BEGINNER: Here are my Instagram screenshots and the numbers for each reel, along with my on-screen hook so we know which reel is which.

EXPERIENCED: You have Windsor connected. Pull my Instagram data for the last 30 days. I want every reel with these fields: views, reach, likes, comments, saves, shares, average watch time, follower vs non-follower reach split, and skip rate and retention if available. If skip rate and retention aren't in Windsor yet, tell me to paste those from screenshots separately.

You are analyzing 30 days of Instagram reel data for my account. Your job is to find the patterns across ALL of my reels, tell me what's actually working, what's dying, and why — based on the actual numbers, not generic social media advice.

For my bottom 10 lowest performing reels, I'll give you the metrics Windsor pulled, plus a screenshot description of my skip rate and retention from Instagram Insights since those aren't in Windsor yet.

Once you have the data, here's what I need you to analyze:

WHICH REELS ARE ACTUALLY WORKING: Not just views. Look at saves, shares, and watch time together. A reel with 500 views and 40 saves is performing better than a reel with 2000 views and 3 saves. Tell me which ones are genuinely landing and what they have in common — topic, length, opening style, anything consistent.

WHICH REELS ARE DYING AND WHY: Break it down by what metric is failing. Low reach with low views means weak early signals. Decent reach but low views means the hook didn't stop the scroll. Good views but bad watch time means the opening worked but the middle lost them. Tell me specifically which problem each underperformer has.

VIEWS AND REACH: Don't just tell me views are low. Tell me if it's a distribution problem or a hook problem. If reach is decent but views are low, the hook didn't stop the scroll. If reach is also low, the early engagement signals were weak.

SAVES ANALYSIS: Saves are the most important signal for a content creator. If saves are under 1% of views consistently, my content is being watched but not felt deeply enough to keep. Flag this pattern if you see it and tell me what type of content is getting saved vs what isn't.

SHARES ANALYSIS: Shares happen when someone feels so seen they have to send it to a friend. Low shares across the board usually means content is too general or too polished to feel personal. Tell me if this is a pattern and which reels bucked it.

WATCH TIME PATTERNS: Look at average watch time relative to reel length. Under 5 seconds on a 45 second reel means they left before I even spoke — hook failure. 15 to 20 seconds and then gone usually means the opening worked and then something in the middle lost them. Tell me where people are leaving and what that points to.

SKIP RATE: If I've provided skip rate data, flag every reel where it's high. High skip rate means the on-screen hook or opening visual didn't stop the scroll. This is almost always a hook framing problem.

RETENTION: I'll describe the retention curve from my screenshot. A steep drop in the first 3 seconds means hook failure. A drop at a specific moment in the middle means something in the script lost them at that exact point. Tell me where it dropped and what that likely means.

FOLLOWER VS NON-FOLLOWER REACH: If most views are coming from existing followers consistently, Instagram isn't distributing my content to cold audiences. Tell me which reels broke out to non-followers and what they had in common.

CONTENT CATEGORY PATTERNS: Group my reels by topic or style based on the hooks I've given you. Which category is performing best? Which is consistently flat?

Go reel by reel first. For each one tell me: what the numbers are actually saying, which single metric is the main problem, and one sentence on what that points to. Then give me a pattern summary across all reels. What's the same mistake showing up over and over.

Be specific to MY numbers. No generic social media advice. Every observation must come directly from what I give you.

At the end give me 3 things: my single biggest recurring problem across all 30 days in one sentence, my top 2 performing reels and what they have in common, and the 3 to 5 lowest performing yapping reels by combined metrics — because those are the ones I'm taking into Step 2 for a full script autopsy.

You are doing a full script autopsy on my lowest performing yapping reels. You already have my 30-day pattern summary from Step 1. Now you're going inside the actual content of the specific reels that flopped to find exactly where and why each one lost people.

Analyze each reel using these standards:

ON-SCREEN HOOK: Does it name a specific behavior she's doing right now — something she physically does or avoids on a regular Tuesday when she's frustrated — or does it describe a topic, a lesson, or something about the creator? Behavior hooks stop scrolls. Topic hooks get ignored. Check: is it under 15-20 words? Does it sound like a thought that escaped out loud, or does it sound like a webinar title? Does it make her feel something about herself in under 2 seconds?

HOOK TO OPENING LINE MATCH: The first spoken sentence must escalate what the hook promised — not restate it, not set it up, not provide context. If the hook promised a confession, the first spoken words must be the confession already happening. Flag every reel where the opening line repeats the hook, explains background, or uses any preamble like "so I wanted to talk about," "what I'm about to share," or starts from the creator's situation instead of the viewer's feeling.

HOW MANY IDEAS: Count separate ideas in the reel. If there's more than 1, flag it hard. One reel, one idea. Every time a second concept appears, you're asking her to reset her attention and she doesn't — she leaves.

LANGUAGE CHECK: Flag abstract language like "step into your power," "own your authority," "amplify your message," or anything that sounds like a coach delivering a lesson. Also flag AI educator language — "tokens," "prompts," "systems architecture" — anything that requires 2 cognitive steps before she even understands what she's watching. The standard is: would she say this exact phrase to her business bestie when she's ranting? If no, flag it.

WHERE THE EMOTIONAL PAYOFF LANDS: Where is the first moment a viewer would think "oh sh*t that's literally me"? If that moment is after 15 seconds, flag it. She needs to feel something in the first 8 seconds or she's already checking out. Find the most specific, most uncomfortable, most alive line in the whole script and tell me exactly where it sits. If it's buried in the middle, that line should have opened the reel.

WHERE IT PIVOTS TO TEACHING: Find the exact moment the script shifts from lived experience or real observation into explanation, lesson, framework, or offer description. This is the most common reason yapping reels lose viewers mid-watch. Flag the exact sentence where this happens.

SCENES VS CONCEPTS: Are there concrete moments and specific details? "I was sitting at my desk at 11pm staring at 30 views spiraling" is a scene. "Many creators struggle with content performance" is a concept. Flag every concept and tell me what scene could replace it.

For each reel give me: hook rating out of 10, the exact sentence where the reel loses the viewer, the single biggest structural problem in plain language, then one specific rewrite suggestion for the hook and the opening line only — not the whole script.

End with a pattern summary: what's the same mistake showing up across all of my scripts that I need to fix before I record another reel.

Here are my scripts:

REEL 1
On-screen hook text (or screenshot):
First spoken sentence:
Full transcript:
Skip rate if you have it:
Retention notes (from IG screenshots):

[REPEAT FOR EACH REEL]