# Marcus Yuen, Founder at Fableshift (B2B SaaS, 6 employees) — read of analytics-stack-template, May 22, 2026

> 8 years building software, currently running a niche project management tool for architecture firms, 2,400 paying users and staring at GA4 like it owes me money.

## How I got here

Searched "self-hosted analytics for indie SaaS under 50 a month" on Sunday evening after the kids went down. My wife had forwarded me an Indie Hackers thread earlier in the week where someone mentioned PostHog self-hosted and I fell into a rabbit hole. This page showed up on page 2. I clicked because the title was blunt enough to not feel like SEO bait.

## What I clicked first

The problem statement pulled me in immediately. "The open-source alternatives feel like you're earning a PhD in data engineering instead of running your business." That sentence is accurate. I spent four hours last February trying to get Rudderstack working and abandoned it. So I kept reading.

The cost comparison table is the right call to open with. Amplitude at $1,000/mo versus this stack at $20/mo is the entire argument and they lead with it. That's a good instinct.

## Where I paused

The 30-day plan. Specifically Week 1: "Instrument your five core user flows. Don't instrument everything. Lean." I actually stopped and copied that into my notes app. It's the kind of advice that sounds obvious but most guides bury under 40 other steps. That section felt like someone who has actually done this, not someone who researched it.

## What I distrusted

Two things.

First, the scoring box at the bottom. "Financial upside: 1/10. Pain intensity: 4/10." That's the product they're selling me, rating itself a 4 out of 10 on pain intensity. If the pain isn't that intense, why am I here? It undercuts the entire problem framing above it. The page opens with "you need to understand how users behave, where they get stuck, why some churn" and then their own scoring engine says the pain is a 4. Which is it?

Second, the pricing structure confused me more than it should. "$5 for the dossier, $99 for the adopt build including working code starter." I genuinely do not know what "working code starter" means for an analytics stack that is fundamentally just three SaaS tools pointed at each other. Is it a docker-compose file? A dbt project skeleton? A Metabase dashboard export? The page does not say. "Brand assets and copy library" is even more mysterious for a backend analytics setup.

Also: "We don't have live customers on this idea yet." Okay, fair for disclosing it. But it makes the $99 buy feel like paying for someone's research doc, not a proven playbook.

## What would convince me

A single real implementation screenshot. Not a stock dashboard, not a generic Metabase screenshot from their website. A real user's retention cohort dashboard from a real app, even anonymized, with a caption like "this is what 8 months of cohort data looks like in Metabase on the $20/mo stack." That would close me.

Or: one paragraph from an actual indie builder describing what they learned from the churn drivers dashboard that they could not have learned from GA4. Specific. "We discovered that every user who churned in month 2 had never touched the bulk import feature, so we added it to the onboarding flow." That kind of thing.

The "working code starter" for $99 needs a spec. What files are in it. What it assumes. What I have to set up myself versus what is pre-built.

## What I'd ask in an email reply

1. The page recommends DuckDB for under 10M events/month. My app is around 400k events/month right now. What does the actual query performance look like at that scale when I'm joining event data against user attributes from Postgres? Is there latency I should plan for in dashboard load times?

2. What does the $99 "working code starter" actually contain? Specifically: is there a schema file, and does it come with pre-built dbt models or do I write those from scratch?

3. You rated pain intensity 4/10 on your own scoring. I'm curious what you think the ceiling on this business is, because if the pain is genuinely low, I want to understand why someone would pay $99 for it versus just reading the free tier and figuring it out themselves.

## Verdict: on-the-fence

The free content on this page is legitimately useful and the 30-day plan earned my trust. But I do not understand what I am buying for $99, and a product that scores its own pain at 4/10 while asking me to spend money needs to work harder to explain the gap between "this is all free-tier software" and "this package is worth $99."

---
*Memo by skeptic persona, generated 2026-05-22. Studio breaks own self-grading loop.*