Short-form video is still the cheapest reach on the internet. A good 30 second clip can reach more new people in a weekend than a month of static posts, and the platforms keep rewarding it because it keeps people watching. The catch is that most brands treat video like a lottery ticket: post, hope, repeat. This is the playbook we use instead, refined across more than four thousand edits for brands in fashion, food, tech and beauty.
Start with the job, not the trend
Every video should have one job. Before anyone picks up a phone, we write a single sentence that finishes the line “After watching this, the viewer will…” Save it for later? Follow for part two? Visit the shop? Tag a friend? If you cannot finish that sentence, the video is not ready to film.
The job decides everything that follows: the hook, the length, the call to action and even the platform. Reach videos are built for the first two seconds. Trust videos earn their minute. Sales videos remove one specific doubt. Mixing all three in one clip is the fastest way to make a video that does none of them.
The first two seconds are the whole video
Viewers decide almost instantly whether to keep watching, and the algorithm is watching them decide. We script the opening harder than anything else in the edit. Three openings work again and again:
- The visual interrupt. Start mid-action: dough hitting the counter, a jacket being thrown on, a product dropped into frame. Motion in frame one beats a talking head saying hello.
- The specific promise. “Three ways to wear one linen trouser” beats “styling tips” every time. Numbers and nouns create a reason to stay.
- The open loop. Show the result first, then rewind to how you got there. People stay to close the loop.
Put the hook in the visual, the on-screen text and the first spoken line at the same time. Most people watch on mute for the first few seconds, so the text has to work alone.
If the first two seconds would not make you stop scrolling, nothing after them matters.
Film in batches, not on inspiration
Consistency beats brilliance on every platform we measure. The brands that grow fastest do not have better ideas, they have a calmer production rhythm. We film in batches: one half-day shoot produces two to three weeks of content, cut for every platform.
A batch day needs three things on paper before it starts: a shot list grouped by location and outfit, a list of hooks to film for each idea (always film more than one), and a list of B-roll moments that can be reused across edits. Hands, textures, products in use and reactions are worth their weight in gold in the edit.
Edit for retention, not for beauty
A beautiful edit that loses people at second four will lose to a scrappy one that holds them. Retention editing comes down to a few habits:
- Cut every pause and every breath that does not add tension.
- Change something on screen every two to three seconds: an angle, a crop, a caption, a cutaway.
- Keep captions in the safe zone, away from the platform buttons and the bottom caption area.
- End on a reason to rewatch or follow, not a slow fade to the logo.
Watch your own retention graphs. A sharp drop at second three means the hook overpromised or the payoff came too late. A slow slide in the middle usually means a section is not earning its place.
One idea, five platforms, five edits
Posting the identical file everywhere is a missed opportunity. TikTok rewards rawer, faster, talk-to-camera energy. Reels lean slightly more polished and aesthetic. YouTube Shorts favour clear payoffs and searchable titles. LinkedIn wants the business lesson spelled out. Pinterest lives much longer, so evergreen tips win there.
We keep one master edit per idea, then make platform versions: a new hook line, different text styling, an adjusted length and a platform-native caption. It takes an extra twenty minutes per video and routinely doubles total reach.
Measure what predicts growth
Views are a weather report, not a strategy. The numbers we watch weekly are the ones that predict next month’s growth:
- Three second hold rate, which tells you whether hooks are working.
- Average watch time as a share of length, which tells you whether the edit is earning attention.
- Shares and saves per thousand views, which tell you whether the content is worth something to people.
- Follows per thousand views, which tells you whether people want more of you, not just more of that video.
When one of these moves, we look for the pattern, not the outlier. A single viral video is luck. Five videos with strong saves on the same topic is a direction.
The playbook on one page
Give every video one job. Script the first two seconds harder than anything else. Film in calm batches with more hooks than you need. Edit for retention, version for every platform, and measure the numbers that predict growth rather than the ones that feel good. Do that for twelve weeks and you will not need a viral miracle. You will have a system that keeps producing small ones.
