The case study from the early days keeps getting told after the company has moved upmarket, and it quietly recruits the wrong buyer. The last one is having no ask. A story that ends without a specific next step leaves the listener with a warm feeling and no obligation, and warmth does not renew. A test to apply to every sentence: does it help the listener understand the problem, believe the path, or take the step. If it does none of the three, it is decoration, and decoration is what you cut first.