Brian Thomas Fidler

• 3x — Website Traffic
• 10,000+ — Monthly Leads

La Mesa RV’s 11-location customer journey depended on a legacy, IT-controlled website and disconnected CRM, content, inventory-search, analytics, and lead-capture systems. Routine changes were slow, inventory discovery was fragmented, and marketing lacked the control and measurement needed to improve acquisition across locations.

As head of marketing, I led the re-platform from discovery through rollout. I defined customer and business requirements, prioritized the roadmap, and aligned executives, sales, IT, an 11-person marketing team, agencies, vendors, and implementation partners. We created a marketing-led platform integrating the CMS, CRM, advanced inventory search, analytics, attribution, and lead capture, while sequencing delivery to keep acquisition running across all 11 locations. This case study reflects my 2021–2024 tenure; the live implementation has since changed.

Across the broader transformation, website traffic tripled and monthly lead capture exceeded 10,000. Marketing gained direct control of the customer journey, and leadership gained clearer visibility by location, vehicle class, source, and campaign.

A dealership’s video ad normally means a camera crew, an agency, and weeks per vehicle — so most inventory never gets one. The raw material that does exist, listing photos, is low-resolution and legally inert: automotive advertising triggers Reg Z / Reg M disclosure requirements that generic video tools don’t understand, and one missing disclaimer makes a spot unairable.

A salesperson pastes a vehicle listing link. The system reads the unit’s details and photographs, treats the photos as evidence — never inventing a vehicle — and re-imagines the actual car at broadcast resolution. A named human signs off on six review regions on every generated frame before a dollar is spent on animation. A disclosure engine pairs every legally triggering term with its required fine print, timed to travel on screen with its own offer, and the finished :15–:60 spot is assembled to a proven ad structure.

A working end-to-end pipeline, pre-launch: listing link to a broadcast-ready, compliance-checked :30 spot in hours — no camera, crew, or agency. The demo spot was produced entirely by the system from seven real listing photos.

Podcast and AI-audio experiences are usually passive. ParodyPod needed to make a user feel like a guest in a fictional show with multiple AI characters. The experience had to coordinate dialogue, interruptions, voices, pronunciation, moderation, and episode structure while maintaining character consistency and cost control, without an established interaction pattern to follow.

I conceived the product and translated the premise into a working experience, defining the user journey, show-pack architecture, conversational rules, and episode lifecycle. I designed and built multi-character orchestration, voice and pronunciation controls, interruption handling, moderation, closing logic, and human quality safeguards. I also developed Stripe fulfillment, referral tracking, analytics, cost metering, an operations dashboard, and automated budget protections, then tested multiple show formats and iterated on prompts, timing, state, and voice behavior.

ParodyPod is not yet in production, so market impact has not been measured. Prelaunch work validated the end-to-end product flow, while pricing research informed the decision to charge $3 per generated episode download.

Local HVAC companies lose calls to rivals who outrank them on the map and in AI assistants — while agencies sell $2,000/month retainers on trust. Build a self-funding engine around a $499 audit, instrumented so even a failed run yields a baseline.

One closed loop: map-index prospecting, the audit run for each prospect before first touch, plain-text sending with server-side UTM attribution, a three-event tag container, click-to-sale joins, and a spreadsheet CRM with instant reply alerts.

Live and early, the engine has already corrected its own strategy from data: caught $400 of $737 in wrong-intent ad spend, exposed a dead tracking container, and filtered bots from intent metrics. Cloning to a new vertical is a config change.