Marketing video is one of the few formats where the viewer actively chooses to keep watching, and that choice is made in the first three seconds. Everything that happens before you press record — the decision, the script, the location, the microphone placement — decides whether that choice goes your way. This guide covers the full pipeline: how to decide what to make, how to shoot it cleanly with modest gear, how to edit it so people stay, and how to test and reuse it once it ships.
Every marketing video should be built backwards from one action you want the viewer to take. A video asked to do three jobs usually does none of them well: the viewer gets a brand story, a product demo, and a discount code in ninety seconds and remembers none of it. Pick one outcome, name the metric that tells you whether it happened, and let that choice drive the format.
| Objective | Natural format | Typical length | Metric that matters |
|---|---|---|---|
| Awareness | hook-driven story or short skit | 15–45s | 3-second hold rate, full-watch rate |
| Consideration | explainer or product demo | 60–120s | average watch time, click-through |
| Conversion | testimonial or offer walkthrough | 45–90s | signups, add-to-cart, replies |
| Retention | tips series or onboarding | 2–5 min | return viewers, feature adoption |
Three practical rules follow from that table. First, write the metric before the script — if you cannot name it, the video is a hobby project with a budget. Second, match length to intent rather than to platform defaults: a three-minute demo is fine when someone is choosing between two tools, and it is fatal when someone is half-scrolling on a commute. Third, plan the distribution surfaces early. A square crop of a widescreen interview loses the hands, the product, and often the punchline, so decide whether the master is horizontal or vertical and frame with the other crops in mind.
Audience research earns its keep when it changes a line of dialogue or a camera angle. Vague personas do not do that, so skip the demographics deck and go where your buyers already talk.
Sales call recordings, support tickets, comment sections, review sites, and search suggestions are the richest sources of raw material. If customers keep saying the tool “fights them on import,” your demo should open on import. If they describe the problem as “chasing approvals across three inboxes,” use that phrase in the voiceover — it signals recognition, and recognition is what keeps a viewer past the first five seconds.
Research only matters when it constrains a creative choice. Three examples:
Write these constraints down before scripting. Constraints save more production time than any piece of gear.
A strong short script usually carries five beats: hook, promise, proof, payoff, and call to action. The hook states a tension in the first sentence. The promise says what the viewer gets. Proof shows rather than claims. Payoff resolves the tension. The call to action is one sentence, spoken plainly.
Read every line aloud and time it. A natural speaking pace runs roughly 140 to 155 words per minute, pauses included. That means a 60-second video is about 140 to 150 words of narration — far less than most first drafts. Cut adjectives, cut setup, cut anything that exists only to sound polished.
Put visuals on the left, audio on the right. This makes it obvious when a line has nothing to show, which is usually a sign the line is filler. It also gives your editor a map: they can see where b-roll, screen capture, or graphics are supposed to land before the footage even exists.
Do not storyboard a talking head. Do storyboard product interactions, multi-shot sequences, graphic transitions, and anything involving movement between locations. Stick figures are enough — the goal is deciding shot order and continuity before the shoot day, not producing art.
Most bad shoots are bad for logistical reasons, not creative ones. Run this list a week ahead.
Send a short call sheet the night before: call time, address, parking, contact number, and the first three shots. People arrive calmer when they know what the morning looks like.
Composition is the cheapest quality upgrade available. Start with the rule of thirds and place the subject's eyes on the upper third line. Leave looking space in the direction the person faces, and keep headroom tight rather than generous — too much space above the head reads as amateur.
Because most marketing footage gets cropped, shoot with safe areas in mind. Vertical delivery needs the subject centered and the top and bottom third kept clear for captions and interface elements. Horizontal delivery tolerates off-center framing better. When in doubt, center the subject and leave clean negative space on one side; that space is where text overlays live later.
Backgrounds matter as much as faces. Move the subject a meter or two away from the wall so they separate from it, and look for depth: a doorway, a plant, a blurred office, a window. Focal length changes how people read on camera. A wide lens held close distorts features; step back and use a longer focal length for a flattering, compressed look. If you only have a phone, step back and zoom slightly rather than pressing the lens into someone's face.
Camera movement should have a reason. A slow push-in adds intensity, a pull-out reveals context, and a lateral tracking shot shows scale. Handheld energy suits testimonials and social clips; a tripod suits demos and instruction. Mixing both in one piece is fine as long as each choice matches the emotional register of the moment. Lock the shot when the viewer needs to read text or follow a click path.
Viewers forgive a soft image. They do not forgive bad sound, and they rarely forgive a face lit from beneath by a laptop screen.
Start with one key light positioned about 45 degrees to the side of the subject and slightly above eye level. Add a fill — a white bounce card or a second dimmer source — to lift the shadow side without flattening the face. Add separation by placing a small light behind the subject to outline the shoulder and hair. That is a complete interview setup, and a window with a diffusion sheet can substitute for the key light in a pinch.
Avoid mixing color temperatures. Daylight through a window plus tungsten ceiling bulbs creates green-orange patches that are painful to fix. Turn off the overheads, match your lights, and use practical lamps in the background purely for depth. Expose so skin tone sits slightly bright rather than clipped, and check the shot on a larger screen than the camera's rear display before you commit.
Place lavalier microphones 15 to 20 centimeters below the collarbone, tucked under fabric without rubbing. If you use a shotgun on a boom, keep it just outside the top of frame and angled down at the chest. Always monitor with wired headphones — a meter cannot tell you that a shirt is rustling or a fridge is humming.
Record 30 seconds of room tone with nobody speaking. That clip is what makes edits between takes seamless instead of jarring. Kill HVAC, fans, and notification sounds before rolling, and choose rooms with soft surfaces: rugs, curtains, and bookshelves absorb reflections that make small rooms sound like bathrooms.
Founders, customers, and teammates freeze in front of a lens. The fix is not more takes; it is better instructions. Give actions, not emotions. “Tell me how you'd explain this to a colleague at lunch” produces better delivery than “be more enthusiastic.”
Break scripts into short chunks so nobody memorizes a paragraph. Ask for the same line three times, then move on. Keep rolling a few seconds after each take — the relaxed, candid moment often arrives right after someone thinks they are finished, and that is the moment editors use.
Always shoot coverage. After the interview, spend fifteen minutes capturing hands on a keyboard, a screen scrolling, a product being unboxed, someone walking through the office, and a wide shot of the space. This b-roll is what makes a rough interview cut watchable, and shooting it during production costs far less than scheduling a second session.
Create folders for footage, audio, graphics, music, exports, and project files. Rename clips by scene, put an “A” in front of interviews and a “B” in front of b-roll, and duplicate the project file before large changes so you can always step back. On slower machines, generate proxy files and edit against those; the final export still uses the originals.
Start with a selects pass: pull the best ten seconds from each interview segment and ignore everything else. Assemble the story from those selects, then attack the first three seconds ruthlessly. If the opening frame does not raise a question or show something interesting, the rest of the edit is wasted effort.
Cut on action rather than on pauses, use J and L cuts so audio leads or trails the picture, and place b-roll every four to six seconds to cover jump cuts in speech. Remove filler words, false starts, and repeated points without mercy; viewers never miss a sentence that was cut.
Set music 18 to 22 decibels below the voice track and duck it further under narration. Add captions to everything — not only for accessibility, but because a large share of viewers watch with sound off. Deliver both burned-in captions and a subtitle file so the same video works on any surface.
Export a high-quality master first, then generate platform versions from it: 16:9, 1:1, and 9:16, each with its own caption placement and safe areas. Use H.264 at 1080p with a generous bitrate for social delivery, and keep variable frame rates out of the pipeline by locking the camera to a fixed frame rate. Name files with the campaign, version, and aspect ratio so nobody publishes the wrong cut.
You can produce convincing marketing video at three levels, and audio quality is the first thing to upgrade in every one.
| Tier | Kit | Best for |
|---|---|---|
| Minimal | Phone, lavalier mic, tripod, window light, free editor | Solo founders, social clips, test concepts |
| Mid | Mirrorless camera, prime lens, LED panel, boom or lav, editing suite | Customer stories, demos, evergreen brand content |
| Studio | Two cameras, three-point lighting, dedicated audio interface, teleprompter | Campaign hero videos, launches, training series |
AI tools are most useful at the edges of the workflow rather than at the center. Draft script variations and hook options to compare directions. Auto-transcribe interviews so you can search for quotes instead of scrubbing timelines. Generate caption files and translated subtitles in minutes rather than days. Clean up backgrounds and remove hum when a location was less than ideal. Use short generative clips as filler when a shot is missing and reshooting is impossible.
Two cautions apply. Disclose synthetic or altered footage when the context could mislead, and review every AI-generated line, caption, and clip before it reaches a customer. Language models write plausible sentences, not true ones, and a confident claim about your own product that nobody checked is a brand risk rather than a time saver.
Match length to the decision the viewer is making. Awareness clips work in 15 to 45 seconds, consideration content in 60 to 120 seconds, and onboarding or training can run several minutes. If a cut feels long, remove a beat rather than speeding up the narration.
No. A recent phone with manual exposure control, a tripod, and a decent microphone outperforms an expensive camera with weak audio for most marketing purposes. Upgrade lighting and sound before you upgrade the sensor.
Use a tripod, prefocus manually, and frame wider than you think you need so you can punch in during editing. Record several takes back to back without reviewing each one, and capture b-roll between takes so the session produces more than a talking head.
Choose based on where the video will earn most of its views, then frame for the other crops. Vertical favors faces and captions; horizontal favors demos, wide scenes, and anything with side-by-side comparison.
Three per line is a reliable default for non-actors. Beyond that, energy drops and the delivery gets stale. If a line fails three times, rewrite it shorter instead of repeating it.
Define one metric before production and check it after 72 hours and again after two weeks. Watch time and completion rate tell you whether the story held; clicks, replies, and signups tell you whether the call to action landed. When a video underperforms, retest the first three seconds before rebuilding the whole piece.

Alexander
Vidu Q4 alternatives compared: Kling 4.0, Seedance 2.5, Wan 3.0, and Runway Gen-4 — 4K support, input modes, durations, audio, and what each costs in credits on Domer.
The top 4K AI image generators on Domer compared: Flux 3 Klein, Nano Banana 2.1, GPT Image 2.5, GPT Image 2, and Nano Banana Pro — 4K credit costs, reference limits, and speed.
Nano Banana 2.1 vs GPT Image 2.5 Flare compared: credit costs per resolution, aspect ratios, reference images, text rendering, and which tasks fit each model.