Portfolio · Case study
Re-platforming an 11-location RV dealer group’s website — before and after.
La Mesa RV sells RVs across 11 locations in four states. As Head of Marketing & Creative Director over 3.5 years (May 2021 → December 2024), I led the replacement of a legacy platform that couldn’t support modern lead capture — or attribute a dollar of a broadcast budget north of $1M a month — with a modern, measurable web stack. Once the digital side could be measured, and broadcast could show no evidence it was working, $500K a month of that budget moved into digital. This page shows the same four page types on the old site and the new one, with the technical differences extracted from archived source, not from memory.
- Rooftops, 4 states
- 11
- Broadcast, no evidence it worked
- $1M+/mo
- Moved to digital
- $500K/mo
- Site traffic
- 3×
- Leads per month
- 10,000+
01 · The starting point
The site as I inherited it.
The baseline captures on this page are archived from May–June 2021. I joined in the last days of May 2021 — so the May captures predate my start entirely: this is the site exactly as I inherited it, recorded by a third-party archive before I’d touched anything.
An 11-location dealer group that had grown through trade shows and word-of-mouth, spending over $1M a month on broadcast with no attribution behind any of it. The website couldn’t support modern lead capture, customer feedback went unmanaged, and each store operated semi-independently with no unified reporting. That was the real ceiling — not an old-looking site, but a marketing operation flying blind: you can’t rebalance spend you can’t measure, and you can’t measure it through a platform with no event layer, no attribution, and no way to connect a lead to a store, a vehicle class, or a campaign.
02 · The re-platform
What changed, and why it was led from marketing.
The re-platform was led from marketing because the problem it existed to solve was a marketing problem. The case to leadership was never “the website is old” — it was that a seven-figure monthly broadcast budget could not be defended, rebalanced, or even honestly evaluated without knowing what actually produced leads and sales. A platform rebuild was the only way to get that answer: the old site couldn’t carry the event tracking, lead capture, or integrations that attribution requires.
So the rebuild was scoped around measurement, not aesthetics. The new platform — a modern server-rendered stack on a headless CMS — was wired so that a web lead became an attributable record: CRM integrated with the dealer management system, every inquiry tied back to its source, and near-real-time dashboards reporting by store, vehicle class, and campaign. Eleven semi-independent rooftops became one measurable funnel.
Measurement is what unlocked the marketing decisions this case study is actually about: with attribution in place on the digital side — and none to be found for broadcast, however hard we looked — the $500K/month was rebalanced from broadcast into digital, defensibly, number by number, while site traffic tripled and lead capture scaled past 10,000 a month. The platform also became the base for what came next: AI-assisted video and email workflows, and customer-feedback automation that lifted store ratings. The re-platform was the foundation; the attribution was the point.
It shipped in three deliberate phases, because a dealer group with eleven stores cannot afford to lose its search traffic while it rebuilds. Phase one, June 2022: the platform moved to the modern stack with the layout, page structure and URLs kept identical — only the colour palette and the new La Mesa logo changed. Same pages, same structure, no rankings put at risk. Phase two, June 2024: the first real redesign, again planned around SEO — no page dropped, and every change to the category and page structure carried a redirect on the back end. Phase three: the redesign that phase two made possible, by a designer, described below.
The clearest proof the foundation worked came at the end. In my final months I brought a dedicated UI/UX designer onto the team — and the platform’s design system and behavioral instrumentation (rage-click and form analytics that simply couldn’t exist on the old site) became the raw material for a data-driven design practice that kept measurably improving the funnel after my departure. A platform you can iterate on with evidence was the deliverable behind the deliverable. That chapter belongs to Sharvari Suresh, the UI/UX designer who joined my team in 2024 — her token-based design system and behavior-driven redesign of the product pages, documented in her own portfolio, drove results she reports at +40% lead submissions and 84% faster form completion.
03 · Before and after
The same four pages, side by side.
Left, the site as archived in May–June 2021, before I’d touched it. Right, the live site, captured September 4, 2026: the design I approved before my tenure ended on December 11, 2024, built out after I left on the platform delivered under my tenure.
04 · The technical delta
Extracted from archived source, not from memory.
Both columns come from third-party-archived HTML of the live site — four page types per era (homepage, inventory, location, unit detail), read straight out of the served source. Baseline captures May 6 – Jun 21, 2021; handoff captures Dec 4 – Dec 11, 2024, within the final week of my tenure — the delivered state, not a later snapshot. The inventory and location rows compare the same URLs in both eras. Where the data shows unfinished business, it says so.
| Metric | Baseline (May–Jun 2021) | At handoff (Dec 2024) |
|---|---|---|
| Homepage <h1> | none | present |
| Canonical tags | absent on every page | self-referencing on inner pages |
| Tag-managed analytics | hardcoded legacy analytics script (deprecated even then); no tag manager | tag-managed on every page |
| Open Graph tags | 0 on all pages | 2–5 per page |
| Platform fingerprint | 2000s-era server platform, jQuery-led front end | modern server-rendered component stack |
| Structured data (JSON-LD) | none | not yet shipped — on the roadmap at handoff |
Methodology: metrics are limited to head-level and structural signals that measure identically across both rendering architectures; DOM-composition-dependent counts (e.g. per-image attributes) aren’t cleanly comparable between a server-rendered legacy page and a hydrated app, and are excluded. Full per-page extraction with exact timestamps lives in the working archive.
Built end to end by Brian Fidler — fractional CMO & marketing technologist, Phoenix, AZ.