Video is worth making when the subject has to be seen in motion. That is the whole test. If you can make the same point in two clear paragraphs on a page, the page is cheaper, faster to change, easier to find, and more likely to be read.
Most of the video budget I see in small B2B and ecommerce companies is spent failing that test. The work is competent. It just does a job that writing already did, at far higher cost and with none of the ability to edit it next month.
Where video genuinely beats text
A physical product in use. Weight, scale, how a mechanism moves, how a fabric falls, how the parts go together. Writing has to describe these and the reader has to reconstruct them. Thirty seconds of footage skips the reconstruction entirely. This is the strongest case there is, and it is the one most product companies underinvest in while shooting lifestyle content instead.
A process that is awkward to describe. Installation, onboarding, a service visit, a workflow inside software with six screens. If your written version keeps needing the phrase “then you”, that is a signal. Watching someone do it once replaces a page of sequencing.
A face on a high-trust purchase. If someone is about to hand a stranger a meaningful amount of money, or let them into their house, they want to see who they are dealing with. One unpolished minute of the person who will actually do the work beats a well-produced piece featuring nobody identifiable.
A customer saying it themselves. This is the one case where video does something text structurally cannot. A written testimonial could have been composed by your marketing team, and your reader knows it. A real person on camera, with their own phrasing and their own pauses, cannot be faked in the same way. Record these badly rather than not at all.
Where it usually does not pay back
The explainer video. The genre exists because it is easy to sell, not because it works. Most explainers restate the homepage over animated shapes, and most viewers leave early because the opening is a problem statement they already agree with. If you want one anyway, lead with the demonstration, not the setup.
The brand film. Drone shot, slow piano, words like “passion” on screen. These are made to be approved internally, which is why they get approved. They rarely answer anything a buyer asked. A company with fifteen people and a positioning problem should fix the positioning in writing first, where it costs nothing to revise.
Social video on a schedule. The worst version is a quota: four clips a week because the plan has a row for it. Output set by a calendar rather than by having something to show produces filler, and filler trains your audience to scroll past you. If you want to post consistently, build a backlog from things that were worth filming anyway, such as real installs, real questions, real unboxings.
What production actually costs
The camera is the cheap part now. A recent phone, a window or one soft light, and a clip-on microphone will cover product demos, process walkthroughs, founder pieces, and customer interviews. Bad audio reads as amateur far faster than soft focus does, so if you buy one thing, buy sound.
A crew earns its fee when the subject genuinely needs it: multiple cameras on one take, a location you cannot control, machinery or motion, anything shot once and never repeatable. Hiring one to film a person talking at a desk is paying for insurance you do not need.
The real cost sits in the two places nobody budgets:
- Scripting. Deciding what the video is for and cutting everything else. This is the part that makes a video watchable, and it is writing. If the brief is vague, no amount of production rescues it.
- Editing. Hours per finished minute, every time you change anything. This is also why video resists iteration. A sentence on a page takes a minute to fix. The equivalent fix in a finished video means going back to an editor.
Budget those two honestly and the comparison with a written page gets clearer, which is usually the point.
The distribution problem
An unwatched video is the most expensive asset in marketing, because the entire cost is paid up front and the salvage value is near zero. A page that underperforms can be rewritten, re-titled, and still get found in search a year later. A video that underperforms is finished footage you either re-cut at cost or leave sitting there.
So decide three things before you shoot:
- Where it lives. A specific page, a specific email, a specific step in the sales conversation. “On our YouTube channel” is not a placement.
- What it replaces. If it is not displacing something weaker, you have added work instead of improving it.
- How it gets reused. Interviews cut into clips. A long demo yields the three moments people actually ask about. Plan the cuts before the shoot and you get them. Plan afterwards and you get one long file and a note to do it later.
That last point is where most of the waste happens. Reuse is almost never discovered after the fact, because the footage was never shot with it in mind.
Deciding in practice
Run a short test on any video someone proposes. What question does this answer? Where would a buyer be when they ask it? Could a page answer it instead? If the page could, write the page, and put the money toward making it easier to find, which is what the conversion work on your own site is for.
When the answer is that it has to be seen, make it, keep it plain, and plan the placement before the shoot.
If you are not sure which of your content gaps are video-shaped and which are page-shaped, that is the kind of question the 30-Day Marketing Diagnostic settles with evidence from your funnel rather than opinion. Talk to a CMO about where your spend is going.
