The hook formula your top 3 competitors are using right now (and how Regent steals it for you)
Every creator I know has burned an evening on this: opening competitor profiles, scrolling their last 30 reels, taking notes on what's working. By the time you've watched 90 reels across 3 accounts, you have 4 pages of notes and zero scripts.
That's the gap Regent fills. The agent does the scrolling. Daily. And writes your week's scripts using what it found.
Here's what the pattern looks like when you actually run it.
#What Regent looks for
Three competitors. That's it. You drop their handles into Regent Insight, and the agent watches their public posts every day, extracting six dimensions:
- Hook timing — at what second does the actual claim land?
- Face cadence — when does the creator's face first appear on camera?
- Caption structure — does the caption lead with the hook or with the CTA?
- Topic clusters — what's the 3–5 recurring themes their best posts share?
- Peak posting windows — which day-and-hour combinations consistently outperform?
- Length distribution — how long are their best posts vs their worst?
Then Regent compares the patterns across all three. The patterns that appear in at least two creators' winning posts are flagged as the "formula" for your niche.
#The pattern that shows up across niches
Across fitness creators, finance creators, indie builders, copywriters, real estate creators — the same skeleton keeps emerging from the data. Different surface, same bones.
Hook delivered before 2 seconds. Not 3. Not "after the intro slide." The literal claim, in the first beat. Examples that appear across niches:
- "Most people don't realize..."
- "The thing nobody tells you..."
- "Stop doing X. Here's why."
- A specific number ("I made $12k last month doing this")
- A direct contradiction of conventional wisdom
The reels that try to hook with "Hey guys, today I'm going to talk about..." consistently underperform. Algorithm reads the high skip-rate at second 2 and chokes the distribution.
Face on camera by 3 seconds. No long intro slides. No animated logo reveal. The viewer needs to see who's talking, fast. Otherwise the post reads as a generic page reposting clips. Algorithm signal: low trust → low distribution.
One idea per reel. This is the one most creators ignore. The best-performing posts cover exactly one thing. Not "three reasons why" — just one reason, delivered with weight. Multi-point reels split attention, lose the algorithm's per-second engagement scoring, drop off at the second point, never recover.
Caption opens with the hook, not the CTA. Algorithm reads the first 30 characters as the preview. Putting "Follow for more!" first wastes the slot. Put the punchline first, save "follow" for the last line. Caption-to-engagement rate doubles, sometimes triples.
Question close. The reels that go viral consistently end with a question. Not always literal — sometimes implied. But the close opens a loop the viewer wants to comment on. Comments are the strongest engagement signal a platform sees. Open the loop.
Posted in the 8–10 PM local window. Not universal, but across most niches the post that lands in early evening hits the algorithm's discovery boost cleanest. There are exceptions (B2B creators win at lunch, finance creators win on Sunday evening). Regent learns yours specifically from your competitors' best windows.
#Why the formula is consistent
A reasonable question: why does the same skeleton work across fitness, finance, real estate, and indie hackers?
Because the algorithm doesn't care about your niche. It cares about the per-second engagement curve. The hook-by-2-seconds pattern works because that's when the algorithm starts deciding whether to show the post to more people. The face-by-3-seconds pattern works because viewers need a person to trust within the first beat. The question-close pattern works because comments are the metric that triggers further distribution.
You're not optimizing for the niche. You're optimizing for the algorithm. The algorithm's behavior is the same across niches, so the formula is the same.
What changes between niches is the content — what you talk about, who you reference, what stakes you raise. That's still yours. Regent doesn't write your point of view. It writes the shape that delivers your point of view to the algorithm.
#How Regent applies the formula to your reels
Once Regent has watched your three competitors for a week, the next time you open the calendar, you see this:
- Topic slots auto-generated for the week ahead (one topic per day, sometimes two)
- Each topic with three hook variants that match the patterns Regent found
- Each variant tagged with the competitor it borrows the shape from ("Inspired by @hormozi's contradiction-first hook")
- Suggested posting time (per platform, based on your competitors' peak windows)
- One-click regenerate if a topic or hook misses
You read the suggestions. You approve. You don't write scripts. You don't watch reels. You don't take notes.
The agent renders each reel in your face and voice, applies the captions in your tone, schedules them at the peak window, and ships.
#What the formula doesn't do
To be honest: pattern-matching is not strategy. Regent finds what's already working. If your three competitors are all stuck in a topic that's about to die — finance creators all hyping a coin that's about to crash — Regent will faithfully reproduce a dying formula.
The human call is picking the right three competitors. Not the biggest. Not the most viral. The ones whose audience overlaps with yours, whose stage you want to be at in 12 months, whose tone you'd actually be proud to share.
Pick them well, and Regent does the rest.
Try the formula on your own competitors. Apply for the Creator Beta — drop three handles, see what Regent extracts, ship reels using the patterns within the hour.
Read next:
