Brief to published video ad in under an hour is achievable, and in a manual workflow it usually is not, because two steps consume most of the hour: finding the right footage and cutting it to length for each placement. Editing a short-form ad from existing footage typically takes 30 to 90 minutes, and a large share of that is search and conform rather than creative decisions. Remove those two steps and the remaining cost is review, at about five minutes per output, which is the floor nothing removes.
The useful way to think about ad production speed is not "how fast can one ad be made". It is "how many published ads per hour of a person's attention", because attention is the thing you are actually short of.
Where the hour goes, with a stopwatch on it
This is a single 20-second ad from footage that already exists, built by someone competent in an editor they know well. Times are what I would expect on a normal day, not a best case.
| Step | Manual | With generation from an indexed library | What the time is actually spent on |
|---|---|---|---|
| Read the brief, decide the structure | 4 min | 4 min | Thinking. Unavoidable and cheap |
| Write the script and hook | 8 min | 1 min | Wording, and rewording it shorter |
| Find the shots | 15 to 25 min | 1 min | Scrubbing timelines, opening the wrong file, remembering a clip exists |
| Assemble the cut | 8 min | included | Dragging, trimming, matching audio |
| Voiceover or audio pass | 6 to 10 min | included | Recording, retakes, levels |
| Captions | 5 to 8 min | included | Typing, timing, restyling |
| Conform to length and ratio, per placement | 10 to 20 min | 1 min | Reframing, re-exporting, checking safe zones |
| Export and file handling | 4 min | 1 min | Waiting, renaming, uploading |
| Review | 5 min | 5 min | Six checks against the brief and constraints |
| Total per output | 65 to 92 min | about 13 min, of which 5 is review |
Two lines dominate the manual column: finding the shots, and conforming to length and ratio. Together they are 25 to 45 minutes, which is roughly half the hour, and neither of them is a creative decision. That is the part worth being annoyed about.
Why footage search is so expensive
Because it is unindexed. Most libraries are organised by shoot, date and file name, which answers "where did this come from" and not "what have I got of someone frustrated by paperwork". So the search becomes linear scanning at 8x speed, and it is slowest exactly when you need it most, which is on the fifth variation when the obvious shots are used up.
The fix is indexing rather than discipline. Clips tagged by the role they can play in an ad (hook, problem, solution, product demo, testimonial, lifestyle, branding, transition, call to action) and searchable by transcript turn a 20-minute scan into a query. Our guide to generating variations covers how that pool feeds a sequence, and the practical point is that the index has to exist before the deadline, not during it.
Why conforming to placements is so expensive
Because it is per-ratio work that feels like it should be free. One 20-second ad delivered at 9:16, 4:5, 1:1 and 16:9 is four reframing passes, four safe-zone checks and four exports, and the caption position usually breaks in at least one of them. Ten minutes if the ad was built with reframing in mind, twenty if it was not.
This is the step where manual workflows quietly cut corners: teams deliver one vertical file and let the platform crop the rest, then wonder why the feed placement underperforms.
What generation removes, and what it does not
Removes, in my experience: the search, the assembly, the caption timing, the voiceover pass, and the per-ratio conform. That is the 45 to 70 minute block in the middle of the manual column, and it goes to roughly a minute of machine time per output.
Does not remove:
- Deciding the offer and the angle. The brief is still yours, and a bad brief produces bad output faster.
- Review. About five minutes per output, and this is the load-bearing point of the whole post.
- Having usable footage. No library, no output. If your first step is a shoot day, your production speed is measured in weeks.
- Judging which output will win. There is no predicted performance score in our product, and I would distrust one anywhere else. The test decides.
- Publishing. Genyad has no direct publishing to Meta or TikTok, so upload and campaign structure stay manual, which is about 3 minutes an ad in bulk.
Genyad is our product, so weigh this accordingly. One variation is 1 credit, about EUR 1.52 on the EUR 99 Growth pack, and uploading footage, editing and re-exporting cost nothing, so the marginal cost of another attempt is roughly a minute and one credit. That changes what you would try, which is the actual value, more than it changes any single ad's timeline. There are also no AI avatars, no static banners, no product-URL import and no feed or CSV template rendering, so if your speed problem is rendering 400 product variants from a spreadsheet, this is the wrong tool and a template renderer is the right one.
Review is the remaining cost, and it is the one that scales
Here is the trap. When production drops from an hour to a minute, throughput does not rise by 60 times, because review does not move. Five minutes per output is stubborn, and every honest speed calculation has to carry it.
Then there is keep rate, published divided by produced, which is what converts review time into real cost per published ad. At a 60 percent keep rate, shipping 40 ads a month means producing 67, so you are paying 67 reviews for 40 launches.
| Keep rate 45% | Keep rate 60% | Keep rate 80% | |
|---|---|---|---|
| Produced to publish 40 | 89 | 67 | 50 |
| Review time at 5 min each | 7.4h | 5.6h | 4.2h |
| Review cost at EUR 45 loaded | about EUR 333 | about EUR 252 | about EUR 189 |
| Effective minutes per published ad | 11.1 | 8.4 | 6.3 |
Loaded hourly cost there means salary plus overhead, not the salary line on its own. The last row is the number to manage: minutes of human attention per published ad. Manual production puts it at 70 to 95. Generation with a sloppy brief puts it at 11. Generation with a constraint-first brief puts it at 6.
That is why the fastest teams I have worked with are not the ones with the best editors. They are the ones whose briefs specify constraints up front, so review is checking rather than deciding. Keep rate is the production metric worth tracking if you want production speed to mean something on the media side.
Getting to under an hour, end to end
A realistic under-an-hour path for a single ad, assuming an indexed library and a written brief:
- Brief already exists (0 min, written weekly, not per ad).
- Specify the variation: angle, sequence, length, ratios (3 min).
- Generate (about 1 to 2 min of processing).
- Review against six checks: claim, disclaimer, offer wording, audio, angle match, ratio and length (5 min).
- Fix or regenerate if needed, one pass only, capped (5 min).
- Export all ratios (1 min).
- Upload, name, attach to the ad set (3 min).
That is roughly 18 minutes of attention, and most of it is review and upload. Our video ad builder page shows the middle of that sequence, and our video ad cost calculator will convert your own review time and keep rate into a cost per published ad.
The thing to protect is step 5's cap. An uncapped fix loop is how a 13-minute output becomes a 40-minute one, and the reason it happens is almost always a brief that did not specify what the reviewer is now trying to fix by hand.
Frequently asked questions
How long does it take to make a video ad?
Manually, 30 to 90 minutes for a short-form ad from existing footage, with 25 to 45 of those minutes spent finding shots and conforming the ad to each placement rather than making creative decisions. With an indexed library and generation, the machine time is about a minute and the human time is about 13, of which 5 is review. Add a shoot day and the answer changes to weeks.
What is the slowest part of video ad production?
Finding the right footage, typically 15 to 25 minutes per ad, followed by conforming the cut to each aspect ratio, typically 10 to 20 minutes. Both are search and formatting rather than craft, and both get worse as you build more variations from the same library. Indexing footage by clip role and transcript is what collapses the first one.
Can you really produce a video ad in under an hour?
Yes, if the brief is written, the footage is indexed and the review is a checklist rather than a discussion. About 18 minutes of human attention per ad is achievable in that setup. What breaks the hour is an uncapped revision loop, which is nearly always caused by constraints that were not in the brief.
Does faster production mean cheaper ads?
Only if your keep rate holds. Cost per published ad is everything you spend divided by the share of outputs you actually publish, so producing three times faster at a 40 percent keep rate can cost more per published ad than a slower workflow at 80 percent. Track produced, published and the ratio between them weekly.