A shot list for ad footage should be organised by clip role, not by scene order, because the output of the shoot is a library that gets recombined rather than one finished film. Plan for around 30 usable clips, three or four different openings for every message you intend to run, and two to three shots in each supporting role. A day organised that way produces a dozen genuinely different ads, while a day organised as a narrative produces one ad and some offcuts.
The shot list, by clip role
Print this as your list and tick roles off, not scenes. The counts assume one product and one main message, so scale up if you are covering a range.
| Clip role | What to capture | How many | Notes |
|---|---|---|---|
| Opening or hook | Something happening immediately, product visible | 4 or more | Vary the shot type, not just the angle |
| Product in use | A person using it, hands in frame, in context | 5 to 8 | The most valuable role, and the most often thin |
| Hands only | Close on hands doing one action | 3 to 4 | Universally reusable, no lip sync, no consent issue |
| Hero product | Clean product, plain background, slow move | 2 to 3 | Worth the lighting time, closes most ads |
| Detail and texture | Material, finish, mechanism, scale reference | 3 to 5 | Cheap to shoot, disproportionately useful |
| Result or outcome | The finished state, the before and after | 2 to 3 | Filmed last, forgotten most |
| Reaction | A real face reacting, no dialogue needed | 2 to 3 | Two seconds each is plenty |
| Environment | Where this happens, wide, no people in focus | 2 to 3 | Establishes context without needing a release |
| Talking to camera | The claim said plainly, three ways | 3 per claim | Needs a proper microphone, see below |
| Packaging | Box, opening, first look | 2 | Easy, and it reads as authentic |
| Deliberate offcuts | Setup, resets, between-takes moments | Keep everything | Often the best B-roll in the file |
That is roughly 30 to 40 clips, which is the band where variations stop repeating their openings. Our definitions for each role are in the clip types documentation if you want to align a crew brief to them.
The scheduling implication is worth stating: shoot roles in batches, not in story order. Get every hands shot while the lighting is set for hands. Story order is a habit from filmmaking, and it produces coverage that only makes sense in the sequence it was shot in.
How many openings should you capture?
Four per message, minimum, and they need to be genuinely different shots rather than four angles on the same moment.
The reason is the failure mode a thin library always produces. You write ten scripts, and because only two shots are strong enough to open on, every ad in the set begins with the same three seconds. The viewer sees one ad, repeatedly. Our benchmark reports CTR dropping 45 percent after a fourth exposure to the same creative, and near-identical openings are how a set of ten ads manages to feel like a fourth exposure by ad three.
Four openings that count as different:
- A close-up of the product mid-action, no context.
- A person starting to use it, framed at chest height.
- The problem state, before the product appears at all.
- A result or finished state shown first, out of order.
Each of those makes a different promise in the first second, which means each can carry a different argument. Four angles of the same person picking up the same bottle cannot.
If you are covering more than one message, multiply. Three messages, four openings each, is twelve opening shots, and that is a morning if you plan for it and impossible if you do not.
The shots teams always forget
After enough of these, the gaps are predictable. Every one of these is cheap to capture and expensive to go back for.
The result. Filming stops when the interesting part is over, which is usually just before the outcome exists. The clean plate, the finished surface, the assembled thing, the happy end state: nobody has it, and half your scripts want it.
Hands doing one small thing. A thumb pressing a button, a lid coming off, a strap tightening. These are the most reusable clips in any library because they have no face, no dialogue and no location, so they can drop into any variation.
Scale. A product photographed alone gives no sense of size. One shot with a hand, a door frame or a coffee cup in it answers the question a third of your comments will ask.
Multiple takes of the same line, said differently. Not for safety, for variation. Three deliveries of one claim are three usable clips, and the loose third take is usually the best.
The wide shot with nothing happening. Two seconds of environment, no people in focus, is the cutaway that saves half your edits, and it takes twenty seconds to grab.
Vertical framing. Not a crop of the wide take, an actual vertical setup. A centre crop of 16:9 keeps about 56 percent of the width and loses whatever was at the edges, which is often the product.
Audio capture priorities
Audio decides whether a clip can be a hook or only a cutaway, so it is worth ranking what to protect.
First priority is clean speech, isolated. A wireless lavalier or a shotgun on anyone saying a claim, recorded as a separate track, in the quietest space available even if that space is less photogenic. Background beauty is free. Background noise costs you the clip.
Second is room tone and product sound. Thirty seconds of the room with nobody talking, plus the actual noises the product makes: the click, the pour, the zip. Real product sound in an ad does more than a music bed, and it is the last thing anyone remembers to record.
Third is nothing. A large share of your library needs no audio at all, because cutaways under three seconds carry no dialogue, and knowing that lets you shoot fast in noisy conditions without worrying.
What not to rely on: the camera's onboard microphone for anything a person says. It sits two metres away, picks up the whole room and produces audio that marks an ad as amateur in the first second, which is exactly where you cannot afford it. If the recording is already made and the sound is poor, the workaround is to keep the picture and replace the voice with a clean scripted read, which is standard practice on archive material and covered on the repurpose existing footage page.
Shoot for both ratios, not one
Decide before the shoot which placements the footage has to serve, then frame accordingly. In-stream is 16:9 at 1920x1080 and Shorts, Reels and TikTok are 9:16 at 1080x1920, and one crop does not serve both.
Two practical approaches. Shoot wide and vertical separately for hero and talking shots, which costs setup time but produces properly composed frames. Or shoot in 4K wide and leave generous headroom so a 1080x1920 crop has somewhere to go, which is cheaper but only works for shots where the subject stays centred.
For everything else, meaning hands, details and textures, shoot tighter than feels right. A close shot crops in any direction. A wide shot with meaningful action at both edges is a 16:9 clip and always will be.
What to do with the footage after the shoot
The shoot is half the job. A drive full of untagged rushes is functionally the same as no footage, because nobody can find the shot they need at the moment they need it.
Genyad is our product, so read this as disclosure. Upload the rushes once and it transcribes and tags every clip into an indexed clip library, then each variation is built as a fresh script, shot selection, voiceover and caption set drawn from that library rather than a re-cut of one timeline. Uploading and re-exporting cost no credits, and a variation costs 1 credit, so a well-covered shoot day converts into a dozen or more ads at a small marginal cost. Exports cover 9:16, 4:5, 1:1 and 16:9 with framing handled per ratio.
What it does not do: it cannot create coverage you did not shoot. If there is no product-in-use footage on the drive, every variation will reach for the same hero shot, and the shot list is the only fix for that. There are also no AI avatars, no static banner formats, no product URL import, no predicted performance scores and no publishing to Meta, TikTok or Google Ads.
Frequently asked questions
How do I plan a shot list for ad variations?
Organise it by clip role rather than by scene: openings, product in use, hands, hero product, detail, result, reaction, environment. Target around 30 usable clips with two to three shots in each supporting role, and shoot the roles in batches while the lighting for each is already set. A narrative-ordered shoot list produces one ad and offcuts.
How many hooks or openings should I film?
At least four per message, and they need to be different shot types rather than four angles on the same action. Close product mid-action, a person starting to use it, the problem state before the product appears, and the finished result shown out of order all make different promises in the first second. Without that, every ad in the set opens the same way.
What audio should I capture on an ad shoot?
Clean isolated speech first, using a lavalier or shotgun microphone on a separate track, recorded in the quietest available space. Then room tone and the real sounds the product makes, which are more useful in an ad than a music bed. Do not rely on the camera's onboard microphone for anything a person says.
Should I shoot vertical or horizontal for ads?
Both, and separately for hero and talking shots. Cropping 16:9 to 9:16 keeps roughly 56 percent of the width, which loses whatever sat at the edges of frame, so a vertical version needs its own composition. Shooting 4K with generous headroom is an acceptable compromise for centred subjects, and close shots of hands or details crop safely in any direction.