Outbound

Video Engagement Follow-Up

Automate follow-up sequences based on Loom video view analytics and engagement signals

Loomn8nAttioInstantlyPostHog
$npx gtm-skills add drill/video-engagement-follow-up

What this drill teaches

Video Engagement Follow-Up

This drill builds automated follow-up workflows that react to Loom video engagement data. Instead of following a fixed time-based email cadence, this drill routes prospects into different follow-up paths based on whether they watched the video, how much they watched, and whether they clicked the CTA.

Input

  • Active video outreach campaign (from video-prospecting-outreach drill)
  • Loom Business account with viewer analytics
  • n8n instance connected to Attio and Instantly
  • PostHog tracking configured

Steps

1. Set up Loom view data ingestion

Loom does not provide webhook-based real-time notifications via API. Build an n8n polling workflow using n8n-workflow-basics:

  1. Create a cron-triggered workflow that runs every 4 hours during business days
  2. For each active campaign video, query the Loom dashboard (scrape or use the SDK viewer data endpoint if available) to get:
    • List of viewers (email if identified, anonymous if not)
    • Watch percentage per viewer
    • CTA click events
  3. Compare against the last poll to identify NEW view events
  4. For each new view event, fire a PostHog event using posthog-custom-events:
    • Event: video_viewed
    • Properties: prospect_email, video_id, watch_percentage, cta_clicked, campaign_slug

2. Build the engagement-based routing workflow

Create an n8n workflow triggered by the video_viewed PostHog event (via webhook):

Branch A -- High engagement (watched >75%):

  1. Using attio-contacts, update the contact: video_engagement: high, video_watch_pct: {pct}
  2. Using attio-deals, create a deal or move existing deal to "Interested" stage
  3. If they clicked the CTA: do nothing extra, the Cal.com booking handles it
  4. If they did NOT click the CTA: pause the Instantly sequence for this prospect using instantly-reply-detection (treat as manual takeover), then generate a personalized follow-up email:
    Subject: Saw you watched my video, {first_name}
    
    Hey {first_name}, looks like you caught the video I sent.
    
    Want me to walk you through {specific_feature} live? Here are a few times: {cal_link}
    
    {your_name}
    
  5. Send this follow-up via the founder's personal email (not Instantly) for authenticity

Branch B -- Medium engagement (watched 25-75%):

  1. Update Attio: video_engagement: medium
  2. Accelerate the Instantly sequence: move Email 2 forward to send within 24 hours instead of the default 3-day delay
  3. Log in PostHog: video_followup_accelerated event

Branch C -- Low engagement (watched <25%):

  1. Update Attio: video_engagement: low
  2. Let the standard Instantly sequence continue on schedule
  3. After the sequence completes with no reply, add to a re-engagement list for a different angle in 30 days

3. Handle "never watched" prospects

Create a separate n8n workflow using n8n-scheduling:

  1. Run daily, 5 days after the initial video email was sent
  2. Query prospects where video_sent: true AND video_watched: false AND no email reply
  3. For these prospects, the video approach did not work. Send a text-only follow-up with a different angle (no video reference):
    Subject: {pain_point_headline}
    
    Hey {first_name},
    
    {two_sentence_value_prop_without_video_reference}
    
    Worth a quick chat? {cal_link}
    
    {your_name}
    
  4. Update Attio: video_approach: no_engagement

4. Track follow-up effectiveness

Using posthog-custom-events, track which follow-up branch produces the most meetings:

  • video_followup_sent -- properties: branch (high/medium/low/never_watched), prospect_email
  • video_followup_replied -- properties: branch, sentiment
  • video_followup_meeting_booked -- properties: branch, source (cal_link/direct_reply)

Build a PostHog insight comparing meeting conversion rate by engagement branch. This data validates whether video engagement scoring is predictive and calibrates the watch percentage thresholds.

5. Feed learnings back to video recording

After 2+ weeks of data:

  • If high-engagement prospects (>75% watched) are NOT converting to meetings: the video content is engaging but the CTA or value prop is weak. Adjust the video script's closing section.
  • If most prospects fall into "never watched": the email subject line or thumbnail is not compelling enough. Test different subject lines via ab-test-orchestrator drill.
  • If medium-engagement (25-75%) is the largest bucket: videos are too long or lose viewers mid-way. Shorten videos and front-load the most compelling content.

Output

  • Automated n8n workflows that route prospects based on video engagement
  • Personalized follow-up emails triggered by watch behavior
  • Attio contact records enriched with video engagement data
  • PostHog tracking of follow-up branch effectiveness
  • Data-driven recommendations for video script and email optimization

Triggers

The polling workflow runs every 4 hours on business days. Follow-up workflows trigger on each new view event. The "never watched" workflow runs daily.