Your marketing agent converted a trial
before the drip campaign fired.

Create a segment for stalled trials

Trial users who created a mailbox but never sent their first email. They got through setup, then stopped. The segment recomputes automatically, so new users who stall will enter the journey on their own. No polling required on your end.

molted segments create \
  --name "Trial Stalled at Setup" \
  --filters '{
    "rules": [
      { "field": "plan", "op": "eq", "value": "trial" },
      { "field": "mailboxes_created", "op": "gte", "value": 1 },
      { "field": "emails_sent", "op": "eq", "value": 0 },
      { "field": "signup_days_ago", "op": "gte", "value": 2 }
    ]
  }'

Create an event-driven journey

This journey fires when a user matches the stalled segment. You add steps next.

molted journeys create \
  --name "Trial Onboarding - Stalled Users" \
  --trigger-event "segment:trial-stalled-at-setup" \
  --mailbox marketing-agent

Add journey steps

A journey is a sequence of steps. send delivers an email, delay waits, branch checks a condition, end terminates. The interesting part below is the branch at position 3: after waiting 48 hours, it checks whether the user actually sent their first email. If they did, the journey ends quietly instead of sending a nudge they no longer need.

# Step 1: Send the nudge
molted journeys add-step jrn_abc123 \
  --type send \
  --template trial-stalled-nudge \
  --position 1

# Step 2: Wait 48 hours
molted journeys add-step jrn_abc123 \
  --type delay \
  --delay-hours 48 \
  --position 2

# Step 3: Did they activate?
molted journeys add-step jrn_abc123 \
  --type branch \
  --condition '{ "field": "emails_sent", "op": "gte", "value": 1 }' \
  --on-true end \
  --on-false 4 \
  --position 3

# Step 4: Second nudge (only if still stalled)
molted journeys add-step jrn_abc123 \
  --type send \
  --template trial-second-nudge \
  --position 4

# Step 5: End
molted journeys add-step jrn_abc123 \
  --type end \
  --position 5

A/B test the welcome email

Two variants of the same template. Molted splits traffic automatically, but here's the part that matters: it tracks which variant drives more activations and conversions, not just opens. Open rates tell you about subject lines. Outcome attribution tells you about revenue.

molted templates create \
  --name trial-stalled-nudge \
  --subject "You created a mailbox — here's the next step" \
  --body "Hey {{name}}, you set up {{mailbox_name}} but haven't sent yet.
Here's a 2-minute quickstart: {{quickstart_url}}" \
  --auto-publish

molted templates create \
  --name trial-stalled-nudge \
  --subject "Your mailbox is ready. Send your first email." \
  --body "Hey {{name}}, {{mailbox_name}} is live and waiting.
One API call and you're sending: {{quickstart_url}}" \
  --variant-of trial-stalled-nudge \
  --split 50

Fire the trigger event

Before the journey sends anything, run a fatigue check and a policy simulation. The fatigue check looks across all agents, not just this one. If the CS agent already emailed Maria yesterday, the marketing nudge gets held until the cooldown window clears. The simulate call shows you exactly which of the 22 policy rules passed and whether anything would block.

# Check fatigue first
molted analytics fatigue --contact maria@startup.io
# → { "recommendation": "safe_to_send", "recentSends": 0,
#     "windows": { "24h": 0, "7d": 1, "30d": 1 } }

# Dry-run the policy check
molted send simulate \
  --to maria@startup.io \
  --template trial-stalled-nudge \
  --mailbox marketing-agent
# → { "status": "allowed", "rulesEvaluated": 22,
#     "passed": 22, "blocked": 0 }

# Trigger the journey
molted journeys trigger jrn_abc123 \
  --contact maria@startup.io \
  --payload '{
    "name": "Maria",
    "mailbox_name": "my-app",
    "quickstart_url": "https://docs.moltedemail.com/quickstart"
  }'

Record the conversion

Maria sends her first email, then upgrades to paid three days later. Record both outcomes. Molted traces them back to the journey step that started the conversation, so you know which nudge (and which A/B variant) drove the conversion.

molted outcomes ingest \
  --contact maria@startup.io \
  --event-type activation \
  --event-name "First Email Sent" \
  --metadata '{ "mailbox": "my-app", "days_since_signup": 3 }'

molted outcomes ingest \
  --contact maria@startup.io \
  --event-type conversion \
  --event-name "Trial to Paid" \
  --revenue 49 \
  --metadata '{ "plan": "starter", "days_since_signup": 6 }'

# See what worked
molted outcomes journey-impact jrn_abc123
# → { "conversions": 24, "revenue": 1176,
#     "attribution": "last-touch",
#     "topVariant": "trial-stalled-nudge:B" }