Every script has a duration whether you've measured it or not. The only question is whether you find out at your desk or in the room.

Timing a script properly takes three passes: an arithmetic estimate, timestamp markers to see the shape of it, and a read-through to catch what the arithmetic missed. Each one catches a different kind of error.

Pass One: The Arithmetic

Word count divided by speaking pace. At 140 words per minute — the default across ReadCalc's tools and the midpoint of the band research puts prepared delivery in — a 1,400-word script runs about ten minutes.

Word count At 130 WPM At 140 WPM At 160 WPM
500 words3:513:343:08
750 words5:465:214:41
1,500 words11:3210:439:23
2,000 words15:2314:1712:30
3,000 words23:0521:2618:45

That's the estimate. It is also, reliably, the shortest your script will ever be — because it assumes you talk continuously at a constant rate and never pause, which no one does. Treat it as a floor.

Which pace to divide by matters more than people expect. The gap between 130 and 160 WPM on a 2,000-word script is nearly three minutes. If you don't know your own rate yet, the research on speaking rate covers where the standard figures come from and how to measure your own.

Pass Two: Timestamp Markers

A total duration tells you whether the script fits. It doesn't tell you whether the shape is right — and that's usually the real problem.

Timestamp markers fix that. Paste the script into the script formatter, set your pace and an interval, and it inserts a marker at every interval so you can see where you'll be at each point.

> $ script-formatter --interval 30 --style inline
The thing nobody tells you about migrations is that the [0:30] hard part isn't the redirects. It's deciding what not to bring across. [1:00] We started with forty thousand URLs and a spreadsheet…

Now you can see structure. If your three-minute introduction is still running at the [4:00] marker, you know exactly which section to cut — and you know it before you've rehearsed a version that doesn't fit.

SettingRangeUse
Speaking pace 80 – 220 WPM Your measured rate, or 140 to start
Custom interval 1 – 600 seconds Match slide changes or segment breaks
Marker style Inline or new line Inline to read past; new line to scan

Pick the interval to match what you're checking. Thirty seconds is right for a conference talk — frequent enough to catch drift, sparse enough to read around. Fifteen is better for a 60-second ad read where every phrase is budgeted. For a presentation, set a custom interval matching your average slide time and put a marker at each slide change instead.

> $ tip --marker-style

Use inline markers if you'll read from the script itself — they sit in the flow and your eye passes over them. Use new line if you're editing rather than performing, since the markers become scannable checkpoints down the left edge.

Pass Three: The Read-Through

The arithmetic assumes constant pace and no silence. Both assumptions are wrong, and the gap between them and reality is where over-runs come from.

Read the script out loud, at performance pace, with a timer running. Not in your head — reading silently runs far faster than speaking, and if you time a script by reading it to yourself you will be minutes out. That gap is covered in speaking time vs reading time, and it's the most common single error in script timing.

Expect the spoken version to run 5–15% longer than the estimate. What accounts for the difference:

What the estimate missesTypical cost
Pauses between sections2 – 4 seconds each
The beat after a point lands1 – 3 seconds each
Slide changes and demos5 – 15 seconds each
Audience laughter or reaction3 – 8 seconds each
Corrections and restartsUnpredictable

None of these are failures. Pauses are how a listener catches up — a reader can go back over a sentence, a listener cannot, so the silence is doing work. The mistake isn't pausing, it's not budgeting for it.

> $ rule --write-short

Write 10–15% fewer words than the target suggests. For a ten-minute slot at 140 WPM, that's about 1,200 words rather than 1,400. The difference is your pauses, and you will use all of it.

If It's Over

Cut words. Don't raise the pace.

Speeding up to make a slot is audible, and it's the first thing that makes a delivery sound rushed. It also compounds: a faster read needs simpler sentences to stay comprehensible, and a script written for 140 WPM read at 170 has exactly the wrong structure for the speed it's being delivered at.

Cut the introduction first — almost every script over-runs at the front — then look for filler. A pass through the filler finder will typically recover 5–10% of the word count without touching a single point you're making, and cutting word count without losing meaning covers the deeper edits when that isn't enough.

Then Rehearse at the Real Pace

Once the timing works on paper, load it into the teleprompter at the same WPM you used for the timestamps. The scroll enforces the pace you planned rather than the one nerves want to give you, and the markers you added line up with where you should be.

If the scroll speed feels wrong once you're reading, that's information about your natural rate rather than a fault in the script — what WPM to set covers how to calibrate it. And if you're working the other way round, from a fixed duration to a word target, the reference tables in how many words is a 10 minute speech give you the starting count for every common length.

Add timestamp markers to your script at your own speaking pace — inline or on their own line.

$ open script-formatter →