In short
- A side project built to fix the founder's own analytics annoyance became the entire product. Marvin Blum built the open-source Pirsch Go library in June 2020 to track his own blog without cookies or Google Analytics — then productized it in October 2020 after his first startup, Emvi, failed to compete with Notion and Roam Research.
- A deliberate two-launch Product Hunt strategy — beta first, full release second — beat treating PH as a single make-or-break moment. The beta launch drew ~75 upvotes and built an audience; mobilizing that audience, including the very first beta customer, for the full release about half a year later drove it to ~329 upvotes.
- MRR grew for five straight years — $76 (2021) → $1,418 (2022) → $11K (2024) → $14.3K (2025) — without a single dollar of paid ads. But the growth rate itself decayed: +83% in 2024 slowed to +24% in 2025, a monthly rate drop from 1.4% to 0.7%.
- The founders' own explanation for the slowdown isn't a hidden growth-hacking myth — they say the AI era made their own playbook copyable. Marvin Blum: “I would assume that AI has lowered the barrier to creating SaaS products that people don't know much about” — a rare founder-sourced admission that the mechanism that built the company is now working against it.
The mechanism
Pirsch's growth chain runs: a founder's own, provable frustration with Google Analytics produced technical writing developers trusted because it wasn't selling anything — that credibility, not luck, is what a two-stage Product Hunt launch and years of transparent annual recaps had to work with, letting a two-person team compound slowly to $14K MRR entirely without paid ads. But every link in that chain was replicable by design (open pain point, open library, open numbers), and the same absence of a moat that let Pirsch enter a crowded privacy-analytics market cheaply is what the founders say let AI-assisted copycats enter behind them just as cheaply — which is why the growth rate, not the growth itself, is what broke.
Personal, provable pain with a well-known incumbent gave the founder's technical writing instant credibility with developers
Marvin didn't set out to build an analytics company — he set out to measure his own blog without Google Analytics or a cookie banner, after noticing his developer-heavy audience used ad blockers that silently drop client-side trackers. Because the problem and the fix were both real and personally tested before being pitched to anyone, the resulting open-source library and blog post read as engineering, not marketing — exactly the tone that earns attention in developer channels that reject anything that smells like a sales pitch.
I was looking for an alternative to Google Analytics to track visitors on a website... I was wondering if I could build something that I can integrate into my Go applications.
A failed first product freed the team to build without minding competition — then a two-stage Product Hunt launch converted that groundwork into paying customers
Emvi's failure against Notion and Roam Research removed any illusion that a good product alone wins a crowded market, so when the team productized the tracking library they spent deliberate effort on distribution mechanics instead. The clearest example is their Product Hunt approach: rather than one launch, they ran an unpolished beta first specifically to accumulate an audience, then treated the full release as a second, separate marketing event they could mobilize that audience for — including asking their very first beta customer to hunt the full launch. That's why the full release (329 upvotes) beat the beta (75 upvotes) by more than 4x: the second launch wasn't cold traffic, it was warmed-up traffic from the first.
The beta launch has about 75 upvotes, compared to the much better 329 upvotes for the full release. The main difference between these two launches is that we could build an audience before ‘wasting’ the main launch.
Radical transparency and a privacy-regulation tailwind built trust for five years — but the same low barrier to entry that let Pirsch in eventually let everyone else in too
Annual recaps with exact MRR, subscriber, and cost figures gave a privacy-focused product something to point to as proof it practices what it preaches, while GDPR-style pressure kept sending Google Analytics refugees toward cookie-free alternatives generally. That combination sustained steady, unspectacular growth for years. But the founders' own 2025 recap names the flip side: a simple, easily explained dashboard is also simple to copy, and by their account AI-assisted development lowered the cost of cloning that dashboard enough that monthly growth fell by half (1.4% to 0.7%) even though the underlying regulatory tailwind hadn't gone anywhere — the moat was never the tailwind, it was being first, and “first” stopped mattering once copying got cheap.
As you can see, growth has slowed down. We experienced increased competition in 2025, and the market appears more fragmented. I would assume that AI has lowered the barrier to creating SaaS products that people don't know much about.
How it went
Personal pain becomes an open-source library
2020-06 → 2020-10While rebuilding his personal blog to track post performance, Marvin rejected Google Analytics and cookie banners and built a self-hosted, Go-based fingerprinting tracker, open-sourcing it as “Pirsch.” Testing it on his own site validated the technical approach before there was any product plan.
Pivot from a failing first product, launch out of beta
2020-10 → 2021-06When Emvi, the team's knowledge-base product, couldn't compete with Notion and Roam Research, the two-person team redirected all effort into productizing the Go tracking library. v1.0 shipped in March 2021 with payment collection built in, followed by 33 feature updates in six months and two backend rewrites — most consequentially a full migration from Postgres to ClickHouse for real-time querying. By June 2021 they had their first 10 paying customers.
Two-launch Product Hunt strategy plus content marketing compounds to 100 customers
2021-06 → 2022-05Rather than treat Product Hunt as a single make-or-break event, the founders launched twice — an unpolished beta (75 upvotes) to build an audience, then mobilized that audience, including their very first beta customer, to hunt the full release about half a year later (329 upvotes). Combined with ongoing technical blog posts and a year-end transparency recap, MRR climbed past server costs by January 2022 ($542 MRR) and crossed 100 paying customers by May 2022 ($1,418 MRR).
One founder goes full-time, growth peaks, then the copycat wall
2024-01 → 2025-12Marvin quit his agency job to work on Pirsch full-time at the start of 2024, while Daniel moved to an external engineering role and contributed part-time; MRR grew 83% that year to $11K. By 2025, growth had slowed to a 24% annual pace (0.7% monthly) as competitors — helped, in Marvin's own telling, by AI lowering the barrier to building a copycat SaaS — crowded the same niche, and enforcing euro-only payments to cut currency losses cost them a few customers.
Milestones
- 2020-06Marvin publishes ‘Server-Side Tracking Without Cookies In Go’; creates open-source Pirsch Go library
- 2020-10Emvi stalls against Notion/Roam Research; team pivots to productize Pirsch
- 2021-03v1.0 released out of beta; payment collection begins; 33 updates shipped in ~6 months
- 2021-0676First blog introduction post: $76 MRR, 10 paying customers, content marketing only, no paid ads
- 2022-01542IH update: $542 MRR, 49 subscribers, 10 months after v1.0; MRR now exceeds server costs
- 2022-051418IH update: 100 paying customers, $1,418 MRR
- 2024-016000Marvin goes full-time; year starts at ~$6K MRR, 432 subscribers
- 2024-12110002024 recap: $6K→$11K MRR (+83%), 432→553 subscribers, 2 new enterprise customers
- 2025-12143002025 recap: $11K→$14.3K MRR (+24%), 553→627 subscribers, growth slows to 0.7%/month
Whether it fits you
Pirsch's loop depends on structural conditions specific to its niche and the founders' patience. Run it if these hold for you; the costs are real and worth naming before you copy the approach.
What it needs
You need a personally-lived, provable problem with a resented incumbent
Pirsch's credibility came from Marvin testing the fix on his own website before writing about it. Writing ‘alternative to X’ content without having actually hit X's specific pain point produces generic listicle content, not the kind developers trust enough to adopt a brand-new open-source library from.
You can sustain single-digit-thousands MRR for years without outside funding
It took roughly 10 months post-launch to clear server costs ($542 MRR, Jan 2022) and about 3.5 more years to reach $14.3K MRR (Dec 2025). One founder worked a day job for over 3 years before going full-time; this path requires either a day job, savings, or a second income for a long runway.
Run it if your core mechanism is simple enough to explain and defend transparently
Open-sourcing the tracking library and publishing exact MRR and subscriber numbers every year built trust specifically because Pirsch competes on privacy, where ‘verify what we say’ matters. That only works if what you're verifying is simple enough to be legible in a blog post — complex enterprise products don't get the same credibility lift from a public dashboard.
What it costs
There is no real moat once your category gets crowded
A simple, well-explained dashboard is also a simple target to copy. By 2025 the founders were competing against ‘copycats’ they say were made cheaper to build by AI, and growth cut in half (1.4% to 0.7% monthly) with no lever to pull that wasn't already being pulled.
Expect a permanent plateau, not a temporary one
Marvin Blum's own description — ‘too rich to die and too poor to live’ — is not a phase before a breakout; five years in, it's the steady state. Anyone expecting this playbook to eventually produce venture-scale growth is expecting something the mechanism was never built to deliver.
Enterprise upside is unreliable and can vanish over compliance, not price
A prospective client generating 150 million pageviews/month walked away in 2021 after compliance concerns surfaced in testing — after the team had already done a week of 12-14 hour days rebuilding infrastructure to serve them. Enterprise deals in a privacy-regulated category can fail for reasons product improvements can't fix.
The numbers we could verify
- monthly growth 2024
- 1.4% (MRR $6K→$11K over 12 months) — S7
- monthly growth 2025
- 0.7% (MRR $11K→$14.3K over 12 months) — S8
- conversion rate 2021
- 1.2%–1.4% website-to-signup, ~29k visitors — S3
- server costs 2021
- 45€/month (early 2021) → 290€/month (end of 2021) after ClickHouse migration — S3
- currency loss pre 2025
- ~10% of payments lost to USD/EUR conversion before euro enforcement — S8
Channels it actually used
- Content marketing (technical blog)
- Product Hunt
- Privacy regulation tailwind (GDPR)
- Build in public / transparent annual recaps
- Hacker News / developer community
- Word of mouth / referrals
- Open-source community (GitHub)
Our read
The name ‘Pirsch’ — German for a quiet, stalking hunt — describes the growth strategy better than any tactic list: content marketing and a two-stage Product Hunt launch are both slow, patient approaches that avoid one big noisy bet. The irony is that the same quiet, easy-to-replicate approach that let two developers sneak into a crowded privacy-analytics market is exactly what let dozens of AI-assisted copycats sneak in behind them by 2025.
The founders' own admission that Product Hunt is ‘just one of many marketing steps, not the key moment’ is a useful correction to the common survivorship-bias story where a single launch ‘made’ a company. Pirsch's real growth curve is a compounding of small, unglamorous levers — content, PH, transparency, referrals — sustained over five years, not a single spike.
‘Too rich to die and too poor to live’ is a more honest data point than most teardowns get: this is what durable, patient bootstrapping looks like when the mechanism itself carries no exclusivity. It produces a stable, modest business, not a breakout — and the founders say so themselves rather than dressing it up.
Sources
- Introduction to Pirsch Analytics2021-06-04
- Server-Side Tracking Without Cookies In Go2020-06-22
- Our Recap of 20212021-12-21
- Our Experience Launching the Same Project on Product Hunt Twice2022-02-17
- We passed $500 MRR!2022-01-26
- We now have 100 paying customers!2022-05-18
- 2024 Recap — Being In the Weird State of a Mid-Sized SaaS2024-12-21
- Our Recap of 20252025-12-30
- Pirsch vs. Google Analytics: How Marvin Blum Built a Privacy-First, Developer-Friendly Alternative (Website Planet interview)2025-07-03
- Introduction to Pirsch Analytics (privacy-friendly web analytics)2021-06-17
