The product demo video is the ninety-second version of your front page

Your front page has about ninety seconds to explain what your software is for. That is the length of a product demo video, and it is enough, provided the film is asked to do one job rather than four.

A demo film is not a shorter version of your documentation, and it is not a recorded sales call. It is a film with an argument: what the software does, who it is for, and what changes when you use it.

A demo film is not a screen recording

A screen recording is evidence of what happened on one machine on one afternoon. A software demo video is an argument about what the software is for. The difference is not polish. It is who decides the order.

A recording inherits the order of the product: the tab you open first, the menu you remember, the dialog that takes four seconds to appear, the cursor hunting for a button that moved last month. The viewer learns a layout, when what they needed was the point.

The film reverses that. It starts from the sequence of the argument, shows the parts of the interface that support it, then lands on the result the viewer came for. If it had to be recorded again because a button moved, it was a recording. A film that survives a redesign shows outcomes.

Decide which one job the film does

Three jobs account for most of the demo films worth making. Write yours in one sentence before anything is drawn, and keep that sentence beside the notes.

• Activation. The viewer has signed up and has not reached the moment where the product is obviously worth the effort. The film shows the shortest route there.
• Upgrade. The viewer uses one part of the product and has never touched the part you charge more for. The film shows what that part makes possible.
• Evaluation. The viewer is comparing you with an alternative and has committed to neither. The film explains what the software does and what adopting it asks of them.

Evaluation is the one that stands in for your front page, and it is the hardest, because the viewer arrives with no context to hang anything on. Activation and upgrade films can assume the product is understood. An evaluation film has to earn that understanding in the first half minute. Where the buyer has never met the category, the film is also doing the work of a SaaS explainer video.

If you cannot choose between two jobs, that is an argument for two films rather than a longer one. The same budget buys one careful film, three related ones or a running arrangement, and one film or three is a decision you can make in an afternoon.

The order of information for a sceptical buyer

A buyer who does not trust you yet reads in a different order from one who does. This sequence survives scepticism, and it is not the order of your navigation or your pricing table.

1. What it is, in one sentence. A stranger should be able to repeat it after hearing it once, in their own words.
2. Whose problem it solves. Name the job they already do badly, in the words they use for it.
3. What changes. Before and after, shown as work rather than as the same dashboard with different numbers.
4. What adopting it costs. Time, migration, training, the habit they have to give up. An objection answered before it is raised never becomes a reason to walk away.
5. Proof that it works. The moment inside the product where the result appears, shown once, at a size where it can be read.
6. What happens next. The trial, the call, the price. If the viewer has to hunt for the next step, the film was entertainment.

Two notes on that order. The price belongs in the film, near the end, because a demo that hides the cost produces a first call that opens with a correction. And the proof should be the product's own output: a diagram of your architecture is proof only if the buyer is buying the architecture.

When to show the interface, and when to abstract it

Nothing requires a software demo video to show software. The rule is about recognition. Show the interface when the viewer will have to recognise it later; abstract when the point is something the interface cannot show.

• Show it when the viewer is about to use that screen themselves, when the objection is that the product looks complicated, or when the placement of one control is the convenience you are selling.
• Abstract it when the value lies in a sequence across several screens, when what changed is invisible (a permission, a reconciliation, a queue emptying), or when the interface is dull. Three tables and a grey sidebar teach nobody anything at film size.

Whatever you choose, judge it at the size the film will be watched. A label captured from a laptop and dropped into a vertical frame is a smudge on a phone, which is where most of these films are seen. If the interface has to be shown, rebuild what matters at film scale: fewer rows, larger type, one thing highlighted.

The interface is evidence. The argument is what the software lets somebody do.

What abstraction must never do is imply a capability that does not exist. If the film shows a button that is not there, the demo call opens with a correction. Abstract the arrangement, never the claim.

What to leave out

Ninety seconds is made in the cutting, and the material that has to go is fairly predictable.

• Settings, permission screens and admin areas can go, unless administering the product is what you sell.
• Your navigation can go, because nobody learns a menu from a film and the menu will change first.
• Empty states, loading spinners and the first-run checklist can go, since they are the parts nobody chose.
• The integration wall can go. One integration shown working is worth more than twenty logos.
• Anything that needs a second viewing can go. A beat that only lands on the rewatch is not carrying its seconds.
• The feature you are proudest of can go, if it is not the feature the film is about.

Films run long because every department was given a slot. A product demo video with one owner and one job survives a hard cut to ninety seconds. A film assembled by committee cannot be cut at all, because every second belongs to somebody. This is where an explainer video for startups usually loses its shape: every feature still feels like part of the story.

The mistake of filming every feature

The most common shape of a bad product tour video is a walk through every tab, in the order the tabs appear. It feels thorough, and it fails for a reason that has nothing to do with length: a feature is an answer, and the viewer has not asked the question yet.

What a viewer takes from a full tour is a map of a place they have never been. They cannot hold it or use it, so the film becomes a list of claims with nothing connecting them, and a sceptic can only doubt them one at a time.

A full tour also ages badly. Every release makes it slightly wrong, and a film with forty features has forty places to go stale. A film about one outcome survives several releases, because outcomes change far more slowly than interfaces. When somebody asks for one more feature, ask which of the current ones should come out.

One film in the sales cycle, the help centre and on the pricing page

The same film can work in four places, provided it was built to do one job. Distribution is not the same as purpose, and none of these need a separate edit.

• The sales cycle. Send it after the first conversation rather than before it. A film that arrives before the buyer has a question goes unwatched; the same film afterwards answers what they went away wondering about.
• The help centre. Put it at the top of the article people reach when they are stuck, and let the text below carry the detail. The film answers what the feature is for; the article answers what to click.
• The pricing page. A film beside a price does more than a film beside a headline, because that is the moment somebody decides whether the number is reasonable.
• The onboarding funnel. The same file in the second email does the job of the paragraph nobody reads, which makes an onboarding funnel video the cheapest place to reuse work you have paid for.

One film travels further than most teams expect, because a Nine film can be saved wide, tall and square from the chamber. The pricing page, the phone, the feed and the slide deck each get a frame of their own. Decide the shapes before the film is written; adding one later means checking every scene again.

There are more placements than these four, and the full list of places a short film earns its keep is worth reading first.

It is also worth knowing the shape of the cost before the first meeting. In Nine that shape is a monthly plan, and the plans differ in how many films a month they hold and how long each can be, which matters for a product that ships often enough that the films have to keep up with it.

How to tell whether it is working

The useful signal is not views. It is where people stop.

Read the drop-offs as notes on the film.

• A fall in the first five seconds means the opening promised nothing. The first line has to name the job, not the company.
• A cliff in the middle usually sits on a feature that mattered only to the people who made the film.
• A drop where the interface appears means the screen was unreadable at that size, or it arrived before the viewer cared.
• Viewers who stay to the last second had already decided, so treat the closing call to action as a courtesy rather than a conversion.

Then compare placements instead of totals. A film that finishes on the pricing page and dies in an email is a film in the wrong place. The quieter check matters as much: if the questions in your inbox and on your first calls start referring to the film, it is doing work. If nobody mentions it, it is decoration.

You can see your colours in motion before paying for anything: Nine's free page reads your website and plays Nine's own films in your colours, with no account and no card. A film of your own comes with a plan, and a film Nine writes is judged before you see it against a bar of nine out of ten; how it works is on the front page.