Forty real photographs, and a darkness chosen one by one

The product description is already written. The ingredients, the size, how to use it — it is all there. And still, the fact that the thing is good does not fit anywhere in the text. What you would notice in half a second holding it takes a long paragraph to write, and a long paragraph does not get read. I think that is what people mean when they say words alone do not carry how good a product is.

We did the same thing with our own sample films. The 15-second films on the KAZENA Motion page are not video files; they are drawn inside the browser with HTML, CSS and JavaScript. In the first version everything was made of those drawn shapes. The coffee cup, the steam, the sea, the room — lines, surfaces and colours laid over each other. Lined up together they looked tidy. Tidy, but the coffee did not look like coffee.

So we decided to put real photographs in, into 7 of the 11 sample films. Kopi Senja for food and drink, orbita for fintech, SONIKA for a music festival, PULSE for fitness, Teluk Nusa for a resort, KAIROS for street fashion, ARUNIKA for property. Forty photographs in all, converted to 1280px WebP.

Before choosing which photographs to use, we decided which ones we would not. First, only photographs distributed under a free licence — the free part of Unsplash. Second, no photograph showing a real brand, venue or artist, however good it looked. If a real company's logo appears in a film for a fictional brand, viewers take it as that company's film, and for a sample that is simply not correct. We keep the credits for every photograph we used in a list in our own internal documents, and the corner of each film says the photographs are borrowed.

The biggest change was Kopi Senja. That one was not adjusted; it was rewritten from scratch. The cup built in CSS was dropped and replaced with photographs cutting quickly from one to the next. I had put time into that cup, so deleting it stung a little. But set the drawn cup and the photographed cup side by side and there was no argument left about which one looked like a drink.

Adding photographs immediately produced a different problem: text on top of a photograph cannot be read. White letters disappear in the bright areas and black letters sink into the dark ones. So we added a way of handling photographs to our film toolkit — slow pushes and pans tied to time, darkening, a colour laid over, black and white, blur — and a background that switches photographs by time range.

Of all of those, the one we ended up doing by hand was the darkening. I wanted a single setting that darkened everything by the same amount, and it did not fit. A bright photograph still swallows the letters at that amount, and a photograph that was dark to begin with just turns muddy when you darken it. In the end we chose the darkness one photograph at a time. What that taught me is that whether text can be read depends less on its size or colour than on whether the photograph you chose can take being darkened.

We fixed how we check, too. Ten frames per film, captured on both a computer screen (1440×900) and a phone screen (390×844), looked at in order. There are 4 things to look at: whether the text over the photograph is readable, whether text or photographs are cut off on the phone, whether the first frame goes blank while the photographs load, and whether the representative frame shown in the sample list is a good moment. Every one of them needs an eye; none of it comes out of the code.

That is our own work, but I think the same thing happens to anyone using photographs of their product. One photograph of the actual thing carries more than a drawing or a heavily decorated image, because the viewer takes it as a sign that the thing exists. And if you intend to put text over that photograph, it saves trouble later to check, while choosing it, whether it survives being darkened.

We decided what we ask customers to prepare on the same basis. Product photographs taken on a phone are enough to work from. Cutting the background out and correcting the brightness are ours to do inside the plan, and we do not come out to shoot. For an app, screenshots are enough. We did not want this to be something you can only order once you have had professional photographs made.

What we ask for is the product and 1 thing you most want to say about it. It does not have to be in finished words. We work out the structure and the wording, show you every scene as a still storyboard, and nothing moves until you approve it. We cannot promise a video will raise your sales. What we can promise is to narrow the single best thing about the product down to 1 point and make it land in a few seconds without anyone reading a line.

If words alone are not carrying how good your product is, start by taking 1 plain photograph. The thing itself, in the light you already have, with nothing arranged around it. Whether such a photograph exists does more than adding another sentence of explanation. What we did with 40 of them was, in the end, a way of confirming that.

Finally, where things stand. KAZENA Motion is not accepting orders, quotes or payments yet. We are partway through adding the video-production KBLI (business licence) to the company, and until it comes through the page is only an introduction. Every figure shown is a planned price. You can ask from the page to be told when we open. I write about licensing only as our own paperwork; I am neither a tax accountant nor a lawyer, so please confirm your own business registration, and your own use of photographs, with a professional.