The video sales letter is back, and long-form copy is the input

The video sales letter was largely written off around the middle of the last decade, mostly for reasons that had nothing to do with whether it worked. It was expensive to produce, impossible to iterate, and closely associated with a particular style of high pressure marketing that consumers had learned to recognise.

Production and iteration are now much cheaper, which removes two of those objections. The remaining work is editorial: restructuring sales copy for speech, choosing narration settings that fit the offer, and recognising the product categories where a video sales letter still does more harm than good.

Iteration became affordable

The production economics changed first. A video sales letter previously required a script, a presenter, a recording session and an edit, which put the cost of a single version high enough that most businesses produced one and left it in place. That made testing impossible, and an untested sales asset is a guess.

Tools that generate video from existing documents removed that constraint. A business with sales copy already written can produce a video version from the copy itself. Leadde accepts PDF, Word, PowerPoint and plain text files, or pasted text, and platforms offering video sales letter software work from the copy document rather than from a separately written script.

The second change is measurement. A written sales page reports scroll depth at best, and scroll depth is a weak proxy for attention. Video reports where viewers stopped watching, which is a considerably more direct signal about which part of an argument failed.

The third objection, the association with aggressive marketing, has not changed and is not a production problem. It is a copy problem, and it is addressed below.

Sales copy and video scripts are structured differently

Long form sales copy is written for a reader who can move at their own pace. Video is not, and this difference determines whether a conversion succeeds.

A written sales page front loads context. It establishes a problem, builds a case, and arrives at the offer some distance down the page. This works because a reader who is not interested leaves without cost, and a reader who is interested reads at speed. Neither behaviour is available to a viewer.

Video requires the claim first. A viewer decides whether to continue within the opening seconds, on much less information than a reader has. Nielsen Norman Group's work on video usability makes the practical point directly: viewers need to know what a video contains early, because they will decide whether to stay before they have received much else.

Three structural adjustments follow from that.

The claim moves to the opening line. Whatever the page argues, the video states in its first sentence, with the supporting case following for viewers who remain.

Proof gets compressed. A written page can carry six case studies because a reader skips five. A video carrying six is asking every viewer to sit through all of them.

The close arrives earlier and repeats. A written page has one call to action at the end and several inline. A video needs the action stated at a point most viewers actually reach, which is well before the end.

Narration has to sound like speech

Generated narration reflects whatever context it is given, and the difference between a usable script and an unusable one usually comes down to a few fields.

Narrative style is the most consequential. Leadde offers Formal, Neutral, Objective, Structured, Analytical, Explanatory, Narrative and Persuasive. For a sales letter, Persuasive and Narrative produce different assets and suit different products. Persuasive fits a well understood product where the argument is about value. Narrative fits a product whose benefit requires a scenario to become clear.

The audience field changes the register more than most users expect. Describing the audience as someone who has already compared two competitors and is deciding on price produces copy that addresses the comparison. Left blank, the same document produces a general introduction.

Detail level should be set deliberately and then held constant. Platforms typically offer several levels from summary through to comprehensive, and a sales letter is usually served by a shorter setting than the source document suggests, because written copy contains context that a viewer does not need.

Generated scripts require reading before publication in this category more than in most. Automatic summarisation removes qualifying clauses ahead of anything else, and in commercial copy the qualifier is frequently what keeps a claim accurate. A guarantee that applies under specific conditions can lose those conditions and become unconditional, which is a compliance problem rather than a style problem. The FTC's guidance on advertising and endorsements is the relevant reference for anyone making claims or featuring testimonials.

Where a generated presenter is used, disclosure is straightforward and worth doing. Businesses that conceal it are exposed to a discovery that costs more than the disclosure would have.

A drop in completion identifies weak copy

The reason to test video versions is that they produce a specific diagnostic which written pages do not.

Leadde reports impressions, total and unique views, shares, watch time, completion rate and interaction counts. Where viewers consistently stop at the same point, the cause is almost always the copy at that point rather than the length overall. A claim that is not supported, a price introduced before value has been established, or a transition that skips a step will show as a drop off across viewers.

This is usable in a way that page analytics generally are not. Duplicating a video, revising one section and running both is a short exercise, since the source is a document rather than an edit project.

Offers that are worse on video

Considered purchases with long evaluation periods are poorly served. A buyer comparing options over several weeks needs material they can return to, search and forward to a colleague. Video obstructs all three.

Anything requiring specification comparison should stay written. Where a decision depends on comparing figures side by side, a linear format is the wrong one.

Products bought on visual quality need real footage. Furniture, clothing and anything where finish and material determine the purchase cannot be sold by a presenter describing them. A generated presenter is particularly poorly suited here, since the format directs attention to a face rather than to the product.

There is also a limit worth stating plainly. This category of tool generates video from supplied content and does not record screens, so a software demonstration requires screen capture rather than a narrated description.

Treat video as another version of the copy

The video sales letter has returned because a second version is now inexpensive to produce. Its persuasive power has not changed; the economics of testing it have.

Keep the written copy as the source, generate video as one version of that copy, and use completion data to decide what needs revision. Producing one video and leaving it untouched recreates the old maintenance problem with cheaper tools.