Video on the product page: what it has to do to earn its place

A video on a product page has one job, and it is narrow: answer the objection your images leave open. Everything else belongs on your homepage, in an ad, or nowhere.

That framing is useful because it turns a vague question ("should we do video?") into a specific one you can answer per product: what is the thing a buyer still does not know after looking at the photographs?

Start with the objection, not the benchmark

You have probably seen the percentages that get quoted about video and conversion. We are leaving them out, because we cannot trace them to research we are willing to put our name next to.

What we can tell you is mechanical and checkable: a buyer who is unsure about something does not buy, or buys and sends it back. If a short clip removes that specific uncertainty, it does work. If it does not address it, it is a nice asset in the wrong place. The rest of this article is about finding the uncertainty, and at the end there is a way to measure the effect on your own catalogue rather than on someone else's.

Where to find the objection

You already have the answer written down in four places, and none of them is a brainstorm.

Your returns reasons. The most direct source there is. "Smaller than expected", "colour different", "did not fit our cabinet", "harder to assemble than I thought". Each of those is a video brief.

Your support inbox. Group the questions that arrive before a purchase by product. The question that recurs is the objection, and the fact that people are asking means the page did not answer it.

Review text, the three-star ones. Five stars tell you what you did right, one star is often shipping. Three-star reviews are where people describe the gap between what they expected and what arrived.

What your dealers get asked. If you sell through dealers or retailers, their staff field the question at the counter all day. This one is usually the fastest to collect and the least used.

Run that for your twenty best-selling products and you will not have a list of twenty different objections. You will have four or five, which also means far fewer clips than you feared.

Four common objections and the clip that answers each

Scale. "How big is it really." Show the product next to something unambiguous in the context where it is used, and show it moving, because a hand entering the frame fixes the size instantly in a way a dimension drawing never does.

Mechanism. "How does it open, close, fold, connect." One continuous take of the real product, no cuts. The cut is what makes people suspicious, because a cut is where the difficult part is usually hidden.

Material. "What does it actually feel like." Raking light across the surface, slow movement, close enough to see texture. This is where fabric, stone, wood and brushed metal either convince or do not.

Use and assembly. "Can I do this myself." Not a full instruction film, just the part people fear: the step that looks hardest in the manual.

Four objections, four clip types, and you can produce them for a whole range once you know which product needs which. Which of those can be generated and which need the real product is the subject of which clips can be generated.

The one-sentence brief

Before anything is produced, write this down and refuse to proceed without it:

This clip exists so the buyer can see that ______.

If you cannot finish the sentence, you do not need a video for that product yet. If the sentence is "our brand is premium", that is a brand film, and the product page is not where it earns its keep. We made the same argument for stills in which image sells where.

What it has to do technically

Short. Ten to twenty seconds for an objection clip. If it is longer, it is answering more than one thing and should be two clips.

Readable with the sound off. Assume silence. If the point only lands with a voice-over, add a line of on-screen text.

High on the page , near the main image, not below the specifications. A video nobody scrolls to has no effect to measure.

Consistent with the photography. Same background logic, same light direction, same crop discipline. A clip that looks like it came from another brand undoes the trust the rest of the page built, which is the whole argument in consistency reads as trust.

One file, several formats. The same clip will be asked for vertically by social and squarely by a dealer feed. Plan for that at production rather than afterwards.

Three ways it goes wrong

It autoplays with sound. Still happens, still costs you the visitor. If it plays by itself, it plays silently, and it is short enough that a loop is not irritating.

It is a brand film in disguise. Thirty seconds, a logo sting, three settings and no answer to anything. That clip has a place, and the place is not three centimetres from the buy button.

It contradicts the photographs. Warmer light, a different white, a slightly different colour on the product itself. The buyer does not diagnose that, they just feel that something is off, and a page that feels off converts worse than a page with no video at all. Generated clips are particularly prone to this, which is why the colour check belongs in your process rather than in your hopes.

Measure it on your own catalogue

Here is a method that costs nothing.

Pick ten products that share one objection. Put the clip on five of them and leave five as they are. Wait for a normal sales cycle, usually four to six weeks. Then compare three things between the groups: return rate, support questions about that specific objection, and conversion. The first two move before conversion does and they are far less noisy, which makes them the better early signal.

That is a number you can defend internally, because it is yours.

Food is a category where the objection is almost always portion and pack, which we worked through in pack, portion, proof. Once you know what your clips need to do, how to have it made is the next question.

Want to see how we produce product video and stills from one setup? See how it works, or try LensLab free with five credits.