How to make a SaaS launch video

What a launch video is for

One job: make a stranger who scrolled past your announcement want to click the URL. Not to explain the product completely, not to close the sale, not to train the user. Every decision in the film is downstream of that, which is why the most common structural mistake is treating it as a compressed demo.

If the viewer already wants the product, they need a demo. If they do not understand the category, they need an explainer. The launch film is for the person who understands the category, has never heard of you, and gave you four seconds.

The six-beat structure

Sixty seconds, six beats. The timings below are a working default rather than a rule, but the ordering is load-bearing: every beat exists to buy attention for the next one. You can watch the same structure at work in the launch films worth studying.

Beat 1: the hook

Five seconds, one line, and the line has to be about the viewer rather than about you. "Close your books in hours, not weeks" is a hook. "Introducing Ledgerly, the modern accounting platform" is a title card. If the first frame is your logo, the hook is missing.

Beat 2: the problem

Ten seconds naming the current state precisely enough that the right viewer flinches. Precision is the whole trick: "manual work is time-consuming" describes nothing, "four people, three spreadsheets, and a Slack thread every month-end" describes a Tuesday. Specific problems are also what make the film feel like it was made for one audience rather than any audience.

Beat 3: the reveal

Seven seconds. Product name, and the promise in one sentence. This is the only moment the film is allowed to be about you, and it works because beats one and two earned it.

Beat 4: capability

Twenty seconds, two or three capabilities, no more. The instinct to list eight features is the single most reliable way to make a launch film forgettable: eight things at 2.5 seconds each is a list nobody retains, three things at 6 seconds each is a product people can describe to a colleague. Pick the three that support the promise from beat three, and cut the rest.

Beat 5: proof

Ten seconds of the most defensible thing you have. A real metric, a customer count, a recognisable logo, a benchmark. If you have nothing yet, say something concrete instead: what it costs, what it integrates with, how fast setup is. An unsupported superlative is worse than no proof beat, because it tells the viewer the rest was marketing too.

Beat 6: the end card

Eight seconds. Logo, URL, one call to action, held long enough to read and screenshot. One action, not three. "Start free at ledgerly.com" beats "Sign up, book a demo, or follow us", which converts on none of the three.

Ten failure modes

1. Opening on the logo. Nobody is waiting for your logo. Open on the hook.
2. Feature lists. Eight capabilities at 2.5 seconds each is a list, not a film.
3. Jargon in the hook. If the first line needs a glossary, the viewer is gone.
4. Unreadable type. Small text over busy footage fails on a phone, which is where it will be watched.
5. Depending on sound. Autoplay is muted. Assume it always is.
6. Stock footage of people in offices. It says nothing about your product and everyone recognises it.
7. Running long. Past ninety seconds you are making a demo with a launch film's structure.
8. Three calls to action. Choose one.
9. Claims you cannot support. The proof beat is where trust is won or lost.
10. One aspect ratio. A 16:9 film cropped to vertical loses the type. Compose for each shape.

Choosing reference films

Whether you are briefing an agency or a generator, references beat adjectives. Pick two or three films you would be happy to be compared to, and prefer launches from products in an adjacent category rather than in yours, so you inherit the craft without inheriting the resemblance.

• Pick films with a similar information density to what you need. A film with four words in it is not a useful reference for one that has to carry three capabilities.
• Pick for pacing you can live with. Fast cutting looks great and is unforgiving of long claims.
• Watch them muted first. If a reference stops working without sound, it will not work for you either.
• Say what you are borrowing: the cadence, the palette, the type treatment, or the tone. Not all four.

supaintro works the same way internally: its reference library is analysed frame by frame into named traits rather than described in adjectives. That is the whole argument of the product, and it is also just good briefing practice regardless of who makes your film.

Deliverables to ask for

Specs per platform, including where the safe area is on vertical, are in the specs guide.

Launch day sequencing

1. Two weeks out. Lock the positioning and the three capabilities. This is the part that cannot be rushed later.
2. One week out. Film approved, cuts exported, captions checked on a phone with the sound off.
3. Launch morning. Vertical cut on social, 16:9 on the site above the fold, still frame as the Open Graph image.
4. Launch day. Post the film natively per platform. Uploading to each beats linking out from all of them.
5. Week two. Re-cut the strongest 15 seconds for paid. This is where a re-renderable project pays for itself.

If launch day is a Product Hunt launch, the film also takes the listing's first gallery slot, which has rules of its own: the Product Hunt launch video guide covers them.