AI Advisor
← Case Studies

Recovering a 3,000-Contact Dormant Database

Worked scenario, modelled on published benchmarks. Figures are conservative and sourced below.

An open filing cabinet drawer packed with hanging paper folders, in a dim office at dusk with a city skyline through the window.

The problem this system solves

Dormant leads are rarely dead — they’re unfollowed-up. The most rigorous audit of this behaviour remains Harvard Business Review’s 2011 study of 2,241 US companies, which found the average first response to an inbound web enquiry took 42 hours, and that 23% of companies never responded at all. Every one of those contacts was paid for. Most are still sitting in the CRM.

A reactivation system doesn’t buy new attention. It works an asset the business already owns.

The build

A three-touch conversational sequence across SMS and email, orchestrated in n8n with GoHighLevel as the CRM and messaging layer:

  1. Scrubvalidate numbers and addresses, strip Do-Not-Contact records, suppress anyone who has ever opted out, and drop records with no lawful basis to contact.
  2. Segment by recencycontacts dormant under 90 days go into a separate, warmer sequence from contacts dormant over a year.
  3. Reference the original enquiryeach message names the specific thing they asked about. This is the single biggest driver of reply rate, and it's only possible because the CRM field is piped into the message body.
  4. Route repliesan if/else branch sends positive intent straight to a booking calendar, negative intent to suppression, and ambiguous replies to a human.
  5. Enforce opt-outs at the workflow levelSTOP-family keywords are caught and written back to the CRM before the next send fires.
A phone lying face up on a desk in low evening light, showing a single unread text message notification.
Three touches, across SMS and email. Each one names the thing they originally asked about — which is what the reply rate below rests on.

Modelled funnel

StageCountBasis
Raw CRM records3,000
Contactable after scrub2,400~20% loss to invalid, DNC and prior opt-outs
Replies~1456% reply rate
Positive intent~58~40% of replies
Appointments booked~35
Attended~2570% show rate
Closed7–830% close rate
Revenue at $1,500 average ticket~$11,000
Messaging cost~$2202,400 contacts × 3 sends × ~$0.03

The point isn’t the revenue figure — it moves with ticket size. The point is the ratio. Treat any claim of 15–20% conversion across an aged database as a red flag. Low single digits, list-wide, is a strong result.

Compliance is part of the build, not a disclaimer

Reactivation touches people who haven’t heard from a business in months. Regulators treat that as high-risk, and the penalties are not theoretical:

Australia
ACMA fined Lululemon’s Australian arm $702,900 in March 2026 for 370,000+ emails without an unsubscribe option, and Latitude Finance $3.96m in April 2026 for messages where “reply STOP” didn’t function. The Spam Act 2003 requires consent, sender identification, and an unsubscribe honoured within 5 business days.
United States
the FCC’s TCPA revocation rules require senders to honour a standard set of opt-out keywords — stop, quit, end, revoke, opt out, cancel, unsubscribe — and to process a revocation within 10 business days. Compliance was due from April 2025, though the FCC has since waived parts of the revocation requirement into 2027; build to the rule, not to the waiver.
European Union
under GDPR and the ePrivacy Directive, contacting a lead who enquired two years ago requires a documented lawful basis; “they’re in our CRM” is not one.

Every one of these rules is a workflow branch. That’s what the blueprint above encodes.

The system described here is the one built on the lead reactivation project.

Talk it through