Algeria AgencyDigital marketing
Video: a demonstration, not a film
A useful video answers a question you already receive. The rest — the music, the editing, the drone — costs a great deal and answers nothing.
The article on content creation defends one simple idea: photography is a process, installed once and then carried out by anybody. Video does not submit to the same description, and it is better to say so straight away.
What gets fixed in a video is not the setting, it is the question. A good commercial video shows one thing a photograph cannot: an object in motion, a size compared, a gesture, a material that folds, a machine that turns. Everything else is filmed padding.
So this article describes the opposite of an aesthetic. It describes what goes wrong — the sound, almost always the sound — what gets read without being heard, what the platform picks on your behalf when you do not pick, and how long a subject genuinely deserves.
It also carries the broadest refusal in this series: most Algerian businesses we meet do not need video. They need six adequate photographs and a published price, and a video produced ahead of that is an elegant expense laid on top of an unsolved problem.
What a video shows that a photograph cannot
Start with the sorting question, because it eliminates half of all projects before they cost anything: what does this shot show that a still image would not? If the answer is "it feels more alive", the subject was never a video subject.
Four answers hold up and they come back over and over. Movement: fabric falling, a door sliding, a mechanism clicking into place. Scale: an object picked up in a hand, which settles in three seconds the question a photograph leaves open and the content-creation article treats at length.
The third is the gesture: how it is installed, opened, cleaned, adjusted. It is the most profitable video that exists, because it replaces an entire conversation and goes on replacing it for two years.
The fourth is proof: the machine turning in your workshop, the real stock on the shelf, the team packing. Those are images a competitor cannot copy out of a catalogue, and it is the one category where showing the premises has commercial value.
Anything that fits none of the four is a decorative shot. It costs filming, editing and attention, and it moves nobody. The sorting rule fits in a sentence: film what moves or what has to be explained, photograph the rest.
What it gets watched on, and who pays for the connection
One decision precedes all the others and it is not artistic: your video will be fetched by a phone, over a mobile connection, paid for by the person you are trying to persuade.
In the second quarter of 2025, mobile subscriptions were 88.71% of the country’s internet subscriptions, against 11.29% for fixed lines. That is not an audience statistic, it is a manufacturing constraint.
Three direct consequences. Length is not free: every extra second is data billed to somebody who does not know you yet. File weight decides load time, and a video that takes four seconds to start has already lost half its viewers.
And legibility is judged on a six-inch screen held at arm’s length in the street. Small burnt-in text, a wide shot where the product occupies a tenth of the frame, a label you have to guess at: all of it reads perfectly well on the editor’s monitor and nowhere else.
The practical rule we apply: if the subject is not identifiable in a screenshot of the first third of the video, shrunk to thumbnail size, the shot has to be taken again. It is a test that takes ten seconds and saves a week.
ARPCE, internet market observatory, second quarter of 2025
A thousand-dinar tripod is the whole equipment budget
The only production difference a viewer registers without thinking about it is stability. A shaking image reads as amateurism; a still one reads as nothing at all, which is exactly the effect wanted.
A phone tripod costs about what lunch costs and replaces the entirety of what a supplier will offer to sell you. A stack of books and a wall do the same job on day one, provided the frame stops moving once recording starts.
The second useful expense, a long way behind, is daylight — that is, no expense. The same rules as for photography apply: a window to one side, the strip lights off, and a fixed hour so two videos in the same series resemble each other.
What is of no use at this stage: the drone, the motorised gimbal, the second camera, the ring light bought before anybody tried the window. Each of those purchases solves a problem you do not yet have, and none of them solves the sound, which is the one you do have.
One piece of field advice worth more than the equipment: put the phone down, step into the frame, and talk to the object rather than to the lens. Somebody handling something forgets the camera within twenty seconds, and that is where the take becomes usable.
Sound: the one mistake that is never recovered
A dark image can be corrected. A crooked frame can be cropped. A false colour can be pulled back. Bad sound cannot be repaired, and it is the only categorical claim in this article.
The noise that ruins a take is almost never the noise anybody is listening to. It is the fridge compressor, the ceiling fan, the air conditioning, the steady rumble of the street through an open door. The ear stops hearing it after ten minutes; in the edit it is there, even, and it makes every sentence tiring.
Three corrections cost nothing. Switch off whatever is running — fridge, air conditioning, fan — for the five minutes of filming, and write it on the shot card so nobody forgets. Get closer: at forty centimetres a phone microphone is adequate; at two metres it records the room.
The third is the place: a room with fabric, boxes and clothes on rails sounds immeasurably better than an empty space with bare tiling. That is not a studio preference, it is reverberation, and a textile shop is acoustically better than an office.
And if the sound is bad anyway, the decision is to shoot the take again, not to rescue it with music laid over the top. A voice drowned under a song does not become audible: it becomes a video people watch without understanding, which is worse than a video nobody made.
Subtitles are not an option, they are the copy
A video published in a feed starts with the sound off. The person is on a bus, in a queue, in a waiting room, sitting next to somebody. They are not going to switch the sound on for an account they do not know, and they are not going to do it for one they do know either.
Which means your subtitles are not a courtesy to deaf viewers — they are that too, and it is one more good reason — but the default reading mode for your message. The burnt-in text is the video; the soundtrack is the supplement.
This is one of the rare places where this series contradicts itself from one platform to the next, and flagging it beats letting it collide. The article on long-form video advertising describes a surface where sound is on by default and the voice carries the message; here, in a feed, the reverse holds, and the same file does not work in both places.
Write the subtitles short, at the top or the middle of the frame, never in the bottom fifth: that strip is covered by the platform’s interface, the handle, the caption and the buttons. It is the commonest and least visible error in editing.
A note on language, particular to here: spoken Algerian and written language do not coincide, and subtitling word for word what is said in dialect produces text nobody reads comfortably. Subtitle the meaning in French or Arabic, keep the voice as it is, and apologise for none of it.
The first frame is a thumbnail, and the platform picks badly
Every video has an opening frame, and that frame is what will be seen by the people who never watch it. It travels through the account’s grid, through search results, through shares. It is worth more than the edit.
If you do not choose it, the platform chooses it, and it chooses without intention: a frame taken at random from the opening seconds, which very often means somebody with their eyes shut, a blurred object, a hand across the lens, a frame still settling.
The correction is to film a still frame on purpose: before starting to speak, hold steady for two seconds with the product well framed and well lit. You get a usable thumbnail without editing anything, and that one second of discipline shows across the whole grid.
The grid is in fact the real stake, as the article on advertising on the image platform explains: nine thumbnails decide before the first video is ever played. Three black or blurred thumbnails in the middle of a grid make the entire account look neglected.
And write as little as possible on that thumbnail. Two or three words at most, large, readable at quarter size: at the size it is actually seen, a sentence is a grey smudge that says nothing.
Length: write for the question, not for the minute
The wrong question is "how long should a video be". The right one is "what question am I answering, and how long does answering it out loud take". Length is a result, never an instruction.
Three lengths genuinely exist in this trade. Five to ten seconds: an object, a price, a hand turning it. That is an announcement rather than an explanation, and it gets produced in batches during the photography session.
Twenty to forty seconds: a demonstration, a gesture, a before and after, a size comparison. It is the format that does the most commercial work, and it is the one almost nobody produces because it requires knowing in advance what is being shown.
Two to four minutes: a full explanation, an instruction manual, an installation, a workshop visit. That length does not get published in a feed — it is kept for the product page, for the reply to a customer who wrote in, for the website. It is a service video rather than a broadcast one.
The test that settles it: cut the first three seconds and the last three and see whether anything is missing. In most company videos, nothing is — because the opening is a greeting and the ending a thank-you, and neither of them is information.
Four formats that repeat without a meeting
The article on social media management shows that every editorial calendar dies around the sixth week, because it assumes the material will exist on the appointed day. Video formats escape that when they are attached to an event that already happens on your premises.
The first: the delivery. The box opens, the item comes out, it gets turned, the price is said. It runs thirty seconds, it gets filmed on the day the goods arrive, and it requires no decision at all.
The second: before and after. A renovated space, a repaired part, a cleaned car, a haircut, a machine brought back into service. Two shots, no commentary needed, and it is the hardest proof there is to argue with.
The third: the instructions. How it is assembled, plugged in, washed, adjusted, maintained. Every question that comes back three times in your inbox is a video to make once, and the return on time spent is the best on this list.
The fourth: the prepared order. The item wrapped, the address written, the parcel closed. It looks trivial and it answers the question nobody writes down: are these people real and does the parcel actually go out. In a country where a great many purchases are settled on delivery, that is a commercial answer.
Vertical framing, and the file nobody should crop
Film vertically. That is not a taste of the moment, it is the shape of the phone the viewer is holding, and a horizontal video published in a vertical feed occupies a fifth of the screen with two black bands around it.
The corollary costs shooting days: a vertical video is not manufactured by cropping a horizontal one. The subject sits off centre, heads get cut, burnt-in text falls outside the frame, and the loss of definition shows. Decide the format before pressing record, not in the edit.
If the same subject also has to serve horizontally elsewhere — a product page, an in-store screen, a long-form video platform — that takes two takes, or a wide framing with margin at the top and bottom planned from the shoot. It gets decided in ten seconds beforehand, or it costs a day afterwards.
One point carried up from the article on short-form advertising: a file bearing another platform’s watermark is shown to fewer people. Exporting from the editing app rather than downloading your own video from another network is not a procedural detail, it is lost distribution.
Finally, keep the rushes. The original file, uncompressed, uncropped, not exported twice, filed by the same rule as the photographs: reference, then take number. A kept rush is a video that can be recut differently in six months; a deleted one is a shoot to start again.
The edit: cutting, and nothing else
Editing a useful commercial video consists of removing, never of adding. You cut the opening, cut the hesitations, cut the ending, add subtitles, export. Thirty minutes of work for a thirty-second video is already generous.
The first thing to remove is the scene-setting. "Hello everyone, today I am going to show you…": those five seconds are precisely the moment the viewer decides, and you spent them announcing that you were about to start. Start.
Transitions, effects, automatic zooms, spinning numbers: they are not neutral, they date the video to the quarter and signal that a template was filled in. A straight cut never dates and never gets noticed, which is the property being sought.
Music deserves a paragraph for one legal reason and one practical one. Legal: a known song laid over a commercial video is unlicensed use, and the takedown arrives at the moment the video starts working. Practical: it covers the voice, and the voice was the content.
Our default position is therefore: no music at all on a demonstration, quiet royalty-free music on an edit without speech. And if the argument "it feels more dynamic" comes back, it is nearly always the sign that the shot itself shows nothing.
What video does not sell
This needs a section because it is the whole profession’s implicit promise and it is false: the view count says nothing about turnover, and a heavily watched video can produce absolutely nothing.
The mechanism is the one described in the article on short-form advertising: distribution is national and young, and a video can be seen a hundred thousand times by people who are neither in your wilaya nor in your market. The number is real and it does not concern you.
The useful judge is elsewhere, and it is the same one throughout this series: messages, calls, people coming into the shop, orders accepted. A video with three thousand views producing eleven serious messages beats one with a hundred thousand producing none, and that happens constantly.
The second thing video does not do: it does not replace missing information. If the price is written nowhere, if the stock is out of date, if nobody answers messages within three hours, a video only increases the number of people who leave without an answer.
That is why we nearly always place video after the photographs and after the published price. It is not a budget question but an order-of-operations one: filming before those two are settled amplifies a conversation that ends badly.
What we do, and what we will refuse to do
What we will refuse first: producing a video for a business whose catalogue does not yet carry six adequate photographs and a published price. It is the broadest refusal in this series, it costs us billings, and it is the right order — a video in front of a mute catalogue sends people to a page that does not answer.
We will also refuse to promise a view count, whatever it is. Nobody controls it, free distribution depends on a ranking we cannot see, and a view target moves the work towards what gets watched instead of what gets sold.
We will refuse to lay a protected song over a commercial video, even on request, and to deliver an edit without the rushes. The first exposes the client to a takedown at the worst moment; the second makes them dependent on us for the smallest change, which is not a service but a hostage arrangement.
What we do: the shot card — which question, which shot, what length, what noise to switch off; a first day filmed with your team rather than instead of it; subtitles and thumbnail treated as copy rather than as decoration; and the handover of the original files, named and filed, alongside the published versions.
And what you can do alone this week, without us: take the three questions that come back most often in your messages, film three thirty-second answers on a tripod, with no music and the fridge switched off. Publish them. If the volume of messages drops, you know what to produce next — and you also know you can produce it yourself.
Frequently asked questions
How long should a product video be?
Length is a result rather than an instruction: five to ten seconds to show an object and its price, twenty to forty for a demonstration or a gesture, two to four minutes for instructions that are not published in a feed but kept for the product page.
Do we need professional equipment?
No. A phone on a thousand-dinar tripod, near a window, strip lights off. The only production difference a viewer registers is stability, and the only unrepairable error is sound — no camera corrects a running fridge.
Why do our videos sound bad?
Nearly always because of a steady noise nobody hears any more: a compressor, air conditioning, a fan, the street through an open door. Switch them off while filming, come in to forty centimetres, and film in a room with fabric rather than an empty tiled space.
Are subtitles essential?
In a feed, yes: the video starts with no sound and the burnt-in text is the message. Write them short, at the top or the middle, never in the bottom fifth of the frame, which the platform interface covers.
Can we use a well-known song?
Not on a commercial video: the use is unlicensed and the takedown arrives once the video starts working. It also covers the voice, which was the content. No music on a demonstration, quiet royalty-free music on an edit without speech.
What view count should we aim at?
None. Distribution is national and your customers may not be: three thousand views producing eleven serious messages beat a hundred thousand producing none. Count messages, calls, visits and accepted orders.
Where we come in
Three filmed answers, and messages counted the week before and the week after: that is the only honest measurement available at your size.
- We read that difference with you, and say what it permits you to conclude.
- We prepare the shooting plan question by question, with the durations.
- We cut the sound nobody hears while filming and everybody hears afterwards.
If your catalogue does not yet hold six decent photographs, video is not your next expense: it amplifies what already exists.
Read next
Filming without closing: shooting inside a business that is open
The equipment is not the problem. The problem is filming while you serve, with customers in shot and one pair of hands.Content creation: photography is not a talent, it is a process
One corner, one surface, four angles, always the same. What decides is not the photographer’s eye but the repetition.TikTok ads: attention is still cheap here, and it will not last
Followers distribute nothing here: every video is judged alone. A new account can beat an old one on its first try.
Let us talk about your project
A free audit, no commitment: we look at your online presence and tell you what is holding it back.