| raw | raw/i-make-5m-year-giving-my-saas-away-for-free.txt |
|---|---|
| url |
TL;DR
Marie and her co-founder/partner Philip built Tally, “the Notion of forms,” to $5M ARR with a team of 10 — by going all-in on free. Forms are viral by nature (every shared form carries a “made with Tally” badge), so an extremely generous freemium tier (no signup, no credit card) is the entire growth engine: 1.8M users, ~16K paying (~2% conversion) pay for the rest. The 5x from $1M→$5M in 12 months came from LLMs (ChatGPT, then Claude) recommending Tally — an accident they then optimized for.
The Founder’s Story
- Marie (marketing background, Belgian) + Philip (full-stack engineer, building since age 11). Started “late,” past 30, after a failed travel-influencer marketplace (Hot Spots) died in COVID.
- Tally was born mid-2020 from their own pain: Typeform was beautiful but expensive; Google Forms ugly. They built “the Notion of forms” — simple, block-based, low-friction.
The Core Principle: Keep It Simple, Keep It Free
- Simplicity is the USP and the constraint — a team of 10 serving ~2M users only works if the product is self-served. Pricing is binary: free or $29.
- Freemium done properly = an actual free tier. No account needed to start building; no credit card; generous limits (unlimited forms + submissions free). You discover Tally by filling out a Tally form — the product is the acquisition channel.
- Say no to features. Marie refused PDF export for 5 years despite it being the #1 request (finally shipped as Tally→PDF). Decision filter: can we ship a simple version in 1–2 weeks? Can it be free? Does it fit “simplest form builder”? Did users ask?
The Distribution Flywheel
- Viral loop: forms are made to be shared → “made with Tally” badge → more users → more free value offered → more discovery. Compounds, hard to stop. Same mechanic as “made with Lovable.”
- The 5x — LLM recommendations. Early 2025 they noticed onboarding “where did you find us?” answers shifting to “ChatGPT.” Their years of Reddit/social listening + a dry, comprehensive help center (built so they’d “never answer a question twice”) were exactly what LLMs cite. They then learned basic SEO (headings, URLs, content gaps) to feed answer engines. Graphs now track new model releases — ChatGPT first, now Claude is the biggest source. See ai-search-optimization.
- Building in public (Indie Hackers, then Twitter/LinkedIn) for transparency and trust — but Marie credits the free product, not BIP, as the real growth driver.
Customer Feedback at Scale
- First 3 years: anyone could join a Slack channel; they answered every DM/email within the hour, logged feedback manually in Notion, and shipped requests next day — turning early users into promoters. Closed it after year 3 (overwhelming, all negative feedback, couldn’t build). Public upvote boards now.
Building the Team
- Hired too late (burned out first). First hire: a support person who was already answering questions in the Slack channel. Remote engineers failed; they moved to an in-house engineering team in Ghent + globally-distributed support for time zones.
Future / Moat vs. the “SaaS Apocalypse”
- Vibe coding helps them — more people building apps want a simple form to plug in. Defenses: low price (not worth cloning), MCP so you can build/query forms from inside Claude without visiting Tally, founder-led brand. “Everyone can build now, but not everyone can sell.”
Regrets
- Invest in SEO earlier; hire sooner; you can’t fully combine new parenthood with max-growth hours — “the more hours you put in, the faster you grow.”
Key Stats
- ARR: $5M ($1M→$5M in ~12 months) · targeting $7M then $10M
- Users: ~1.8M · Paying: ~16K (~2% conversion)
- Team: 10 · Early infra bill: <€1,000/mo (profitable from the start)
Related Frameworks & Playbooks
- product-led-growth — freemium viral loop, product-as-acquisition-channel
- ai-search-optimization — winning LLM recommendations (the 5x driver)
- micro-saas — single-feature focus, tiny team, say-no discipline
- seo-growth — the help-center content that LLMs cite