/funnel-diagnosis

Stage-by-stage funnel comparison for any A/B test. Find exactly where the variant diverges from control: Add to Cart → Begin Checkout → Contact Info → Address → Purchase.

Works with both active and ended tests.

Step 0: Workspace Check

Check if the shared workspace exists and is ready:

test -d ~/intelligems-analytics/venv && test -f ~/intelligems-analytics/ig_client.py && echo "READY" || echo "NEEDS_SETUP"

If NEEDS_SETUP: Run the /intelligems-core skill first to set up the workspace.

Step 1: Get API Key

Check if an API key is already configured:

source ~/intelligems-analytics/venv/bin/activate && python3 -c "
from dotenv import load_dotenv; import os
load_dotenv(os.path.expanduser('~/intelligems-analytics/.env'))
key = os.getenv('INTELLIGEMS_API_KEY', '')
print('HAS_KEY' if key and key != 'your_api_key_here' else 'NO_KEY')
"

If NO_KEY: Ask the user:

"What's your Intelligems API key? You can get one by contacting support@intelligems.io"

Then save it:

echo "INTELLIGEMS_API_KEY=<user's key>" > ~/intelligems-analytics/.env

Never hardcode or assume an API key.

Step 2: Copy Funnel Script

Copy the funnel diagnosis script from this skill's references into the workspace:

cp references/funnel.py ~/intelligems-analytics/funnel.py

Step 3: Select Test

The script accepts an optional test ID as an argument.

If the user already specified a test ID or name:

• Pass the ID directly: python3 funnel.py <test_id>

If no test was specified:

• Run the script without arguments — it will list all active experiments and ask the user to pick one.

Step 4: Run Analysis

Execute the funnel script with the workspace virtual environment:

cd ~/intelligems-analytics && source venv/bin/activate && python3 funnel.py [optional_test_id]

The script will:

1. Fetch overview analytics for the selected test
2. Extract metrics for each funnel stage (Add to Cart → Purchase)
3. Compare control vs. variant at each stage
4. Identify the biggest gain, biggest drop, and breakpoint
5. Print a stage-by-stage comparison with plain-English diagnosis

Step 5: Present Diagnosis

Read the script output and present results conversationally. Structure your response like this:

1. Quick Summary

One sentence: where the variant wins and where it struggles.

"The variant lifts Add to Cart by +8.2%, but loses ground at checkout — conversion only up +1.3%."

2. Stage-by-Stage Table

Present the funnel stages as a clear comparison:

3. Key Findings

Highlight the most important findings:

• Biggest gain — Which stage improved most
• Biggest drop — Which stage lost most (if any)
• Breakpoint — Where behavior diverges (e.g., "More people add to cart, but they drop off at checkout")

4. Diagnosis

Plain-English explanation of what the funnel data means:

"The variant is great at getting people interested (add to cart), but the checkout experience needs work. Consider testing checkout-specific changes next."

5. Next Steps

Specific suggestions based on the funnel findings.

Step 6: Set Up Slack Automation (Optional)

If the user wants funnel diagnoses delivered to Slack:

1. Get a Slack webhook URL

"To send results to Slack, you need an Incoming Webhook URL. Go to https://api.slack.com/apps → Create New App → Incoming Webhooks → Add New Webhook."

2. Copy the Slack module

cp ../intelligems-core/references/ig_slack.py ~/intelligems-analytics/ig_slack.py

3. Test it

cd ~/intelligems-analytics && source venv/bin/activate && python3 funnel.py <test_id> --slack "<webhook_url>"

Notes

• Funnel stages come from overview analytics — no extra API calls needed beyond the standard overview.
• Missing stages: Some tests may not have data for all 5 funnel stages. The script handles this gracefully by showing "No data" for missing stages.
• COGS awareness: When COGS data exists, revenue metrics use Gross Profit per Visitor instead of Revenue per Visitor.
• Slack output: Uses Slack Block Kit formatting with a clean stage comparison.