Full Rewrite
Landing Analytics is a complete rewrite of L Analytics. The things I learned from L Analytics were used in this second rewrite. Not only tech-wise, but also about user experience and developer traceability
Concept
Landing Analytics is a tool designed for founders and indie hackers to track engagement on their landing pages and wishlists. All other platforms are bloated, with strict privacy regulations and too much noise, leaving analytics focused on four core metrics: time to conversion, traffic source, retention on the landing page, and flag events to correlate traffic.
Technical Architecture
After the complete rewrite, the domain’s architecture was far superior and more resilient. Events were filtered, batched, and sent to Cloudflare before landing in the database, causing a deadlock. With Cloudflare protection to filter malicious events and a strict API-like control, the application’s security was far superior. [CENSORED] did most of the heavy lifting of the application traffic-wise. It had excellent concurrent writing / reading capabilities, plus support for premade views. The backend, previously on PocketBase, was replaced entirely with Supabase + Clerk. Techstack: Deno.js, Nuxt, Next
Monetization
Never offer a free plan again. Trial is a far better approach. The conversion/leverage you have to convert free into paying customers is really zero. A free plan also attracts abuse from less affluent users and overloads backend services with unnecessary traffic.
Infrastructure Costs
- Domain costs: $10/year
- Backend: $0 - Deno Cloud
- Frontend: $0 - Vercel Despite 100s of users and dozens of active users, the app worked just fine on the free tier for Supabase and Clerk
What I Learned
- Never offer a free plan. Always do a trial from the start, you have much better control and leverage on new users signing up
- Questionnaire at the start. Always collect additional information on your users when they sign up. Where did they first hear about the platform? What are their expectations? What feature do they want to see?
- Onboarding. 80% of the time, people will use their first 5 - 10 second experience to determine if they stay or leave the application
- Have a basic logging system for the front end. You need to know what features of your app get the most interest from users, what is the path they take, whether someone is struggling, or just looking around
- Never use pocketbase again. After the first 50 users Pocketbase starts breaking. Growth in apps is exponential. This means your back office database will break overnight
- People do read the docs. The application docs are one of the places users spend the most of their time. More than the landing page, before making a purchase decision.