In many small firms, video is treated as something you get to once there is spare time and spare budget — which means rarely. Yet the value for a small business is not in a polished production. It lies in answering a question customers ask anyway: how does this actually work with you? Sixty seconds of phone footage with clear audio, proper captions and a transcript underneath often answers that better than half a page of text. This article describes four formats that have proven themselves in small businesses, shows how to record them without a studio, and treats captions and loading behaviour as what they are: required parts, not optional extras.
Four formats that actually generate enquiries
The viewing habit is long established. Online video reaches most of the population: the largest video platform reaches 72 percent (ARD/ZDF media study 2025) of people aged 14 and over in Germany, the public broadcasters' media libraries more than 60 percent (ARD/ZDF media study 2025). 85 percent (ARD/ZDF media study 2025) of respondents use more than one streaming offering, on average 2.5 (ARD/ZDF media study 2025) offerings per person, while total daily media use stands at 6.5 hours (ARD/ZDF media study 2025) with a cross-media daily reach of 98 percent (ARD/ZDF media study 2025). Nobody needs to be told how a play button works.
That very familiarity creates a misleading benchmark. Next to professionally produced content, a company image film with drone shots and a music bed quickly looks like a weaker copy of something made elsewhere with entirely different means. The advantage of a business video lies somewhere else: it shows something no broadcaster and no platform can deliver — your own process, your own rooms, your own people. Videos with that purpose need no equipment, only a clear question they answer.
Four questions come up especially often in small firms. Each has a format that can be produced with reasonable effort and that works at a specific place on the site — not as decoration on the home page, but where the question arises.
How a first consultation works
The sequence from the first call to the written quote, in the order customers experience it. It removes the most common hesitation before making contact, because nobody has to guess what they are getting into. Place it directly above the contact form.
Team and workshop in 60 seconds
A walk through the premises with two or three people each saying one sentence about their job. It replaces the stock phrases found on many about pages with faces and voices. Place it on the about page, below the first block of text.
Before and after a real job site
The same room, the same camera position, two points in time, plus three sentences on the task. The strongest format for trades work, because the result speaks for itself. Place it on the services or reference page.
Arriving and finding the entrance
From the car park or the bus stop to the door, including the bell panel and the stairs or ramp. It saves phone calls, prevents late arrivals and doubles as honest information about accessibility. Place it on the contact or directions page.
These four formats share one trait: they tell nothing you could not also write down — they simply prove it faster. A process that reads like a statement of intent in text looks self-evident on video, because viewers can see that it genuinely exists. If you are still working out whether the site needs rebuilding at all, the groundwork is in the article on when a new website pays off; if you want to sort out audience and service descriptions before the camera runs, start with the copy on your service pages.
Where to start
Producing with what is already there
The equipment question is settled faster than most people expect. A smartphone from the past four or five years delivers footage that is good enough for a company website — provided it is held in landscape, kept steady and used in daylight. Landscape, because the video sits in a wide area on the page: a portrait recording there produces either black bars or a crop in which the top of the head is missing.
Audio beats picture
- Smartphone in landscape, on a tripod, a clamp or, failing that, a stack of books
- Lavalier microphone on the speaker's collar, roughly a hand's width below the chin
- Daylight from the front: the person faces the window, the camera has its back to it
- Quiet room with no running machines, radio or telephone; ventilation and compressor switched off for the recording
- Record a 20-second test and listen to it on headphones before starting the real take
- Check storage and battery beforehand; one minute at high resolution quickly takes several hundred megabytes
- Record every sentence twice so that the edit has an alternative ready
| Common assumption | What works in practice | Why |
|---|---|---|
| A camera is needed first | The phone you own is enough, the microphone is not | The image quality of current phones exceeds what a website delivers; the distance to the built-in microphone is the real problem |
| The script has to be memorised | Five bullet points on a sheet next to the camera | Memorised text sounds recited; freely spoken sentences read as credible and are easier to cut |
| The shoot needs a free day | 45 to 60 minutes per format | With three setups and two takes each, the material is in the can before concentration drops |
| Music makes the video look professional | No music bed under speech | Music makes comprehension harder, raises licensing questions and makes captions more important rather than replacing them |
| An intro sequence belongs at the front | Start with the first sentence | The opening seconds decide whether people stay; a logo intro consumes exactly those seconds |
Editing stays just as simple. In most cases it is enough to trim the start and the end, cut out slips of the tongue and append a text card with phone number and address. Elaborate transitions, colour filters and animated name plates cost time without deciding anything. If the video ends up below a service description anyway, sharpen that text first — the article on website copy customers actually read explains how.
Length and structure: 30 to 90 seconds
Short is not merely a preference, it is common practice on the web. Of all videos embedded on web pages, 21 percent (HTTP Archive Web Almanac 2024) run up to ten seconds, 37 percent (HTTP Archive Web Almanac 2024) between ten and thirty seconds and 21 percent (HTTP Archive Web Almanac 2024) between thirty and sixty seconds. Only 12 percent (HTTP Archive Web Almanac 2024) fall between one and two minutes, 9 percent (HTTP Archive Web Almanac 2024) exceed that. More than half of all embedded videos are therefore shorter than half a minute. A business video may run longer than an animated preview, but it has little reason to pass the ninety-second mark.
The reason is not a lack of attention span but the situation: a video on a company website is rarely opened out of interest in film, but to settle a specific question. Once the question is answered, attention ends and the remainder plays to nobody. So the content is ordered with the answer at the front rather than at the end of a dramatic arc.
- First sentence: who is speaking and what the video is about — no greeting formula in front of it
- Second to fourth sentence: the answer in the order in which it happens in real life
- Middle section: one concrete number, deadline or example that backs the statement up
- Second-to-last sentence: the next step from the customer's point of view
- Final sentence plus text card: how to get in touch, with phone number and location
- Then cut — no closing sequence, no afterthought, no repeated goodbye
A logo intro is the one second of a video that nobody asked for.
For the same reason, automatically playing the next video does not belong on a company website. Whoever has the answer should be able to read on or pick up the phone without having to stop something first. A single well-placed video beats a media library of six items, five of which nobody opens.
Captions are a required part, not a bonus
The Web Content Accessibility Guidelines are unambiguous here. Success Criterion 1.2.2 requires that „captions are provided for all prerecorded audio content in synchronized media", with a single exception for media that is itself an alternative for existing text and clearly labelled as such (W3C, WCAG 2.2). That criterion sits at Level A (W3C, WCAG 2.2), the lowest of the three conformance levels — squarely in the area of baseline requirements rather than nice-to-haves. For the legal framing in Germany and the EU, see the article on accessible websites under the accessibility act.
The group of people for whom this decides between accessible and inaccessible is large: an estimated 16 million people (German Association of the Hard of Hearing) in Germany live with a hearing impairment, and the number of people who are hard of hearing in the narrower sense is put at 3 to 4 million (German Association of the Hard of Hearing). On top of that comes a much larger group that benefits from captions without any hearing impairment: people watching on the move, on a train, in a waiting room or in an open-plan office, who keep the sound off out of consideration. For them, a video without captions is simply a moving still image.
What belongs in the captions
Caption track as a separate file
A WebVTT file next to the video, referenced through a track element. It can be switched off, it does not interfere with the composition of the frame, it can be corrected afterwards, and it still works when the video is embedded at a different size later.
Transcript as a text alternative
The full spoken text as running copy below the video, extended by whatever you have to see in order to understand the content. Success Criterion 1.2.3 explicitly allows such a media alternative (W3C, WCAG 2.2), which also covers visual information.
Speak in plain language
Short sentences, one thought per sentence, technical terms explained the first time they appear. What applies to written text applies doubly to spoken text, because listeners cannot jump back. The captions become easier to read as a side effect.
The transcript is the part that does double duty. It satisfies an accessibility requirement and it is readable page content at the same time: anyone who prefers to skim rather than watch finds the answer in text, and search engines and AI answer systems find material that would otherwise be locked inside the video file. A page with video and transcript therefore carries more substance than a page with video alone — without anyone having to invent extra copy. How to write text that actually gets read is covered in the article on accessible, clear website copy.
WEBVTT
00:00:00.400 --> 00:00:03.900
You call us and we find a slot
for the first consultation together.
00:00:03.900 --> 00:00:07.600
We bring a tape measure, a sample case
and a price list to that appointment.
00:00:07.600 --> 00:00:09.200
[Sound: workshop door closing]
00:00:09.200 --> 00:00:13.400
BERGER: You will have the written quote
within three working days at the latest.- Treat automatically generated captions as a draft and correct them line by line before publishing
- Check proper names, place names, figures, measurements and prices with particular care — that is where the embarrassing errors appear
- Show at most two lines at a time and break them at sense boundaries, not in the middle of a phrase
- Use punctuation: a full stop at the end of a caption makes reading along considerably easier
- Mark speaker changes with a name or a colon and add meaningful sounds in square brackets
- Do not burn captions into the picture; deliver them as a track that can be switched off
- Re-check the timings after every new edit, because every following caption shifts
Embedding without wrecking your load time
Video is no longer a fringe element on the web: 6.7 percent (HTTP Archive Web Almanac 2024) of pages fetched on mobile contain at least one video element, a relative increase of 32 percent (HTTP Archive Web Almanac 2024) over 2022. The way it is embedded lags behind that adoption. A poster image — the still frame shown before playback — appears on only 3 percent (HTTP Archive Web Almanac 2024) of video elements, a deliberate preload setting on 16 percent (HTTP Archive Web Almanac 2024), while 23 percent (HTTP Archive Web Almanac 2024) start automatically and only 8 percent (HTTP Archive Web Almanac 2024) of those are muted.
This matters because the first painted frame of a video counts as a candidate for the largest contentful element and therefore feeds straight into the most important loading metric (Google web.dev). Without countermeasures the browser also fetches more video data than needed: the default preload behaviour pulls part of the file before anyone has pressed play. For videos meant to start on user action, Google explicitly recommends preload="none" together with a poster image, because the attribute only concerns the video data while the preview image is still shown (Google web.dev).
<!-- Preview image immediately, video data only after the click -->
<video controls playsinline
poster="/videos/first-consultation-poster.jpg"
preload="none"
width="1280" height="720">
<source src="/videos/first-consultation.mp4" type="video/mp4">
<track kind="captions" srclang="en" label="English" default
src="/videos/first-consultation.en.vtt">
Your browser cannot play this video.
<a href="/videos/first-consultation.mp4">Download the video file</a>
</video>
<!-- Transcript right below, expandable and readable in the source -->
<details class="transcript">
<summary>Read the transcript of this video</summary>
<p>Hello, my name is Sandra Berger ...</p>
</details>- No automatic start with sound: Success Criterion 1.4.2 requires a way to pause, stop or separately control audio that plays automatically for more than 3 seconds (W3C, WCAG 2.2)
- Set a poster image and treat it like any other photo: appropriate output size, modern format, width and height declared
- Set preload to none when playback starts on click; use metadata only if the running time has to be visible up front
- Declare width and height or a fixed aspect ratio on the element so that nothing jumps while loading
- Fetch the video file after the click instead of holding it ready in the background
- Ship MP4 as the base format — 68 percent (HTTP Archive Web Almanac 2024) of video files on the web use it, and support is correspondingly broad
- Deliver the caption file as a separate .vtt file and give the track element a language and a label
- After embedding, check which element is actually measured as the largest contentful element
Keep the video out of the stage area
Embed or self-host
That leaves file size. One minute at 1080p with moderate compression lands somewhere around 8 to 15 megabytes depending on how much movement is in the frame — too much for an automatic start, but unproblematic when the file is only fetched after a deliberate click. If you want extra headroom, ship a smaller 720p version as well. The principle is the same as with photos, only with bigger numbers; the basics are in the article on image formats and output sizes. How the result feels on a small screen is best checked on a real device, which is the subject of the article on mobile usability.
One video per quarter: a rhythm a small firm can sustain
The most common mistake with video is not technical but a matter of resolve. Anyone who sets out to publish a video every month stops in month three and is left with a page holding two outdated items. Four videos a year, by contrast, is a workload that survives alongside day-to-day business — and after two years there are eight videos on the site, each working at its own place.
The topics do not have to be invented. They sit in phone notes, in follow-up questions about quotes and in the contact form messages of the past few months. If the same question has been asked three times, it is a video topic — without anyone having to weigh up whether it is interesting enough.
- Week 1: collect the three most frequent customer questions of the past quarter from notes and enquiries
- Week 1: pick one question and write down five bullet points for the answer, no more
- Week 2: schedule a one-hour shoot, prepare the location, check microphone and storage
- Week 2: record, two takes per setup, and review the material straight afterwards
- Week 3: edit, generate and correct the captions, derive the transcript from the captions
- Week 3: pick a poster image from a calm frame and prepare it like any other photo
- Week 4: embed, verify on a mobile device, check loading behaviour, publish
- Week 4: add the link to the quote email and to the contact form reply template
| Quarter | Trigger | Format | Place on the site |
|---|---|---|---|
| First quarter | Enquiries come in hesitantly, the process is unclear | How a first consultation works | Above the contact form |
| Second quarter | Season starts, many new prospects | Before and after a job site | On the services or reference page |
| Third quarter | Hiring push, applications are not coming in | Team and workshop in 60 seconds | On the about and careers page |
| Fourth quarter | Appointments on site, customers get lost | Arriving and finding the entrance | On the contact and directions page |
Budget the effort realistically
And a video is not a monument. If the process changes, the old video is replaced rather than supplemented, so that the page does not carry two contradictory statements. It pays to give media a fixed slot in the maintenance plan for the site — what such a plan looks like is described in the article on routine website maintenance.
Video as a finished block instead of a craft project
The list of technical points from the previous section is manageable, but it has to be worked through again for every new video. That is where the quarterly rhythm fails in practice: not at the filming, but at the point where someone copies the embed code from video three into video four, forgets the poster image and cannot find the caption track any more. The way out is not another tool but a building block in which those decisions have already been made.
Poster and loading behaviour preset
The video block expects a poster image, fetches the video file only after a deliberate click and carries its dimensions with it. The business uploads a file without touching attributes for preloading or aspect ratio.
Caption track and transcript as fields
Caption file and transcript are part of the block, not an afterthought. The transcript is rendered as readable page content and therefore stays usable for people who would rather not play the video at all.
Placeholder instead of a silent embed
If an external video does have to be embedded, a placeholder with an in-place release appears first. Without that release, no connection to the external provider is opened.
The result is delivered like the rest of the page: as static files, with no framework loaded afterwards and no embed code for anyone to write around. The overview of features lists the building blocks available; the process at a glance shows how a page gets from first description to publication, and the example sites show what finished pages with media look like. If you are still weighing up self-build, agency and AI-assisted setup, the criteria are in the comparison of routes to your own website.
The short version
Sources and studies