Decoded by AI · Instagram · Updated October 2, 2026

Kallaway's best Reels, decoded.

Kallaway (@kallawaymarketing) is one of the most studied short-form educators on Instagram. We pulled his latest 100 Reels, transcribed every one, and had Jev label each script on eight dimensions: hook, structure, opening, evidence, emotion, advice, topic and call to action. Then we compared the labels against his plays. Here is what the data says, and what it doesn't.

Reels collected

100

Scripts decoded

Median plays

29.9K

Top Reel

1.1M

His formula

What his top 20% have in common

Hook

Result

Structure

Steps

Opening

Claim

Evidence

Personal

CTA

Comment

The most common label in each category among his best-performing fifth of Reels, by plays.

The numbers

Median plays by hook, evidence and structure

Compared with his account median of 29.9K plays.

Hook

Mistake4 Reels · small sample34.7K+16%

Result52 Reels32.4K+8%

Curiosity15 Reels31.7K+6%

Direct7 Reels24.9K-17%

Contradiction5 Reels23.7K-21%

Recognition3 Reels · small sample20.2K-32%

Evidence

Personal24 Reels45.4K+52%

Numbers3 Reels · small sample36.2K+21%

Example19 Reels27.9K-7%

Reasoning40 Reels24.4K-19%

Structure

Comparison4 Reels · small sample34K+14%

Steps42 Reels29.2K-2%

Explanation28 Reels28.2K-6%

Problem → solution12 Reels25.6K-14%

5 findings

What actually moves his views

01
He runs one format on repeat
60% Result hooks · 84 of 86 end with a comment CTA
Most of his Reels promise a concrete outcome in the first line ("here's how to..."), teach it in steps or a short explanation, and close by asking viewers to comment a keyword. 52 of 86 scripts open with a Result hook and 72 open with a plain claim. The consistency is the strategy: viewers know what they'll get, and the comment CTA turns every Reel into a lead capture.
Limits: Because almost every Reel uses the comment CTA, this data can't tell whether it helps or hurts reach.
Try this: Pick one hook type and one CTA and run them for 20 posts before you judge them.
02
Personal proof is his strongest lever
45.4K vs 24.4K median plays
Reels where he backs the claim with his own experience (24 scripts) reach a median of 45.4K plays, about 1.9x the Reels that only argue the point with reasoning (40 scripts, 24.4K). His account median is 29.9K. Among label groups with at least five Reels, the type of evidence separated his views most clearly.
Limits: Personal stories may also cluster around certain topics, so treat this as a strong pattern to test, not a law.
Try this: Add one line of "I did this, here's what happened" to your next teaching Reel.

03
AI-tool Reels climb to the top, and he remakes the winners
Claude in 32% of top-quarter openings vs 5% of the bottom quarter
Openings that mention Claude sit far more often in his best-performing quarter than his weakest. He also re-shoots what works: variations of "turn Claude into a social media machine" appear four times in his top 12 Reels, at 822K, 342K, 127K and 92K plays.
Limits: Topic interest shifts fast. This reflects the moment these Reels were posted.
Try this: When a Reel wins, remake it with a new first line within two weeks instead of moving on.
04
Structure matters less than you'd think
Steps 29.2K vs Explanation 28.2K median plays
Numbered steps (42 scripts) and straight explanations (28 scripts) perform almost identically for him. Problem-to-solution scripts trail a little at 25.6K. The first sentence and the proof behind it move his numbers more than whether the body is a list.
Limits: Comparison scripts look strongest (34K) but there are only 4 of them, too few to trust.
Try this: Spend your editing time on the first sentence, not on reformatting the middle.
05
Naming the technique isn't the magic
"It's called..." in 41% of his top quarter and 52% of his bottom
He often names his method: the addiction loop, the desire hook, comprehension maxing. The habit is just as common in his weakest Reels as in his best, so a catchy name doesn't carry a Reel on its own. What changes between his best and worst is the outcome promised and how it's proven.
Limits: Names still help memorability and sharing, which this data doesn't measure.
Try this: Keep naming your frameworks, but judge hooks on the outcome they promise.

Top 12 Reels

His best-performing openings

#11.1M plays44.9 likes/1K

“Here's how to make your storytelling so addictive, it is scientifically impossible to skip. This is called the dopamine addiction loop.”

Result hookStepsWatch

#2821.6K plays35 likes/1K

“This is how you turn Claude into a social media machine. Most people don't know this, but you can just type /analyze…”

Result hookExplanationWatch

#3455.2K plays32.9 likes/1K

“This is the ultimate script writing format for short-form video. First, you start with the hook, then you go into the lock-in…”

Direct hookStepsWatch

#4342K plays28.7 likes/1K

“This is how you turn Claude into a social media machine. Most people don't know this, but you can type forward slash…”

Result hookExplanationWatch

#5305.4K plays30.2 likes/1K

“This is the best database in the world for short form video references. There are over 100 different collections of videos across…”

Direct hookExplanationWatch

#6295.3K plays46.5 likes/1K

“There's a new type of content that is starting to crush on social media. It's called application content.”

Curiosity hookComparisonWatch

#7256.2K plays35.4 likes/1K

“Here's how to write a hook so good it is basically unskippable. This is called the desire hook method.”

Result hookStepsWatch

#8165.8K plays33.6 likes/1K

“These are the only six words you need to make your hooks impossible to skip. They're called the six hook power words,…”

Curiosity hookStepsWatch

#9148.2K plays54.4 likes/1K

“This is how to get people addicted to your storytelling, even if you talk about super boring topics. It's called the illusion…”

Result hookStepsWatch

#10136.9K plays47.2 likes/1K

“This is the best storytelling hack for educational content. It's called comprehension maxing.”

Result hookStepsWatch

#11127.3K plays29.6 likes/1K

“This is how to turn Claude into a social media machine. Now most people don't know this, but you can just type…”

Result hookExplanationWatch

#1292.2K plays28.1 likes/1K

“This is how to turn Claude into a viral content machine. Most people don't know this, but you can actually have Claude…”

Result hookProblem → solutionWatch

Script anatomy

Inside his top Reel: 1.1M plays

A promise of storytelling that is "scientifically impossible to skip", a named method, a second hook tied to casino psychology, then four numbered steps (stakes, the big question, the head fake, the re-hook), each with a quick example, and a comment CTA to close. Jev labeled it a Result hook with a Steps structure and sequence-level advice.

Sentence roles, in order

HookSetupAdviceExamplePayoffCta

1 Promise an outcome
2 Name the method
3 Second hook
4 Step 1 + example
5 Step 2 + example
6 Step 3 + example
7 Step 4
8 Comment CTA

Steal the structure

4 script templates from his patterns

Outcome + method

Here's how to [specific outcome] even if [common obstacle]. It's called [method name]. Step one: [action]. Step two: [action]. Step three: [action]. Comment [WORD] and I'll send you [resource].

The tool machine

This is how you turn [tool] into a [result] machine. Most people only use it for [basic use]. Instead, [specific feature or prompt]. Here's what it gave me: [result]. Comment [WORD] for the setup.

Proof first

I [did a specific thing] for [time period] and got [result]. Here's the exact [number]-step process. One: [step]. Two: [step]. Three: [step]. The step most people skip is [step].

Remake the winner

[The promise from your best Reel], but faster. Last time I showed you [old method]. Here's the [number]-minute version that gets the same result.

Written by us from the patterns above. Swap in your own outcome, proof and keyword.

The full dataset

Get all 86 decoded Reels as a spreadsheet

Every Reel with plays, likes, its hook, structure, opening, evidence, emotion and CTA labels, and the opening line.

Method and limits

FAQ

How was this analysis made?

We collected his latest 100 public Instagram Reels, transcribed the 86 that contain speech, and had Jev (a typed AI judgment model) label each transcript on eight dimensions. Jev only reads the words, never the view counts. We then compared median plays across the labels.

Is this affiliated with Kallaway?

No. This is an independent analysis of publicly available Reels. We only quote short opening lines and link to the original posts.

Do these patterns cause more views?

Not necessarily. These are medians within one account, with no significance testing. They are patterns worth testing on your own content, not proof of what causes reach.

Independent analysis of public Reels by AI Topia, not affiliated with Kallaway. Data as of October 2, 2026: 100 Reels collected, 86 with speech decoded, 14 without speech skipped. Plays and likes change over time.

Want a content engine built on what works?

We turn findings like these into a weekly video system for founders.

Book a call

Unlock the full Kallaway breakdown

All findings, top Reels, script templates and the spreadsheet. Free, opens instantly.

Kallaway's best Reels, decoded.

What his top 20% have in common

Median plays by hook, evidence and structure

What actually moves his views

He runs one format on repeat

Personal proof is his strongest lever

His best-performing openings

Get all 86 decoded Reels as a spreadsheet

FAQ

Want a content engine built on what works?