yt Retention

    1

    Read a YouTube Studio audience-retention CSV export and find where viewers actually leave (hook leak, cliffs, slide), then name the single biggest leak and one change.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    yt-retention

    Example session with this skill installed

    Analyze retention.csv for a 12-minute video. Use transcript.srt to find the biggest leak and suggest one fix.

    • Read your context and instructions
    • Compiled the yt-retention

    The retention analysis for your 720-second video is complete. I have mapped the CSV data against your transcript to identify the primary cause of viewer attrition.

    Retention Leak Audit: "Why I Switched to Neovim"

    MetricValueStatus
    Hook Retention (30s)68%LEAK DETECTED (Target >75%)
    Primary Leak TypeCliffOccurred at 04:12
    Drop Magnitude-1.2%Significant
    Context"Before we get into the config, let's talk about my desk setup..."Topic Shift

    Biggest Leak: The "Desk Setup" Cliff
    The data shows a sharp 1.2% drop at the 04:12 mark. This coincides exactly with the unsignposted transition from technical Neovim benefits to your physical desk setup. This is a classic "setup cliff" where viewers who clicked for software advice felt the value proposition changed without warning.

    Suggested Change
    Cut the 45-second desk setup segment entirely or move it to the end of the video. If the segment must stay, add a verbal signpost at 04:10: "I'll show you the software in a second, but the hardware is why this works."

    Next steps

    • Apply the suggested cut to your master project for future re-uploads or shorts.
    • Check if your next script has a "bridge" sentence at every major topic transition.
    • Run the title/thumbnail check if your hook retention remains below 70% after script changes.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Analyzing YouTube retention charts manually is tedious and often leads to vague conclusions about "pacing." This skill processes YouTube Studio audience-retention CSV exports to identify exactly where viewers drop off and provides specific script or editing fixes based on the data.

    What it does

    • Identify Hook Leaks checks the first 30 seconds for excessive viewer loss and flags script issues.
    • Locate Retention Cliffs finds specific moments with drops over 0.8% caused by sudden topic shifts or long setups.
    • Analyze Pacing Slides detects steady viewer bleeds across the middle of the video that require tighter editing.
    • Transcript Mapping syncs retention data with .srt files to show exactly what was said during each drop.
    • Actionable Prioritization isolates the single biggest leak and provides one concrete fix instead of a generic list.

    How it works

    1. Export Data from YouTube Studio Analytics (Engagement > Audience Retention > Download CSV).
    2. Execute Script using retention.py with the CSV file and the video duration in seconds.
    3. Optional Context can be added by passing a --transcript file to map data points to specific dialogue.
    4. Review Report to see the primary leak type and the suggested corrective action.

    Frameworks & tools

    Python 3 script. Requires standard YouTube Studio CSV exports and optional .srt transcript files.

    Why this beats prompting it yourself

    Generic LLM prompts cannot accurately parse raw CSV time-series data or apply specific retention heuristics like the 0.8% cliff threshold. This skill uses a dedicated Python parser to ensure data integrity before providing a strategic analysis.

    Use cases

    • Post-Mortem Analysis of a video that underperformed despite high click-through rates.
    • Script Optimization by identifying which sections of a series consistently cause viewers to leave.
    • Editing Refinement to determine if a video needs content cuts or better signposting.

    Known limitations

    Requires at least 8 data points in the CSV to function. Thresholds for "leaks" are based on specific heuristics, not official YouTube platform rules.

    Requires CSV with at least 8 data points. Heuristics like the 0.8% cliff are specific benchmarks, not YouTube API constants. No direct account access; requires manual CSV export from Studio.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 14 days ago

    • Free with an Agensi account

    Requires

    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Listed14 days ago

    What's inside

    Frequently Asked Questions

    This skill automates the analysis of YouTube Studio CSV exports, mapping specific retention drops to your video's timeline or transcript to identify exactly which script lines or editing choices caused viewers to leave.

    You will need a Python 3 environment to run the analysis script, along with standard audience retention CSV files exported directly from your YouTube Studio dashboard.

    Your purchase includes the `retention.py` script, comprehensive documentation on how to export the correct data, and a guide for interpreting the prioritized leak reports.

    Yes, if you provide an optional .srt transcript file, the skill will sync the retention data to your dialogue, showing you precisely what was said at the exact moment a 'retention cliff' occurred.

    The skill requires at least 8 data points within the CSV to generate a report, and while it works for all video lengths, it is most effective for long-form content where pacing slides are more common.

    New in Social Media