Evidence
Not marketing work done for someone else — a marketing platform designed, built and tested from zero.
Design a free, voluntary ad-exchange network where local businesses barter ad space on each other's social pages instead of paying platforms, monetized by optional paid services (post generation, automated posting, analytics) around the free core.
Built the platform at localcircle.ca: a 20-niche onboarding flow that auto-generates sample ads per industry, a bronze/silver/gold credit system to balance who-posts-for-whom, and a referral program. Built the engine to run it: a coded-creative tracking system (every ad gets a PROJ-FMT-YYMM-SEQ code and hook/catch/CTA tags before publishing), an automated reel/carousel production pipeline, and scheduling automation. Then ran a real paid test — Wave 1, real estate niche.
A small, honest test, read the way the analysis itself reads it: CAD $8.06 spent on a 2-second-view objective across five creatives → 1,495 impressions, 617 video views (41.3% blended hook rate), 5 link clicks. Site-wide: 245 GA4 page views, 38 unique users on the /tune-up/ landing page but only ~10 seconds average engagement, and zero conversion events recorded — because none were wired up yet. The analysis diagnosed its own gap (no GA4 key events, no Meta Pixel conversion event) and prescribed the fix before any scaling call was made.
The transferable skill here isn't the result — it's the discipline: hook-rate analysis that normalizes out uneven ad-platform budget allocation, a funnel read that catches a landing-page leak before blaming the ads, and refusing to scale spend until the conversion event actually fires. That method applies to any new paid-channel test, on any budget.