My weekly workflow
Each week I save the top videos in the brand's niche from the TikTok Creative Center, the Instagram Reels tab and the Meta Ad Library into a swipe file. I note views, hook and length.
Why did it work? I note the hook line, the structure (POV, listicle, before/after), the pacing, the sound and the CTA. I copy the format, never the video itself.
I rebuild the format with your b-roll or AI footage. The script, product shots, voiceover and captions are matched to your products and offer.
Hook in 0–2 s, a cut every 2–4 s, word-by-word captions, product shown early, text inside the safe zone, then one CTA at the end.
A caption that opens with the hook, a reason to save or share, a CTA and a short hashtag set. A pinned comment carries the offer.
After 48–72 hours I check views, watch time, shares, saves and link clicks. Winners get new hooks, and weak videos get re-cut or dropped.
"POV: your 'gold' necklace turned your neck green"
- Brand
- Aurelle, a fictional everyday-jewelry store. Jewelry visuals were made with AI as code.
- Viral format
- The "POV" relatable-pain video: a problem the viewer has had, then the fix. These spread because people tag friends who've had the same problem.
- Structure
- Hook (0–2 s): dull, green-tinted chain + POV line.
- Switch: the fix in one line, plus the material.
- Proof: Gym ✓ Shower ✓ Beach ✓ stacked on the beat.
- Product: the set the viewer can buy.
- Offer + CTA: "15% off this week", Shop the set.
- 3 hooks to test
- POV: your "gold" necklace turned your neck green 🤢
- Jewelry I can shower, swim and sweat in
- Stop buying jewelry that turns green. Here's what to look for.
- Caption
- If your jewelry has ever turned your skin green, this one's for you 💛 The Everyday Set: shower-proof, sweat-proof, made to stay on. Save this for your next jewelry haul, or send it to the friend whose rings always turn green 👀 15% off the set this week. Tap the link in bio. #everydayjewelry #goldjewelry #waterproofjewelry #jewelrytok #giftideas
- Pinned comment
- "Code SET15 works on the whole Everyday Set until Sunday 💛"
- What I'd track
- 3-second views (is the hook working?), average watch time, shares (the "tag a friend" goal), link clicks and code uses.
"Make your home look expensive (no renovation)"
- Brand
- Terra & Tide Home, a fictional home decor store. With your b-roll, each tip would show your product in the room.
- Viral format
- The "X swaps that make your home look expensive" list. People save these to use later, and saves tell the algorithm the video is worth showing.
- Structure
- Hook (0–2 s): a big promise + "(no renovation)" to remove the effort.
- 4 numbered tips: one room per tip, numbered so viewers stay to the end.
- Voiceover + word captions: works with sound on or off.
- CTA: "Save this" plus "Shop the look".
- 3 hooks to test
- Make your home look expensive (no renovation)
- 4 swaps under $50 that make a rental look rich
- Why your living room looks "cheap" (and the easy fix)
- Caption
- 4 small swaps that make any room look expensive 🏡 No renovation, no landlord drama. Which one are you trying first: 1, 2, 3 or 4? Tell me below 👇 Save this for your next weekend reset, and shop every piece through the link in bio. #homedecor #homedecorideas #apartmentdecor #renterfriendly #cozyhome
- Pinned comment
- "Everything in this video is in our Shop the Look collection, link in bio 🛋️"
- What I'd track
- Saves and comments (the "1, 2, 3 or 4?" question), how many viewers reach tip 4, profile visits and link clicks.
Weekly test plan for each brand
| Day | What I'd do | Why |
|---|---|---|
| Mon | Find 10 new viral videos in the niche and pick the best 3 formats | Fresh formats every week |
| Tue–Wed | Re-create and edit 3–5 videos per brand, each with 2 hook versions | The hook decides most of the result, so I test it first |
| Wed–Sun | Post at the brand's best times, with performance captions and pinned comments | Steady posting gives the algorithm more chances to push a video |
| Every 48–72 h | Compare each video against the account's own average: views, watch time, shares, saves, clicks | Keep what beats the average and drop what doesn't |
| Fri | Short report: top video, why it worked and next week's plan | Next week's videos build on what won |
The numbers stay blank in a sample. With your accounts, the report would use your real data.