Full GHL system for a migration & education business.
The migration consulting business was juggling five sub-accounts, thousands of leads across broken pipelines, and a team drowning in manual follow-ups. Two months later they were running 11,000+ opportunities through a single, coherent system. The team was focused on high-value conversations instead of chasing DMs.
The business was growing faster than the system could handle.
The client operated five distinct sub-brands under one migration and education group: Aussie Migration, L2C Education, CourseApply, Luma Labs, and Mariam Medicare. Each brand ran its own channels, qualification, and tracking. Spreadsheets, half-configured GHL sub-accounts, manual follow-up chains.
The user problem: leads went cold before anyone got back to them.
The business problem: the team couldn't tell where anyone was in the funnel.
The objective: one coherent GHL setup, so the team could focus on qualifying calls instead of chasing leads.
Strategy, build, and continuous maintenance.
I led the entire automation architecture: from initial discovery through eight weeks of build and into ongoing management. I collaborated with the client's operations director and their sales team, with occasional input from their marketing lead.
Understanding how the business ran.
Before touching any GHL settings, I spent the first week interviewing the sales team, shadowing the operations director, and mapping the current-state lead journey across all five sub-brands.
Interviews
- Sales team (4 people) — where leads got stuck, what took the most time, what they wished existed
- Operations Director — reporting needs, forecast requirements, team performance visibility
- Marketing Lead — ad performance data, funnel drop-off, campaign attribution
Key findings
- Response time was the killer — average first response was 4h 20m during business days; leads who waited >1 hour were 68% less likely to convert
- Duplicate tracking — the same lead often existed in 2–3 sub-accounts if they engaged with multiple brands
- No handoff logic — leads went from marketing to sales via WhatsApp screenshots (yes, really)
- No lead scoring — the team treated every enquiry the same, even obvious tyre-kickers
Defining what to solve.
"The business had five funnels. It needed one system, five brands."
Primary problem statement
Leads across five sub-brands were being lost to slow response times, duplicated tracking, and the absence of any single source of truth. The team was overwhelmed with volume they couldn't measure.
Design principles I set for the build
- Every lead lives in one place — no matter the brand
- Every stage triggers itself — no manual moves
- Sales sees only what they need — hot leads, ranked, with context
- Operations sees the whole picture — reporting and forecasts across brands
- Edge cases don't break it — duplicates, channel conflicts, off-hours leads
Sketching the system.
I mapped the target architecture in a whiteboard session before any GHL configuration. Each sub-brand kept its own funnel intake, but everything flowed through a shared master pipeline where lead scoring, routing, and follow-ups were centralised.
Information architecture
- 5 sub-accounts (one per brand) preserved for brand-specific automation
- 1 master pipeline with 8 stages: Captured → Contacted → Qualified → Called → Proposed → Won/Lost
- Cross-account tagging for lead attribution
- Automated de-duplication using phone + email
Building the eight-stage pipeline.
Over six weeks of build I created the pipeline stages, custom fields, tagging structure, sub-account permissions, and every automated transition. The team reviewed each stage as it went live and flagged mismatches with how they operated.
Workflows built
- Lead capture & routing — enquiries route to the correct sub-account based on ad source, form field, or channel
- First-response automation — instant WhatsApp acknowledgement, calendar link, brief qualification questions
- Show/no-show handling — different sequences depending on whether the lead attends their scheduled call
- Nurture sequences — long-form email + WhatsApp for leads not ready to book
- Re-engagement — leads dormant for 30 days receive tailored re-engagement based on original enquiry
What else was built.
Beyond the main pipeline, eleven components filled in the operation across GoHighLevel and n8n. From lead capture through invoicing.
Workflows & sequences
AI components
Reporting, funnels & campaigns
Testing with real leads, not sandbox.
Each phase went live to real leads under close monitoring before I moved on. This surfaced edge cases that no amount of pre-planning would have caught.
What we caught & fixed
- Duplicate lead flooding — 12% of leads engaged with 2+ brands. Added phone-based de-dupe with brand-preservation logic.
- Off-hours WhatsApp mismatch — automated messages felt cold at 2am. Added quiet-hours routing.
- Calendar timezone bugs — international leads seeing wrong times. Fixed timezone detection in booking flow.
- Team over-notification — sales team drowning in Slack pings. Batched notifications into hourly digests.
The system after six months.
Business impact
- Response time collapsed from 4h to 2 minutes for the automated first touch. The biggest single conversion improvement
- Team focused on qualified conversations only — nurture and cold outreach handled automatically
- Cross-brand reporting exists for the first time — operations can forecast and reallocate spend
- Zero data loss across five sub-accounts since launch
"Islam brought structure and clarity to a complex product challenge. The final system was not only elegantly built but significantly easier for our team to operate."
— Operations Director, Migration Consulting
What I'd do differently.
Two things stand out looking back at the build.
Duplicate detection should have been phase one, not phase four. Once the de-dupe logic was in place, half the "weird bugs" we were seeing in the first three phases disappeared retroactively. I should have architected for it from the start.
The team training was the most valuable single deliverable. A working system that nobody understands still fails. I now spend more time on team walkthroughs and less on documentation nobody reads.
What comes next
- Adding an AI qualification agent as the new first-response layer
- Predictive lead scoring based on 12 months of pipeline data
- Cross-brand lead attribution modelling
CRM migration turning into a bigger project than you expected?
That's usually the shape of it. Book 30 minutes and I'll tell you what actually needs rebuilding and what can be lifted across.