Re-Engagement Campaign Builder

Identify dormant Mailchimp subscribers by signup source and interest tags, then generate personalized multi-step re-engagement email sequences using Mavera Agent and Generate

Scenario

Subscribers who haven’t opened an email in 90+ days are effectively lost — unless you can win them back. You pull inactive subscribers from Mailchimp, analyze their original signup source and interest tags, then use Mavera to generate personalized re-engagement email sequences per segment. The result is a multi-step re-engagement campaign that addresses why each subscriber went dormant.

Architecture

Code

Python

import os, requests, time
from datetime import datetime, timedelta
from collections import defaultdict

MC_KEY = os.environ["MAILCHIMP_API_KEY"]
MC_DC = os.environ["MAILCHIMP_DC"]
MV = os.environ["MAVERA_API_KEY"]
MB = "https://app.mavera.io/api/v1"
MH = {"Authorization": f"Bearer {MV}", "Content-Type": "application/json"}
MC_BASE = f"https://{MC_DC}.api.mailchimp.com/3.0"
mc_auth = ("anystring", MC_KEY)

lists_r = requests.get(f"{MC_BASE}/lists", auth=mc_auth, params={"count": 1})
LIST_ID = lists_r.json()["lists"][0]["id"]

cutoff = (datetime.now() - timedelta(days=90)).strftime("%Y-%m-%dT%H:%M:%S+00:00")

inactive = []
offset = 0
while len(inactive) < 500:
    r = requests.get(
        f"{MC_BASE}/lists/{LIST_ID}/members",
        auth=mc_auth,
        params={
            "count": 100,
            "offset": offset,
            "status": "subscribed",
            "since_last_changed": "2025-01-01T00:00:00+00:00",
            "fields": "members.email_address,members.stats,members.source,members.interests,members.merge_fields,members.timestamp_signup",
        },
    )
    r.raise_for_status()
    members = r.json().get("members", [])
    if not members:
        break

    for m in members:
        avg_open = m.get("stats", {}).get("avg_open_rate", 0)
        if avg_open < 0.02:
            inactive.append(m)

    offset += 100
    time.sleep(0.5)

print(f"Found {len(inactive)} inactive subscribers (90d+ no opens)\n")

source_groups = defaultdict(list)
for m in inactive:
    source = m.get("source", "unknown") or "unknown"
    source_groups[source].append(m)

source_summary = "\n".join(
    f"- {source}: {len(members)} subscribers"
    for source, members in sorted(source_groups.items(), key=lambda x: -len(x[1]))[:8]
)

mave = requests.post(f"{MB}/mave/chat", headers=MH, json={
    "message": f"""Analyze why email subscribers go dormant and recommend re-engagement strategies per signup source.

INACTIVE SUBSCRIBERS (no opens in 90+ days): {len(inactive)} total

BY SIGNUP SOURCE:
{source_summary}

For each source group:
1. Why do subscribers from this source tend to go dormant?
2. What originally attracted them?
3. Best re-engagement angle (nostalgia, new value, curiosity, FOMO, direct ask)
4. Should we try to re-engage or clean the list?"""
}).json()

print("--- Dormancy Analysis ---")
print(mave.get("content", "")[:1500])
print()

for source, members in sorted(source_groups.items(), key=lambda x: -len(x[1]))[:3]:
    gen = requests.post(f"{MB}/generations", headers=MH, json={
        "prompt": (
            f"Create a 3-email re-engagement sequence for {len(members)} dormant subscribers "
            f"who originally signed up via '{source}'. They haven't opened an email in 90+ days.\n\n"
            f"For each email in the sequence:\n"
            f"1. Subject line\n"
            f"2. Preview text\n"
            f"3. Body (100-150 words)\n"
            f"4. CTA\n"
            f"5. Send timing (days after previous)\n\n"
            f"Sequence strategy:\n"
            f"  Email 1: Soft re-introduction (don't be desperate)\n"
            f"  Email 2: Value demonstration (show what they're missing)\n"
            f"  Email 3: Last chance (unsubscribe if no response)\n\n"
            f"Tone: Warm but not pushy. Acknowledge the gap."
        ),
    }).json()

    print(f"{'='*50}")
    print(f"Re-engagement: '{source}' subscribers ({len(members)})")
    print(f"{'='*50}")
    print(gen.get("output", gen.get("content", ""))[:800])
    print()
    time.sleep(0.5)

Example Output

Found 1,847 inactive subscribers (90d+ no opens)

--- Dormancy Analysis ---
## Source: "Import" (620 subscribers)
These were likely bulk-imported from another tool or event list. They
never had an organic relationship with your emails — re-engagement
rate will be low (~3-5%). Strategy: one "do you still want to hear
from us?" email, then clean the list.

## Source: "Signup Form" (480 subscribers)
Organic signups who went dormant. They once cared enough to fill out
a form. Re-engagement angle: remind them what they signed up for.
Show what's changed since they left.

==================================================
Re-engagement: 'Signup Form' subscribers (480)
==================================================
**Email 1: Soft Re-introduction (Day 0)**
Subject: We noticed you've been quiet
Preview: No guilt trip — just checking in.

Hi [NAME],

It's been a while since you opened one of our emails. No judgment
— inboxes are overwhelming.

Here's what happened while you were away: we shipped [Feature],
redesigned [Page], and published [Popular Post Title].

If any of that sounds interesting, just click below. If not, no
hard feelings.

→ [Show Me What I Missed]

**Email 2: Value Demo (Day 4)**
Subject: This saved our users 6 hours last week
Preview: The feature you might not know about.

**Email 3: Last Chance (Day 10)**
Subject: Should we stop emailing you?
Preview: Click to stay — or we'll unsubscribe you automatically.

Error Handling

Filtering inactive subscribers

Mailchimp doesn’t have a direct ‘last open date’ filter in the Members API. The code uses avg_open_rate < 0.02 as a proxy for dormancy. For more precise filtering, use Mailchimp’s built-in segment builder with ‘Did not open’ conditions, then query that segment.

Pagination with offset

The Members API uses offset pagination (not cursor-based). For lists with 100k+ members, this can be slow. Consider using the batch endpoint or exporting via the Export API for large lists.

Re-engagement compliance

Always include an unsubscribe link in re-engagement emails. The “last chance” email should honor the implicit opt-out — if they don’t engage, move them to a cleaned or archived state.

---

What’s Next

Mailchimp Integration

Back to Mailchimp integration overview

Interest Content Generation

Generate content by audience interest

Open Rate Prediction

Test subject lines before sending

Mave Agent

Full reference for POST /api/v1/mave/chat