Case studies·Tools

Fathom Analytics

Privacy analytics — Laravel, then Vapor, then SingleStore

Jack Ellis and Paul Jarvis rebuilt Fathom on Laravel, hit Heroku limits, moved to Laravel Vapor on AWS Lambda, then swapped RDS/Redis/Dynamo for SingleStore when database pain arrived. Still a tiny team, still anti-cookie-banner analytics.

2018–2019

Heroku Laravel, privacy as product

Early SaaS · privacy-focused GA alternative

Fathom launched as simple, privacy-friendly analytics — a product bet, not an infra flex. Their moved-to-Vapor post describes early Heroku hosting and the usual Laravel stack before traffic and ingest volume got interesting.

They publicly debated “does Laravel scale?” with the honest answer: most apps die from database shape, not framework choice.

Lesson

Pick the framework that lets you ship; pick the database when you know your write pattern.

2019–2020

Ground-up rebuild on Laravel Vapor

Separate collector env for high-volume ingest (2020 blog)

Ellis rebuilt Fathom from scratch and migrated to Vapor — serverless Laravel on AWS — to ditch always-on servers while keeping autoscaling ingest. They split “dashboard” and “collector” environments with shared databases via vapor.yml.

Their one-year Vapor review (August 2020) reads like a love letter: zero server maintenance, automatic scaling, and candid notes about AWS service limits (Parameter Store throughput, etc.).

Lesson

Serverless is not sin if your workload is spiky and your team hates pager duty. Split environments before you split monoliths.

2021–today

SingleStore replaces the database zoo

High-traffic analytics · hundreds of millions of pageviews/month on collector (2020 blog)

In “Does Laravel scale?” they describe graduating from managed Postgres/Redis/Dynamo pain to SingleStore for analytics-shaped queries — still Laravel at the edge, different storage engine underneath.

The stack page and blog continue as living docs: WAF, ALB, Vapor-managed AWS primitives — complexity added only where revenue justified it.

Lesson

Swap the analytics store when queries hurt, not when Twitter says Laravel “doesn’t scale.”

Sources

Facts drawn from public engineering posts and interviews. Numbers are approximate where sources disagree — we're stack advisors, not historians.

Get the right stack for YOUR scale — not theirs

Fathom Analytics's stack made sense for Fathom Analytics. Yours won't — and copying theirs is how you end up with a $4k/month bill and three unused Kubernetes clusters.

Plan my stack — free