What We Measure When We Measure Speaking

"You were great" is the least useful sentence in communication. It is kind, it is unfalsifiable, and it gives you nothing to change. The alternative is not harshness. It is specificity: a small number of things that can be observed, counted, and improved.

Here is every dimension we measure, what it actually means, and why it matters more than it sounds.

Pace — words per minute

Conversational English sits around 140 words per minute. Presentation delivery that lands tends to run slower, roughly 120 to 150, because listeners need processing room they do not need in dialogue. Above about 170, comprehension starts to drop before the listener notices it is dropping — they feel busy rather than confused, and they blame themselves.

The aggregate number matters less than its distribution. Almost nobody speaks fast throughout. People speed up on the part they are least certain about, which is reliably the part that mattered most. That is why we surface rushed segments individually rather than only reporting an average.

Pauses — and the three kinds

A pause is not empty. It is punctuation the listener can hear. We count three kinds separately because they do different jobs: micro pauses under half a second, which mark clause boundaries and make a sentence parseable; emphasis pauses of roughly half a second to a second, which tell the room that the thing you just said was the point; and dramatic pauses beyond a second, which create anticipation and are the single most underused tool in professional speech.

Most nervous speakers have plenty of micro pauses and almost no emphasis pauses. The result sounds fluent and lands flat.

Silence ratio

The proportion of your recording that is not speech. Very low silence — under about ten per cent — reads as pressured, as though you are speaking to avoid being interrupted. Very high silence can read as hesitancy. The range in between is where the listener has room to agree with you.

This is the metric that most surprises people. Speakers who are told they need "more confidence" usually need more silence.

Energy arc

We bin the recording into ten-second windows, score each one on a composite of volume, pace variation and pitch movement, smooth the curve, and then name its shape: building, front-loaded, peaked early, inverted-U, strong open that fades, flat, variable.

The most common arc among capable professionals is a strong open that fades. You prepared the first ninety seconds carefully and improvised the rest. The room hears exactly that.

Structure

Three questions. Did the point arrive early, or after a runway? Did the middle hold a shape a listener could follow? Did the close land on a statement, or dissolve into an explanation of the statement?

Endings fail more often than openings. The last two sentences of most professional speech are an apology for the length of the previous ones.

Listener's memory

The hardest measure and the most honest one. Not what you said — what a listener would carry out of the room. We separate what was retained from what quietly fell out, and the gap between the two is usually the real diagnosis. You can be entirely clear and still be forgotten, because clarity and memorability are different properties.

Why numbers instead of adjectives

Because adjectives cannot be practised. "Be more confident" has no first step. "Your pace ran at 178 words per minute through the pricing section, and you used no pause longer than 0.4 seconds in three minutes" has an obvious one.

The point of measurement is not to score you. It is to convert a vague feeling of having spoken badly into two or three specific behaviours you can change on Tuesday.

Start your first session — record three minutes and see how you actually land. It is free and takes about three minutes.