The five months before coaching, next to the months since. One client’s account told month by month, dips included, then the same method run again from zero. I’d rather show you one real run in full than a highlight reel you can’t verify.
01The client account
Ski Hugger came in posting about once a week inside a tight niche within skiing: 23 reels in five months, 142,868 views all together, best video 49,597. Exactly one video had ever passed 25,000. Not failing, plateaued: the worst place to be, because everything looks almost right and nothing tells you what to fix.
The gap wasn’t effort. Nobody on the team could look at a finished video and name what was holding it back, so every post was a guess.
No framework PDF. We went straight into the loop: plan, film, send me the video, get exact notes back, revise, post. The loop had barely started when one video did 394,308 views, against a previous best of 49,597.
That’s the number everyone wants to lead with, and it’s the least useful number on this page. A spike proves the old ceiling was fake, and that is all it proves. The video sits at 394,308 today, a few hundred views added in almost five months: spikes stall. One video can always be luck. What can’t be luck is what the months after it did.
1,649,735
views in under six months of coaching
122 reels between mid March and early September, every one through the review loop before it posted. The five months before: 23 reels, 142,868 views.
Client account
Five months before
The coached months
Published grid only. 21 trial reels, 474,234 views, sit outside it.
May fell back to earth and June barely moved. This is the stretch a highlight reel would cut, and it’s where the actual coaching happened: output went from 1.1 videos a week to 4.3, every video got notes before it posted, and every miss got a reason on the weekly numbers call.
Then the reps paid. July ran 1.8x June. August did 446,760 views on the public grid, the biggest month the account has ever had, and it beat April without April’s crutch: April was 90.1% one video, August spread its total across 24 posts with the biggest at 28.5%. The hit rate did the work: before coaching, one video in 23 ever passed 25,000 views. Since coaching began, one in ten across all 122 reels. The audience followed the views: the account has more than quadrupled its followers since March, 4.3x.
The newest lever sits off the grid entirely. Instagram trial reels go only to non-followers and never land on the public grid, which makes them a free test bench: 21 of them added 474,234 views in about two weeks, averaging 22,583 against 11,639 for the 101 published reels. Five of the 21 cleared 25,000, one in four, against seven of 101 on the grid. That is the discovery engine, and the chart above counts none of them.
April437,665one video is 90.1% of it
May57,633the dip
June77,949
July140,152
August446,760biggest video is 28.5% of it
Views by calendar month, Instagram data for the client account, published grid only, pulled August 29. The 21 trial reels, 474,234 views read September 7, sit outside it. Coaching began mid March; April is the first full month.
An older video on their page had done 800 views. I sent back six specific changes: cuts, a couple of re-records, a different opening line. They made the changes and reposted the same idea. It did 13,000. In their niche, that’s a monster number.
Nothing about the idea changed. The execution did. That’s the whole method in one anecdote, and the account’s climb since March is that same pass, run 122 times.
800 → 13,000
on one revision pass
02From zero
In February 2026 I started a brand new account from zero. Same loop I coach: film, review, revise, post, repeat. Ten weeks later it had passed 2M views.
Starting from zero was deliberate. No audience, no leg up, no excuses. If the method only worked for people who were already big, I wanted to know before I charged anyone for it.
2,219,908
views between Feb 1 and May 1, straight off the receipt below
97.9% non-follower views
30 minutes. I’ll look at your content and tell you what’s keeping it stuck.
2 of 4 client slots open
