Show What a Product Demonstration Actually Demonstrates

Start with the sentence you want the viewer to say when the ad ends. Not the shot. The sentence: that lid stays shut. Or: nothing leaks in my bag. The shot is the thing you're building; the sentence is the thing you're claiming. A treatment that opens with its best image has usually skipped the sentence, and the sentence is the only part anyone can check.

So write both — the claim and the shot — and keep them together through every revision that follows.

One invented product runs through this piece: a lunch container with a hinged lid and a single latch on the front. It doesn't exist. It hasn't been tested, filmed, or reviewed, and the proposals below are writing exercises rather than results. That's deliberate. A treatment is written before anything is captured, so it can't establish that a closure happened, that the container is representative, or that a claim has support. What it can do is say what the finished shot is meant to mean, and what remains to be confirmed before anyone films it.

Put the intended inference beside the visible action

Here's a complete proposal, short enough to fit on a treatment page.

Product: lunch container, hinged lid, one front latch. (Invented.) Starting condition: container open on a counter, lid up, empty. Action: a hand presses the latch down. The latch closes. Camera: static, counter height, about a foot back. Lid and latch stay in frame. Copy: "Closes with one press." Intended inference: in this shot, the latch closed when a hand pressed it.

That last line carries more weight than the copy does. It states what a viewer is meant to take from the frame, and it stays inside what the frame actually shows: an empty container, one press, one closure, all in view. Someone reading the treatment knows what to film and, just as usefully, knows what the finished shot is allowed to mean.

Now write down what the shot doesn't establish, because a viewer may supply some of it without being asked:

• whether the latch stays shut while the container is carried, dropped, or filled
• whether the container in the shot came from regular production or was prepared for filming
• what happens on the second press, or the hundredth
• how many attempts the closure took before one looked clean

None of those are failures of the proposal. They're the list the proposal hands to whoever is responsible for the product's claims, and naming them makes the treatment more useful, not less.

The distinction to hold onto is between two kinds of sentence. The latch closed in this shot describes an observable event. This container won't leak is a claim about the world. Images carry claims too, and not only headlines do. The US Federal Trade Commission's advertising guidance for small businesses discusses claims the words state directly and claims the ad conveys overall, and it asks whether the evidence offered relates to the claim being conveyed. A filmed action and the sequence it's cut into both count as part of that overall meaning. US federal guidance is one frame of reference rather than a worldwide rule, and it doesn't decide whether any particular demonstration is adequate — that call belongs to whoever reviews claims for the product, with the actual evidence in front of them. (FTC guidance)

What you can do as the person writing the treatment is narrower and simpler: name the inference, at the resolution the shot supports, and hand it to the reviewer.

There's a second half to this that's easy to skip. An observed instance is one instance. Even with every condition disclosed, a single filmed closure wouldn't establish how the latch behaves in general — on a shelf, in a bag, after a winter of daily use. That's the difference between a demonstration and evidence of typical performance, and a treatment that blurs the two has promised something the shoot can't deliver.

Write the conditions into the proposal

A demonstration is an action that happens under conditions. If the treatment doesn't say what they are, the shot has them anyway — the audience just can't see them.

For the container, the conditions worth writing down are plain. What's inside it. What state it starts in. Who handles it. How long each step takes. How many units appear on screen. What stays in frame and what doesn't. And whether the shot you're watching is a single attempt or the accepted one out of several, which matters more than it sounds: a demonstration cut from take fourteen looks identical to one cut from take one.

Duration and quantity are conditions too, and they're usually left vague. If the container sits closed for two minutes before the bag shot, the claim being supported is a claim about two minutes. If the "after" image is captured three hours later, that's a different condition and a different claim. If the shoot uses one container, the treatment can say so and let the production team decide whether to film more; if it uses three, the treatment should say which one appears where, because a viewer watching three containers close will read that as three.

Don't invent a test protocol to fill the gaps. If you don't know the conditions, write them as questions addressed to the product owner: Does the container we're filming ship with the same latch as the retail unit? Was it adjusted or warmed before the shoot? What was inside it during the bag sequence, and how long had it been closed when that shot was taken?

Before-and-after is where this gets sharpest. Two images placed next to each other invite the reading of one continuous event, whether or not they were one. Suppose the treatment proposes a "before" shot of the container upright in a tote bag and an "after" shot of the bag tipped over with the container still shut. The pairing invites the conclusion that the container stayed shut through whatever happened in between. But if the two shots were made at different times, or use different containers, or the tipping was never photographed at all, then the continuity is supplied by the cut rather than by anything the camera saw. Production order isn't edit order either — the "after" can be filmed days before the "before," and the finished sequence will still read as cause and then result.

Framing sets another limit, and it's a limit on the claim rather than on the picture. A leak claim is bounded by the smallest thing the lens resolves. A wide shot of a bag can't show a spoonful of soup soaking into a lining. A treatment that plans to carry a no-leak claim from a wide shot has a camera decision to reconsider, not only a copy decision.

A demonstration shows behavior; an illustration shows an idea

A functional demonstration asks the audience to understand something about how the product behaves. A stylized illustration expresses a feeling or an idea without documenting behavior. Both are legitimate. They promise different things.

Take the latch closing in extreme slow motion, room tone dropping to a single low note as the lid seats. That can be a lovely piece of filmmaking. It's also no longer a record of anything; it's a composed impression of security. Which is fine, if the treatment says so and the copy travels with it.

The distinction collapses when the finished piece doesn't signal which one it is. Real hands, a real counter, natural light, no stylization markers — the finished piece invites the reading that it's documentation. Surrounding context shapes what an ad conveys, which is the same ground the FTC guidance works from. And here's the awkward part: writing illustrative in the production notes settles nothing for the viewer, who never sees them. It settles things for the person reviewing the treatment — and only if the rest of the proposal agrees with the label.

The hybrid is the common case, and it's where most of the ambiguity lives. A captured press, then slow motion, then layered sound, then a grade that turns gray plastic into brushed metal. Each of those moves the shot partway from document toward illustration. None is wrong by itself. But the treatment should say where the shot lands, because the words have to match the landing. If the picture is an impression, the copy can describe an impression. If the picture is offered as a record of an action, the copy can't outrun the action.

The useful question isn't "is this shot fake?" It's "what is this shot asking someone to accept?" A slow-motion press asks to be believed as a press. A slow-motion press with a designed click and a swelling chord is asking to be believed as reassurance. Those are two different requests, and they belong to two different parts of the approval chain.

Follow the cut: what framing and editing add

Every compression, comparison, and effect changes the claim. A few, applied to the same container:

Cutting from the latch closing straight to the bag removes the interval. Nothing in those two shots establishes that the lid was still shut in between; the cut asks the viewer to assume it. If the assumption is the point, say so. If it isn't, film the interval.

Adding the click in post is worth flagging rather than deciding at the desk. If the sound that tells someone "that's shut" was layered in by an editor, the finished piece is asserting a mechanical result the captured action didn't produce. That belongs in front of the reviewer, who can look at the raw footage and the sound design notes together.

Slow motion makes a short drop look like an impact. If the toss is the only thing standing behind a durability impression, the speed ramp is doing the claiming.

Any comparison implies conditions on both sides. A side-by-side of two containers means the same fill level, the same drop height, the same surface, unless the treatment states otherwise. Unstated conditions aren't neutral. They're whatever the viewer supplies.

None of this gets fixed by polish. A crisp image of one event is still one event. Resolution, grade, and mix change how convincing a shot feels, not what it establishes. When a proposed effect would carry more than the captured action can, you have two moves: revise the proposal, or get the confirmation that would let the shot mean what you want it to mean. Whether there's adequate support for a claim about typical performance is a separate question from this editorial pass, and it goes to the product owner and whoever is qualified to review the claim.

Two honest revisions

Stay inside the frame. Drop the toss. Keep the press, the closure, and the copy about closing. The intended inference never leaves what the camera saw, and the treatment is complete without a single open question. It's a smaller ad. It's also one nobody has to defend.

Route the bigger claim. Keep the toss as an illustration of carrying a lunch, labeled that way, and move the durability and leak claims to where they can actually be answered. The treatment then proposes a shot, states what the shot is for, and states plainly what is still unconfirmed — who will confirm it, and what a demonstration supporting the broader claim would need to show.

Either way, the page reads better when three lines stay separated:

• Proposed action: what will be filmed.
• Intended inference: what a viewer is meant to conclude from it.
• Unresolved confirmation: what nobody has established yet, and who has to.

A demonstration proposal is a plan for showing something, and plans are allowed to be exact about their own limits. The impressive version of the shot is still available. It just isn't finished business, and it shouldn't be written as though it were.

Frequently asked questions

What should a product demonstration treatment claim first?

Start with the sentence the viewer should say when the ad ends, such as that lid stays shut or nothing leaks in my bag, not the opening shot. The shot is the thing being built; the sentence is the claim, and the two should stay together through revisions.

Why include an intended inference line?

It states what the viewer is meant to take from the frame and stays inside what the frame actually shows, like an empty container, one press, one closure. It tells someone what to film and what the finished shot is allowed to mean; it does not establish broader claims.

What does a single filmed closure not establish?

It does not establish whether the latch stays shut while carried, dropped, or filled; whether the unit came from regular production or was prepared for filming; what happens on the second or hundredth press; or how many attempts preceded a clean closure. An observed instance is one instance, not evidence of typical performance.

How should conditions and before-and-after edits be handled?

Write conditions into the proposal: contents, starting state, who handles it, duration, number of units, what stays in frame, and whether it is a single attempt or the accepted take. Two images placed together invite a continuous-event reading even if shot at different times or with different containers; unstated conditions are supplied by the viewer.

What separates a demonstration from an illustration?

A functional demonstration asks the audience to understand how the product behaves. A stylized illustration expresses a feeling or idea without documenting behavior. Both can be legitimate, but they promise different things, so a hybrid treatment should say where the shot lands and make the copy match that landing.

Further reading

• Side-by-Side Products Make a Comparison—even Without Comparison Copy — When two containers share a frame, this shows how timing, frame share, and edit order produce a comparison the pictures assert without any comparison copy.
• Make the Commercial’s Hero Product Look Like the Product People Will Receive — When a proposal films a prepared unit rather than a retail one, this separates presentation choices from what the frame implies a buyer will receive.