Jet Sports Trader national watercraft marketplace

What we delivered

• Hero search with make, model, location and price entry points, featured placement rail and new arrivals rail
• Faceted search across make, model, year range, price, condition, radius and dealer or private seller, with fuzzy matching so misspelled model names still resolve
• Saved searches with email alerts so buyers are notified when matching inventory appears
• Listing detail with de-duplicated image gallery, structured spec block, quote dialog, view tracking and Product plus Vehicle structured data
• Dealer directory of 124 dealers with profile pages, inventory tabs, state and brand slices and a moderated dealer application flow
• Three step selling funnel: sign up, guided multi step listing form with AI assisted description drafting, optional paid upgrade
• 30 day listing lifecycle with trigger stamped expiry, hourly automated retirement and an owner scoped renew action
• Stripe revenue lines: $50 featured listing placement and recurring dealer subscription plans
• Valuation calculator estimating market value from year, make, model, hours and condition
• Financing calculator producing monthly payment, total interest and total cost, with emailable quotes and rich result markup
• Buying guides plus one used jet ski guide per state, and brand and model landing pages with live inventory
• Programmatic location page families covering city, used, PWC and brand plus city intent across roughly 1,700 cached pages
• Hybrid sitemap architecture of 1,525 URLs with RSS feeds and IndexNow pings, plus an in app SEO dashboard and audit page
• Automated content engine publishing a new article every two days with a normalisation and sanitisation pipeline
• Contextual auto linking that maps 25+ commercial phrases into article bodies at render time, so the whole back catalogue benefits from every new landing page
• Operator audit tooling for internal links and article formatting, including a crawler view that fetches the live URL as Googlebot
• Fully responsive from 320px with a PWA manifest, service worker and prioritised image pipeline

Engineering highlights

1. 1

   Static generation first: real HTML is produced per route so every commercial page is crawlable and fast, with a Cloudflare and Prerender.io layer serving rendered HTML to bots for anything not statically built.

2. 2

   Content is pre-generated and cached in the database, never generated at request time, with a cache dashboard, batch automation and job queue managing generation at scale.

3. 3

   On page discipline enforced in code: exactly one H1 containing the primary keyword, titles clamped under 60 characters, descriptions between 50 and 160, canonicals normalised to the apex domain and sanitised JSON-LD.

4. 4

   Access control treated as a feature: row level security on every user facing table, roles in a dedicated table checked through a security definer helper, and billing columns unwritable by users.

5. 5

   Payment integrity closed at the data layer. Users can only insert pending rows at the fixed amount, and only server side Stripe verification can promote a payment to completed.

6. 6

   Automation over headcount: hourly listing expiry, scheduled publication, saved search alert checks, cache maintenance and search engine pings all run on cron.

Technical stack

Front end

React 18, TypeScript, Vite, Tailwind CSS and a shadcn/ui component layer with semantic design tokens and route level code splitting

Static generation

vite-react-ssg with per route static path generation from the database and a custom hook injecting per page head metadata

Backend

Managed Postgres with 148 migrations, row level security, triggers and cron jobs

Serverless

88 edge functions across content generation, sitemap and RSS, cache and job orchestration, payments, transactional email, alerts, moderation and audit tooling

Payments

Stripe Checkout for one off featured listings and recurring dealer subscriptions with server side verification

Delivery

Custom apex domain with www redirecting to apex, fronted by a Cloudflare worker and Prerender.io for bot traffic

ROI model

Growth comes from owned organic surface area rather than paid acquisition, while revenue is captured on upgrades instead of access, so listing supply and search demand compound together.

Inside the platform