Win back the customersyou already earned.
Four Claude Skills read your HubSpot or GoHighLevel contacts, rank who's worth reviving, and draft the re-engagement sequence in your voice — while refusing anyone you can't reach or who opted out. It drafts; you clear consent and send.
Update notice — version 1.2.0 (2026-11-14). This package replaces version 1.1.0 of the CRM Win-Back System. What did not change: the four weights (35 / 25 / 20 / 20), the score thresholds (65 and 40), the contact-permission precondition, the Suppression gate, the shaky-channel cap, the list gate (40%), how the contact to start with is chosen, and every result name. Scores do not change. What changed: three things. A contact that scores 65 or more with no gate fired and any of the four scored controls marked 0 now reads NURTURE instead of PRIORITIZE FOR WIN-BACK REVIEW (the top-band rule). No win-back score is shown at DO NOT CONTACT or PERMISSION REVIEW REQUIRED, and the result sentence there no longer opens with the score band. Every contact's result now names its priority control. Re-run earlier rows: yes. A row that read PRIORITIZE FOR WIN-BACK REVIEW with a scored control marked 0 now reads NURTURE, and the list's GO / HOLD verdict and the contact to start with can change with it.
Your warmest pipeline is the list you already stopped emailing.
Past customers who already trusted you and paid you are parked in your CRM, untouched. They're easier to win back than a stranger is to win — if you act on the right ones.
Email the whole dead list and you hit spam traps, torch your sender reputation, and message people who asked you to stop. The list gets worse, not better.
Some lapsed contacts are gold; most aren't. Working them by hand — or not at all — means your effort lands everywhere except where it pays.
Every contact ranked — and the one to start with.
This is the win-back scoring core, live. Set the contact-permission precondition, tune the four levers, and watch where a contact routes — then see how the dormant list ranks, whether it's worth working today, and who to start with. The precondition and the gates are the honest part: a contact without documented permission, or without a working channel, is held back, whatever their value.
A working email address or phone number does not establish permission to contact.
Contact permission and suppression status reviewed is marked “Opted out, do-not-contact, or another known prohibition,” which triggers this instrument’s Contact permission precondition gate and forces the final result to DO NOT CONTACT. No score is shown at this result.
Priority control: Contact permission and suppression status reviewed
Contact permission is a precondition, not a score input: only a mark of 2 lets a contact proceed to priority scoring, subject to the sender’s review. The engine and workbook enforce the same rule.
PRIORITIZE FOR WIN-BACK REVIEW: This is a marketing-priority result, not permission to contact. Confirm the applicable contact basis, channel rules, and suppression status before sending or calling.
This is a marketing-priority result, not permission to contact. Confirm the applicable contact basis, channel rules, and suppression status before sending or calling.
Demo resets on reload. Bring your own HubSpot / GoHighLevel MCP for the live list. It ranks and drafts; you clear consent and send. No reactivation rate promised. Email and SMS rules vary by jurisdiction and channel. This result ranks marketing priority; it does not grant permission to contact.
- PRIORITIZE FOR WIN-BACK REVIEW — score 65 or more, with contact permission 2, channel reachability 3–5, and none of the four scored controls marked 0.This is a marketing-priority result, not permission to contact. Confirm the applicable contact basis, channel rules, and suppression status before sending or calling.
- NURTURE — contact permission 2 and score 40–64; or contact permission 2 and a score of 65 or more with channel reachability 2 (shaky), which caps the result here; or contact permission 2 and a score of 65 or more with any of the four scored controls marked 0 (the top-band rule).
- LET GO — contact permission 2 and score under 40.
- PERMISSION REVIEW REQUIRED — contact permission is 1. Not prioritized until permission is documented. No score is shown at this result.
- DO NOT CONTACT — contact permission is 0, or contact permission is 2 and channel reachability is 0–1. No score is shown at this result.
Top-band rule. A contact that scores 65 or more with no gate fired, but with any of the four scored controls marked 0, reads NURTURE instead of PRIORITIZE FOR WIN-BACK REVIEW, and its result names each control marked 0.
Priority control. Every contact's result names one. At contact permission 0 or 1 it is the permission control. Otherwise it is the lowest-marked of the four scored controls; ties go to channel reachability when a gate or the cap read it, then to the heaviest weight, then to question order. With all four at 5 the result reads “No scored control is below the top mark.”
Contact permission and suppression status reviewed: 0 — Opted out, do-not-contact, or another known prohibition. 1 — Permission or applicable outreach basis has not been verified. 2 — Permission or applicable outreach basis has been documented, and suppression lists were checked. This remains a user-supplied assessment, not a legal determination.


How the gate works, in one image
Four levers scored into a 0–100 win-back number, and the contact-permission precondition that forces DO NOT CONTACT on an opted-out contact regardless of value — the same math the demo runs, as a diagram you can share or embed anywhere.
View & embed the full diagramOne pipeline, four Skills, from dormant list to revived revenue.
The scout reads your dormant contacts via the HubSpot or GoHighLevel MCP, records a contact-permission mark for each, and sorts each into one of the five results listed under the demo above, then names the one to start with.
The offer architect builds an honest, segment-matched reason to return — real value first, sized to how good a customer they were, with no fake urgency.
The drafter writes the multi-touch re-engagement in your voice, grounded in the real relationship, opt-out built into every touch and any offer flagged [decision needed]. It never sends.
You clear consent and send. The re-engagement tracker suppresses opt-outs first, flags who came back, and stops the cadence for them so a won-back customer isn't over-messaged.
Plus a runnable Python win-back engine with a worked sample list, a spreadsheet that reproduces the same math and gate, the win-back sequence template, and two playbooks — the end-to-end loop and the CRM MCP setup & compliant win-back guide.
Three rules the system holds to.
Scoring runs on your actual CRM through the MCP. It never fabricates spend, history, engagement, or consent status — and stops rather than guessing if the connection is missing.
It reads and drafts only — it never sends, edits, re-subscribes, or deletes a contact on its own. Every message is yours to review and send.
Contact permission is a precondition, not a score input: a known prohibition is DO NOT CONTACT and unverified permission is PERMISSION REVIEW REQUIRED, whatever the score. With permission documented (mark 2), a contact with no working channel is DO NOT CONTACT too. A working email address or phone number does not establish permission to contact.
An orchestration layer on top of your CRM — honest about its limits.
- A repeatable loop that ranks who's worth winning back and drafts the touches in your voice.
- The orchestration layer — Skills, scoring, a template, and playbooks — on top of your CRM + Claude.
- Built to protect your sending reputation: it refuses the unreachable and the opted-out.
- Not a CRM or a source of contacts — you bring your own HubSpot or GoHighLevel account and MCP.
- Not an auto-sender: it reads and drafts only, and never messages, edits, or deletes on its own.
- Not permission to contact and not a promised reactivation rate — and not legal advice.
This ranks who's worth winning back and refuses the unreachable — it does not grant permission to contact. Run a compliance gate (we recommend the Database Reactivation Readiness Engine) and confirm consent before you send, and honor every opt-out. Email and SMS rules vary by jurisdiction and channel. This result ranks marketing priority; it does not grant permission to contact. This is a marketing workflow, not legal advice.
Built for people with a list they've let go quiet.
- A founder, operator, or marketer with past customers parked in HubSpot or GoHighLevel.
- Sitting on a list you know has value but are afraid to blast.
- Willing to connect your CRM's MCP, clear consent, and stay the sender.
- Want a bot that auto-blasts your whole old list and sends on its own.
- Won't connect a CRM, clear consent, or review drafts before sending.
- Expect a guaranteed reactivation percentage rather than a disciplined workflow.
Clear the list, understand it, and speak to it.
The compliance gate that runs first — the formal consent verdict before you win anyone back.
Understand why they went dormant in the first place, so the win-back actually lands.
Codify your voice so every win-back touch sounds like you, not generic AI.
The honest answers, before you buy.
Yes — one of them, connected to Claude via its MCP. This system is the orchestration layer on top of your CRM: it reads your dormant contacts, scores them, and drafts the re-engagement sequence. It works with either HubSpot or GoHighLevel. You bring your own account; the kit doesn't store, forward, or resell your contacts.
Each contact is scored 0–100 from four levers whose weights sum to 100: relationship value (35), engagement signal (25), fit now (20), and channel reachability (20). One more mark is a precondition, not a score input: contact permission and suppression status reviewed. A permission mark of 0 (opted out, do-not-contact, or another known prohibition) is DO NOT CONTACT, and a mark of 1 (permission or applicable outreach basis has not been verified) is PERMISSION REVIEW REQUIRED, whatever the score. At a mark of 2 the score decides: 65 or more is PRIORITIZE FOR WIN-BACK REVIEW, 40–64 is NURTURE, under 40 is LET GO — with two reachability rules that apply at a permission mark of 2: channel reachability at or below 1 (no working channel) forces DO NOT CONTACT, whatever the win-back score, and channel reachability at 2 (shaky) caps the result at NURTURE — and a top-band rule: a score of 65 or more with any of the four scored controls marked 0 reads NURTURE. No score is shown at DO NOT CONTACT or PERMISSION REVIEW REQUIRED. Every contact's result also names a priority control: the permission control at a permission mark of 0 or 1; otherwise the lowest-marked of the four scored controls (ties go to channel reachability when a gate or the cap read it, then to the heaviest weight, then to question order), or “No scored control is below the top mark.” when all four are at 5. About PRIORITIZE FOR WIN-BACK REVIEW: This is a marketing-priority result, not permission to contact. Confirm the applicable contact basis, channel rules, and suppression status before sending or calling. The runnable engine, the workbook, and the demo on this page all use exactly this math, so a result you see in the demo is the result you'll get.
No. It drafts and prioritizes — it never sends, schedules, edits, re-subscribes, or deletes a contact on its own. Every message is yours to review and send. The re-engagement tracker reads results and flags who came back and who opted out, but you make the changes in your CRM.
No — and that's the point. This ranks who's worth winning back and refuses the unreachable, but it does not grant permission to contact. The formal consent decision belongs to a compliance gate (we recommend the Database Reactivation Readiness Engine) and your own process. Run that first, confirm consent, and honor every opt-out. The contact-permission mark you supply remains a user-supplied assessment, not a legal determination. Email and SMS rules vary by jurisdiction and channel. This result ranks marketing priority; it does not grant permission to contact. This is a marketing workflow, not legal advice.
Four Claude Skills (segment scout, offer architect, sequence drafter, re-engagement tracker), the runnable Win-Back Engine (Python, no dependencies) with a sample list, the same scorer as an Excel workbook, a win-back sequence template with opt-out built in, and two playbooks — the end-to-end workflow and a CRM MCP setup & compliant win-back guide.
No, and any tool that promises a reactivation rate is guessing. Win-back depends on who your dormant contacts are, why they left, and your offer. This system points your effort at the contacts most likely to come back, keeps you off the ones you shouldn't touch, and drafts honest touches in your voice — then reports the real re-engaged / silent / opted-out numbers rather than a flattering one.
Your dormant list isn't dead.
Win back what you earned.
One purchase, lifetime access, 12 months of updates. $149, once. Bring your own HubSpot or GoHighLevel MCP.
Sold by RedHub AI LLC · Secured by Stripe · redhub.ai