Marie and Filip sent direct messages to people they found on Twitter, Product Hunt, and Indie Hackers — specifically targeting builders, founders, and designers who they thought might be interested in a form builder. Because Tally was free with no account required to start, the barrier to try it was essentially zero, which made it easy to get responses. They explicitly asked for feedback on the product and invited anyone who replied to join a private Slack channel. Once users were in the Slack channel, Marie would personally DM them asking what they thought and whether they'd tried specific features. When users requested a feature, the team would ship it — sometimes the very next day — and then personally notify that user. This "we actually listened and acted" experience turned early users into active promoters who shared Tally with colleagues and friends. This hands-on feedback loop via Slack was maintained for the first ~3 years of the product.
Tally
Simple, free-first form builder with Notion-like UX for anyone
9 moves, in order
- Pre launch / ValidationCold dm twitter producthunt indiehackers
Scraped Twitter, Product Hunt, and Indie Hackers for potential users (builders, founders, designers). Sent cold DMs asking for feedback on the idea and early product. Invited anyone who responded into a private Slack channel for ongoing conversation.
Built first cohort of engaged early users who became promoters - Early Growth (Year 1–2)Product led growth viral badge
Offered a genuinely generous free tier — unlimited forms and submissions, no credit card, no account needed to start building. Every published form shows a 'Made with Tally' badge, exposing the product to every form respondent.
Product became its own primary acquisition channel; ~2% free-to-paid conversion rate sustained economics - Early Growth (Year 1–3)Slack community feedback loop
Maintained an open Slack channel for all users. Personally answered every message within the hour, tracked all feedback in a Notion database, shipped requested features rapidly, and notified the specific user when their request went live.
Turned early users into vocal promoters; shaped the entire roadmap for 3 years; community grew to 3,000–4,000 members - Early Growth (Year 1–3)Building in public indiehackers blog
Documented and published milestone posts on Indie Hackers (e.g., revenue milestones, what changed from 4M to 5M ARR). Later cross-posted to Twitter and LinkedIn. Framed as transparency and giving back, not pure marketing.
Grew an audience of indie founders; reinforced 'underdog bootstrapped' brand differentiation; drove some new user discovery - Ongoing / Years 1–4Reddit social listening
Actively monitored Reddit and other forums for relevant conversations about forms, and participated or pitched Tally in those threads over multiple years.
Built a persistent Reddit presence that later contributed to LLM training data recommendations - Year 3+Public voting board user feedback
Launched a public feature upvoting board. Accumulated years of data on user requests which became the primary tool for roadmap prioritization, replacing the manual Notion tracking system.
Systematic feature prioritization; PDF export became most-requested feature (shipped after 5 years) - Early 2025 (breakthrough phase)Llm ai search chatgpt claude
Noticed in March 2025 that users were reporting finding Tally via ChatGPT. Added 'AI' as an option in their onboarding attribution question, then added a follow-up asking which AI tool and what prompt was used. Confirmed massive influx from ChatGPT, later Claude. Attributed it to years of Reddit activity + large help-center content base that got indexed by LLMs.
Revenue went from ~$1M ARR to $5M ARR in roughly 12 months; LLM referrals became the dominant new acquisition channelMRR $417k Users 1.8M users - 2025 (post discovery)Seo llm optimization help center
Retroactively cleaned up the large help center (which had been written purely for users, not search) by fixing headings, URLs, and basic on-page SEO. Identified content gaps and wrote targeted simple pages to capture LLM recommendations. Did not try to compete with large content teams (e.g., Jotform's 100-person content team) — focused on niche gaps only.
Accelerated LLM referral growth; Claude traffic grew alongside ChatGPT traffic following each new LLM release - 2025 (future proofing)Mcp llm integration
Built an MCP (Model Context Protocol) integration so users can build and manage Tally forms directly inside Claude and other AI assistants without visiting the Tally interface. Goal is frictionless access via AI agents.
Validated by interviewer who confirmed seamless Tally+OpenCrow form creation experience
Complementary co-founder pairing: Filip is a full-stack engineer who has been coding since age 11 with multiple prior startups; Marie has a marketing background from working inside a 100-engineer digital product studio. Together they could build, design, and market without hiring. They also had prior failed startup experience (HotSpots marketplace) that taught them lean validation and direct outreach.
LLM referrals (ChatGPT and Claude recommending Tally) combined with years of Reddit presence and a large help-center content base that got picked up by AI training data
Paid ads — explicitly tried and said they "didn't really work for us." SEO was also neglected for years, which Marie explicitly called a mistake. Remote engineering team via Upwork also failed and was abandoned in favor of in-house hiring.
I Make $5M/Year Giving My SaaS Away for Free
Barn2 Plugins
Portfolio of 19 WordPress plugins and 1 Shopify app for WooCommerce stores
Data Fetcher
Airtable extension that connects and auto-syncs external APIs into your database
Sync2Sheets
Google Sheets add-on that syncs Notion databases into spreadsheets