How to Handle Mistakes During a Speech
You said the wrong word. You lost your place. Your mind went completely white. What you do in the next three seconds matters far more than the mistake did — and it is a skill, not a personality.
Everybody plans the talk. Almost nobody plans the mistake.
Which is strange, because one of those two things is guaranteed. Mistakes during a speech are not a sign that something went wrong — they are a normal feature of people talking, including the people you admire. What separates speakers is not who makes fewer. It is who recovers faster, and that part is learnable in an afternoon.
OneYour standard is wrong
Start here, because most of the suffering comes from measuring yourself against something that does not exist.
A team of researchers led by Heather Bortfeld recorded ordinary people having ordinary conversations and counted the stumbles: the ums and uhs, the false starts, the repeated words, the sentences abandoned halfway and rebuilt. Not learners — adults speaking their own language, comfortably.
The rate came out at roughly six per hundred words.
Six stumbles in every hundred words. In your first language. On an easy topic. Relaxed. That is what normal human speech is made of.
You are not failing to reach a standard. You are measuring yourself against a standard nobody has ever met.
And it gets worse for the perfect-speech theory: the same research found that disfluency went up when the topic was harder to describe. Stumbling is a signal that your brain is doing something demanding. Which, if you are presenting in a second language, it certainly is.
TwoWhat “uh” is actually doing
Here is a genuinely surprising one.
Jean Fox Tree tested what happens to a listener when a speaker says “uh.” The expectation was interference — a filler is noise, noise should get in the way. The result went the other direction: hearing “uh” helped listeners recognize the word that came next.
Be careful with this, because the finding is narrower than the internet version of it. “Uh” helped. “Um” did not — it neither helped nor hurt. Fox Tree suggested the difference is timing: “um” tends to come before a longer delay, and listeners may drift during a long wait.
So the honest takeaway is not “fillers are good, use as many as you like.” It is this: a short filler before a hard word is not damage. It is a small signal that something is coming, and your listener uses it. Removing every single one from your speech is not a useful goal.
ThreeThe blunder that makes people like you more
In 1966, Elliot Aronson, Ben Willerman and Joanne Floyd ran an experiment that has been misquoted ever since. People listened to a recording of someone answering quiz-show questions. In some versions the performer was excellent; in others, mediocre. And in some versions, at the end, the performer clumsily spilled a cup of coffee.
The excellent performer who spilled the coffee was rated more likable than the excellent performer who did not. A small human failure made a strong performer easier to like. It has been called the pratfall effect ever since.
Now the half that usually gets left out, and the half you need.
When the mediocre performer spilled the coffee, it did not help. It made things worse.
Read that carefully, because it changes the advice completely. The stumble did not create the warmth. The skill created the warmth, and the stumble released it. People like a well-prepared speaker who slips a little.
So the real advice is simple, if not very exciting: prepare properly, and then stop being afraid of the small slip. Preparation is what earns you the right to be forgiven for it.
FourWhere a mistake costs you least
One of the oldest findings in memory research is the serial position effect, demonstrated by Bennet Murdock in 1962. Give people a list to remember and they do not remember all of it equally. They reliably remember the first few items and the last few items. The middle is the part people forget.
Murdock used word lists, not speeches, so treat the next step as sensible implication rather than proven result. But the implication is hard to argue with: your opening and your closing are the parts that survive. The middle is the cheapest place in your talk to fumble.
Two practical consequences.
- Over-rehearse the first thirty seconds and the last thirty seconds. Those are the parts that get remembered and the parts you deliver at peak nerves. The middle can be looser.
- If you stumble, do not let it become the ending. A mistake followed by a clean, calm finish is remembered as a clean, calm finish. Your recovery happens after the mistake, so it is the newer memory. That works in your favor.
There is one more piece of good news you have probably already been given but not believed: the audience can see far less of what you are feeling than you think. That has been measured too, and it is the subject of our article on controlling nerves before a presentation.
What to say in the next three seconds
One line each. Short on purpose — you have to be able to say it while your brain is still catching up.
- Wrong word
- Sorry — I mean …
- Not … — … That’s the one.
- Wrong number or fact
- Let me correct that. It’s actually …
- Lost the thread completely
- Let me come back to that in a moment.
- The point I want you to take away is this …
- Total blank
- Give me one second. (Then take it. Silence is allowed.)
- Technology fails
- While that loads, let me tell you why it matters.
- You have run over time
- I’ll leave it there — happy to take questions.
Want the ones you use before things go wrong — buying a second, restarting a sentence, finding a word that will not come? Those are in the nerves article. This box is only for the moment after.
FiveThe one habit that does the damage
Almost every recovery goes wrong in the same way, and it is not the mistake. It is the apology — specifically, the second one.
The first apology is fine. Brief, useful, human. The second one tells the room that something bad happened, and the third invites them to feel sorry for you. A mistake most of the audience half-noticed becomes the thing they are now watching for.
Worse, too many apologies take the pratfall effect away from you. The excellent performer spilled the coffee and carried on. If he had spent thirty seconds explaining that he was terribly sorry and does not normally spill things, nobody would have found him charming. They would have found him anxious.
Say sorry once, at most. Fix it in five words. Keep your volume where it was. Go on.
The three seconds after a mistake
Practice this until it is boring. It is a sequence, not a feeling.
- Stop. A short silence reads as control. Rushing reads as panic.
- Breathe out. It slows the sentence that comes next, which is the one you will rush otherwise.
- Fix it in five words or fewer, if it needs fixing at all.
- Decide fast whether it needs fixing. A wrong number does. A clumsy sentence does not.
- Apologize once, or not at all. Never twice.
- Do not explain the mistake. Nobody wants the story of why it happened.
- Return at the same volume you had before. Dropping your voice signals defeat more clearly than the error did.
- Finish cleanly. The ending is what they carry out of the room.
The short version
Six stumbles per hundred words is what ordinary speech looks like, so your standard was never real. A short filler is not damage. A small slip makes a prepared speaker more likable — but only a prepared one. Guard your opening and your closing, spend your mistakes in the middle, and apologize once at most.
Then finish well. That is the part they take home.
Practice recovering
You cannot rehearse a mistake, but you can rehearse the recovery. Speaking without preparation is the fastest way to get the practice — that is exactly what Module 7 and the Impromptu Generator are for.
Keep reading
Where this comes from
Every claim on this page is drawn from published research. No numbers have been invented, and where a finding has been stretched in popular retellings, we said what it actually showed.
Bortfeld, H., Leon, S. D., Bloom, J. E., Schober, M. F. & Brennan, S. E. (2001). Disfluency rates in conversation: effects of age, relationship, topic, role, and gender. Language and Speech, 44(2), 123–147.
Fox Tree, J. E. (2001). Listeners’ uses of um and uh in speech comprehension. Memory & Cognition, 29(2), 320–326.
Aronson, E., Willerman, B. & Floyd, J. (1966). The effect of a pratfall on increasing interpersonal attractiveness. Psychonomic Science, 4(6), 227–228.
Murdock, B. B. (1962). The serial position effect of free recall. Journal of Experimental Psychology, 64(5), 482–488.