I'm Paul Singh — 8+ years turning ad spend into revenue with Google & Meta strategy, AI-powered automation, and attribution systems built from scratch. The eagle view: big-picture strategy, executed down to the pixel.
Five engagements where a business handed me the budget — and the metrics moved. Click any card for the full breakdown.
Full-funnel rebuild across Google & Meta with server-side tracking and AI-assisted qualification — sustained qualified lead growth without inflating CPL.
Diagnosed and rebuilt a broken tracking setup with server-side CAPI before misattributed spend spiraled further.
Restructured targeting and creative for a luxury real estate brand — cost per lead dropped from $80 to $25 while volume grew 169%.
Local service campaign engineered around booked appointments — 54 service calls on the calendar within the first 30 days.
Unified bidding, creative, and reporting across a 12-location franchise portfolio — consistent performance with local nuance intact.
Not campaign management — engineering. Systems and experiments built with my own hands to remove bottlenecks the ads couldn't fix.
A custom React quoting tool for insurance agents — instant quotes instead of manual lookups, saving each agent hours every single day.
Automated document delivery triggered while the agent is still on the phone — the brochure lands in the lead's inbox before the call ends.
A 30-day A/B test routing outbound texts through iMessage connectivity vs standard SMS inside GoHighLevel — higher connection rates, more replies, more booked appointments.
In-house and agency roles across health insurance, real estate, home services, SaaS, and franchise brands — always in high-accountability environments.
About the case studies, the projects, and how an engagement actually works.
Case studies are client engagements where a business handed over a budget and a business metric moved — lead growth, ROAS, CPL. Projects are things I personally designed and built: software, automations, and experiments. They're labeled separately on purpose, so you always know whether you're looking at campaign results or engineering work.
Yes — every case study on this page is a real engagement I ran hands-on, and I can walk through the account structure, the decisions made, and the reporting behind each number in detail. Client names are withheld here out of confidentiality, but the work is mine and I can speak to all of it live.
Because I hold my own work to the same standard I'd want from a report. It was a 30-day A/B test — the iMessage route clearly outperformed standard SMS on connection rates, replies, and bookings — but it ran with platform limitations around number porting in GoHighLevel and wasn't a full production rollout. Strong signal, honestly labeled.
The specifics are vertical-dependent, but the method transfers: clean server-side tracking first, so the platforms optimize on real outcomes; campaign structure built around intent; and automation that removes the human bottlenecks after the lead arrives. That same playbook produced results in insurance, real estate, HVAC, and a 12-location franchise portfolio — four very different businesses.
Both. I'm open to Director/VP-level full-time roles as well as contract engagements — audits, tracking rebuilds, campaign turnarounds, or ongoing management. Reach out with what you have in mind and I'll tell you honestly whether I'm the right fit.
I've managed $50M+ in cumulative ad spend, from local-service budgets (the HVAC engagement) up to multi-location portfolios. What matters more than size is accountability: I work best where spend is tracked to revenue and the numbers are expected to be real.
Tracking audit first — most accounts are optimizing on bad data, and no strategy survives broken attribution. Then account restructure around intent, then testing velocity. The HVAC case study is the pattern in miniature: 54 booked service calls inside the first 30 days because measurement and structure came before scaling.
No — the Agent Portal, the mid-call automation, and the server-side CAPI setups on this page were built by me. I collaborate well with engineering teams when they exist, but I don't wait on one to ship.