A 30-second talking head cut takes about 40 minutes. A 10-minute one takes 5 to 10 hours.

Thirty seconds: about 40 minutes of hands-on work for one vertical cut. That figure is the audit further down this page, run on a real ad rather than derived from a rule.

Five minutes: 3 to 5 hours. Ten minutes: 5 to 10 hours. Twenty minutes: 10 to 20 hours. Those follow from the published rule of 30 to 60 minutes of editing per minute of finished video (Tasty Edits, read 1 September 2026).

Two revision rounds with a stakeholder roughly double every figure above, because each round costs the fix, the reload, the re-watch and the re-export.

Now the part where the rule and the audit disagree. Forty minutes for half a minute of output works out at 80 minutes per finished minute, above the top of the published range. Neither number is wrong.

Short ads carry more decisions per second, not fewer. Use the rule for anything over about two minutes and the audit below for anything shorter.

One thing changes all of them. If the first pass arrives already assembled, the audit collapses to the two steps that carry a decision, which is the last section here.

People estimate editing by output length. That is why the estimates are always wrong. Work scales with the number of separate decisions. A short ad contains more decisions per second, not fewer.

A thirty-second ad might carry twenty cuts, six clip placements and about seventy-five words of captions to verify. That is roughly fifty decisions. Each one is a scrub, a judgement and a click.

Compression is work. Removing the fat is the part that requires taste. A shorter ad often takes longer per second than a longer one.

Count honestly when you run this on your own edit. The minute hunting for the folder. The two minutes waiting for a preview. The six minutes lost to a phone call.

Every round trip with a stakeholder adds the fix, the reload, the re-watch and the re-export. Two rounds can cost more time than the original assembly.

A second multiplier only shows up in batches. Editing one ad takes an evening. Editing twelve takes fifteen. By the eighth you are making inconsistent choices about pace and caption style.

Then you go back and redo the earlier ones to match. That consistency work is invisible on a single edit and dominates a batch.

Published turnaround guides put simple talking-head work at 30 to 60 minutes of editing per finished minute and heavily styled work at one to two hours, read 1 September 2026. Run the audit with a stopwatch and you can see which band you are actually in.

Time your own next batch rather than trusting this one. Three numbers are enough. Minutes on the first ad. Minutes on the eighth. Minutes spent going back to make the first three match the eighth.

That third number is the one nobody expects. On a batch it is usually the largest.

Two steps dominate for most people and neither is the creative part.

Searching for footage is the first. It is open-ended and it invites browsing. It is where a forty-minute edit becomes a ninety-minute one. Decide which phrases need footage before you look. Then look only for those.

Caption timing is the second. Done by hand it is a per-word chore with no creative upside. Nobody has ever won an auction because their caption timings were hand-tuned.

The parts you should not rush are choosing the opening line and deciding what to cut. Those move the numbers. They are about ten percent of the clock.

Test it on your own last edit. Add up the minutes spent choosing words against the minutes spent dragging clips. The ratio is usually about one to nine.

If the first pass arrives assembled, the audit collapses to the two steps that matter. Read the transcript. Decide what stays and what leads.

That is the shape of Cutroom. Upload one take of up to three minutes. It transcribes the read, drafts the edit, finds the footage, times the captions and returns a finished 9:16 MP4.

The searching goes. The dead air goes. The caption timing goes. Those are the three longest bars in the audit above and none of them contained a creative decision.

Changes happen on the transcript. Highlight a phrase and a clip lands over exactly those words. Delete a line you do not want and the cut rebuilds around it. Change the pace and it re-cuts.

Nothing else collapses the whole list in one step. A stock library shortens the search. A caption app shortens the timing. This hands back the finished vertical file with all of it done.

One step stays yours and it decides the ad. Choosing which three minutes to record. No assembly speed rescues a take with no argument in it.

One step should never be automated. Watch the export before it runs anywhere.

Read the captions for a mis-heard product name. Check the first frame is not a blink. Then play it once on a phone at arm's length. Thirty seconds, every time.

The ads that embarrass companies are always the ones nobody watched all the way through.

The facts a buyer needs before the first upload. The source is a person speaking. Three minutes is the ceiling, so an eight-minute ramble gets recorded again rather than trimmed.

Every export is a 9:16 MP4. There is no timeline underneath when a clip lands half a second wrong. You mark a different phrase instead, which is coarser and about thirty-five minutes faster.