Notes for owners · Digital marketing
How to make short form video ads that survive the feed
Vertical, muted, in one hand, with a thumb already moving. A feed video is judged on its first frame, so the hook is the media buy and everything else is the film.
The GullySales team · Updated 21 Sept 2026 · 7 min read
A short form video advertisement is watched vertically, in one hand, with the sound off, by somebody whose thumb is already moving to the next thing. It is not skipped after five seconds. It is skipped after one. So the first frame is what you are actually buying, and the rest of the film only matters to the people that frame stopped.
The first frame, not the first five seconds
Open on the thing that makes somebody pause: a face mid-sentence, a hand doing something unusual, the problem itself. A logo sting at the start is the most expensive second in Indian advertising, because it is the second where everybody leaves.
Start with the problem in the viewer's words, on screen, in text large enough to read at thumbnail size. Four to six words. "Knee pain when you climb stairs." Then the film can breathe.
Sound off is the default, so the picture has to carry it
Most of this viewing happens in a bus, in a queue, in a waiting room, or next to somebody sleeping. If the offer only exists in the voice-over, the offer does not exist.
Burn the captions in. Not the automatic ones, which get Indian names and place names wrong, and not a wall of text at the bottom where the app's own caption sits. Put short lines in the middle third of the frame, timed to the speech, and keep them on long enough to read twice.
Then mix the audio properly anyway. The people who do turn the sound on are the ones closest to buying.
The screen is not all yours
The app draws its own interface on top of your film. Roughly the bottom fifth carries the caption, the handle and the button, the right edge carries the icons, and the top carries the profile strip. Anything important you place there is covered.
So the rule is simple: keep your text and your product inside the middle of the frame, with a generous margin all round. And shoot vertical from the start. Cropping a 16:9 film to 9:16 leaves the person small in the centre and cuts the product out of the side of the frame, which is why re-purposed television footage looks wrong in a feed even when nobody can say why.
A shape that works for a small business
| Second | What is on screen | What it is doing |
|---|---|---|
| 0 to 2 | The problem, as text over a face or a hand | Stopping the thumb |
| 2 to 8 | The thing itself: the clinic, the machine, the dish, the flat | Proving it is real |
| 8 to 16 | One reason to believe: a customer, a price, a timing, a guarantee you actually give | Answering the obvious doubt |
| 16 to 22 | Where you are and what happens next | Turning attention into an enquiry |
| Last 2 | The name and the locality, held still | Being remembered by the ones who did not tap |
Twenty to twenty-five seconds is enough for almost any small business. Longer films work when something is genuinely being shown, such as a full procedure or a factory line.
Shoot the hook five times
This is the whole production method, and it takes one morning.
Write one body: the demonstration, the proof, the ask. Then write five different openings for it. Shoot all five, cut five films that share everything after the third second, and run them against each other. The platform will spend the money on the one that holds attention, and you will learn something about your buyer that no amount of discussion in the office would have produced.
For example, a physiotherapy clinic in Jayanagar spends one morning filming: the physiotherapist explaining why a knee hurts on stairs, a patient on the table, the reception, and five different first lines. Six films leave that morning. Two of them get used again in the next quarter, with the Kannada voice swapped in for one and the English kept for the other.
Where the tap lands
The film is half the job. The other half is what opens when somebody taps, and a feed viewer who lands on a home page is gone in a second. Send them to a page about the one thing the film showed, or straight into WhatsApp with the message already written, which for a clinic or a dealer is the shortest route to a real conversation. Decide that destination before the shoot, because it changes the last line of the script.
What to look at afterwards
Ignore the view count. A view is counted after a second or two, which is exactly the behaviour you are trying to beat. Look at how many people reached the end, how many tapped, and how many of those became a call, a WhatsApp message or a booking. Our work on performance advertising creatives is built around that loop: new hooks against the same body, monthly, with the enquiries counted rather than the impressions.
Say the uncomfortable part out loud too. If you sell rubber mouldings to forty purchase managers in Peenya, this is not your medium, whatever your agency says about Reels. Short form video is for things a stranger can want on sight.
What to do next
Open your own phone, go to the feed your buyers use, and scroll for two minutes. Notice which videos you stopped at, and look at what was on the screen in the first second of each. Then write five openings for your business against that, film them tomorrow morning, and let the money decide which one was right. Video production and marketing covers the shoot day and the edit if you want a crew for it, and the free audit is where we check whether the enquiries these films produce are being answered fast enough to be worth making.