Competitor intelligence
Every account you compete with is running experiments in public. Clipmine watches their feeds, flags the videos that beat their own baseline, and hands you the winners — with the pattern explained.
Watch the baseline. Catch the breakout.
Illustrative tiles. Multipliers are account-relative — a video’s views against that account’s own recent median, not a global leaderboard.
Watchlists
Pick the accounts worth watching.
Group the competitors, adjacent brands, and creators you learn from into watchlists. Clipmine keeps an eye on what they publish so you don’t have to scroll their feeds every morning.
- Organize accounts by competitor, niche, or campaign theme
- Tracked-account limits scale with your plan — the backend decides, not the client
- No follows, no DMs, no engagement — observation only
Watchlist · Skincare rivals
- @lumi.skincare.demo142 videos tracked
- @glowtrack.demo89 videos tracked
- @routine.studio.demo61 videos tracked
- @calm.coach.demo47 videos tracked
Outlier detection
A breakout, not a big account.
A million-view account doing a million views isn’t news. Clipmine measures each video against the account’s own recent median, so a clip running 8.4× above its baseline stands out — even on a small account. That’s the signal that the hook, not the follower count, did the work.
- Relative multiplier vs. the account’s recent median views
- Small accounts surface real winners — no algorithmic bias toward the big names
- The mechanism gets explained, so you adapt the pattern instead of copying the post
Math is honest: 102K ÷ 12.1K ≈ 8.4. The multiplier is just the ratio — no black box.
Confidence levels
We tell you when we’re not sure.
A multiplier means little without enough history behind it. Every outlier carries a confidence label so you know how much weight to put on it.
The account has a stable posting history and a clear median. A big multiplier here is a real breakout worth adapting now.
Enough data to flag the pattern, but the baseline is still moving. Treat it as a strong lead, not a sure thing.
New account, sparse history, or a metric we couldn’t verify. We surface it, but we’d rather under-claim than oversell.
Weekly digest
One email. The week’s breakouts.
Instead of checking every account daily, get a digest that ranks the week’s outliers across your watchlists by relative multiplier and confidence. Skim it, pick the ones worth your time, and send them straight to a brief.
- Ranked by account-relative multiplier and confidence
- Each entry links to the analysis and a one-click “adapt” action
- Cadence and recipients are up to you
- 8.4×
The confession hook
@lumi.skincare.demo
High - 6.7×
One-ingredient teardown
@glowtrack.demo
High - 5.1×
Day-1 vs day-30 split
@habitloop.app.demo
Medium - 4.2×
Whispered “don’t do this”
@calm.coach.demo
Medium
Adapt the winners
From their breakout to your brief.
Spotting the outlier is half the job. The other half is doing something with it. Send any flagged video into the same analyze-and-adapt loop as the rest of Clipmine: keep the mechanism that made it work, change everything that makes it yours.
- Reuse the hook mechanism, structure, and pacing
- Swap the product, audience, examples, and CTA for your brand
- Get five hooks, a script, a shot list, and cut-downs — production-ready
Keep
- Hook mechanism
- Emotional sequence
- Structure
- Pacing
Change
- Product
- Audience
- Examples
- CTA
Honest about the setup
Live indexing depends on a data provider.
We won’t pretend this runs on magic. Continuously indexing competitor feeds means pulling public video data through a connected provider, and that requires setup and a provider account before live watchlists turn on. The tiles and digests on this page show the concept; what your account sees depends on the sources you connect and what those platforms make available. We’d rather be clear about that than imply real-time coverage we can’t promise on day one.
- Observation only — public video performance, no private data
- Coverage and refresh cadence follow your connected provider
- Where data is thin, we label it “Limited data” instead of guessing