Knowledge Bank

3 min read

The speed coach, and why retention graphs cannot diagnose pacing

A drop-off looks the same whether you talked too slowly or the idea was weak. Measure the pacing instead.

Not sure if Produce is your weak phase? The free 3-minute workflow audit will tell you.

Flat illustration of a microphone.
ProduceLinked task: Shoot the A-roll · step 14 of the 47

Almost nobody knows how fast they talk. The two guesses creators make are wrong in the same direction: the ones who think they rush are usually fine, and the ones who think they are being clear are the ones losing people. The speed coach measures it instead of leaving you to guess.

Paste the script
Pick the format
Read it into the mic
Read the four numbers

Retention graphs cannot diagnose this

This is the reason the tool exists. A drop-off on the graph looks identical whether your pacing was too slow or the idea was weak, so staring at retention will never tell you which problem you have.

Pacing is measurable before you film. Every other fix for a slow video happens in the edit, after you have already spent the afternoon recording it.

The four numbers you get back

Words per minute across the whole take, your pace with the pauses removed, how many pauses there were, and the longest one.

That second number is the one people misread. A comfortable overall pace can hide a delivery that sprints between long gaps, and the gap between those two figures tells you which you are doing. If the two are close, your rhythm is even. If the pause-free number is much higher, you are talking quickly and covering it with silence.

The longest pause is worth a glance too, because the one you never notice while filming is always longer than you think.

The word count is exact

The tool never transcribes you. The word count comes from the script you pasted, so the figure the whole result depends on is not a guess.

That is a deliberate design choice rather than a limitation. A speech-recognition error would land directly in the denominator and quietly corrupt every number on the screen.

The band depends on what you are making

FormatRough band
YouTube explainer150 to 180 wpm. Fast enough to hold attention, slow enough for an idea to land
Vlog or talking headA little under that
Podcast or interviewSlower again
Talk to a roomSlowest of the lot

The instruction it gives is the useful part

When your pace lands outside the band, it tells you the clock time the same script would take inside it.

Being told to slow down is advice nobody can act on in the moment. Being told this script should run nine minutes and you just did it in seven is something you can hear yourself doing on the next take.

No audio goes anywhere

Nothing is recorded, saved or sent. The tool reads how loud the microphone is twenty times a second and keeps those numbers, and that is the entire dataset.

When to run it

Once, on the script, before the first take. That is where it pays, because pacing you fix in the read costs nothing and pacing you fix in the edit costs an evening.

It is also worth a run when you switch format. The pace that works on an explainer is noticeably too quick for a podcast, and most people carry their usual delivery across without noticing.

Keep reading

The read-aloud pass belongs to writing the script, and the take itself is covered in shooting the A-roll. If the problem turns out to be energy rather than speed, being less boring on camera is the better read.