Skip to main content

How did Tally.so grow its traffic via ChatGPT? A Case Study in Generative Engine Optimization (GEO)

TL;DR: Tally.so captures 9.6% of its total traffic (approx. 125k visits) from ChatGPT by prioritizing high-intent comparison pages and community mentions. This article breaks down the 3-step GEO strategy they used to reach $4M ARR without ad spend and how you can replicate it using DECA.
The search landscape has shifted. Clicks are vanishing, but conversations remain. While many marketers tremble at the “Zero-Click” threat, the online form builder Tally.so quietly discovered a new goldmine: Generative Search. In May 2024, 9.6% of Tally.so’s total web traffic originated from a single source: ChatGPT. According to data analyzed by Foundation Inc., this amounts to approximately 125,000 visits. Even more remarkably, these weren’t just casual visitors—they were “high-intent users” with significantly higher conversion rates. This article reverse-engineers how Tally.so conquered the AI search market without spending a dime on ads and presents a GEO (Generative Engine Optimization) strategy you can apply to your brand.

1. How much traffic does Tally receive from ChatGPT?

Tally.so generates 9.6% of its referral traffic through ChatGPT recommendations as of May 2024. This success is not accidental but the result of a shift in how information is consumed. Tally’s public data proves that AI has become a new “primary traffic source.”
  • 9.6% Share: The portion of total web referral traffic coming from ChatGPT. While Google search declines, AI-driven traffic has exploded.
  • 4 million in Annual Recurring Revenue (ARR) with a team of just ~13 people. Co-founder Marie Martens attributes a significant part of this organic growth to their “Build in Public” strategy, which naturally feeds AI models.
  • 70% Dominance: When asked, “Recommend a free form builder,” ChatGPT mentions Tally.so with a probability of nearly 70%.
This is why we must focus on GEO (Generative Engine Optimization). Beyond ranking at the top, you must ensure AI “recommends” your brand.

2. Reverse Engineering Tally’s Secret: What Did They Do Differently?

Tally’s AI recommendation strategy centers on three factors: high-intent comparison pages, community-driven brand mentions, and ‘building in public’ transparency. By analyzing Tally.so’s content and activities, we identified these core patterns that made AI fall in love with them.

① Boldness in Comparison (High-Intent Comparison Pages)

Tally.so aggressively targeted high-intent keywords like “Jotform Alternatives” and “Typeform vs Tally.”
  • Why it works: AI users ask, “What is the best free form builder?” AI constructs its answer by synthesizing documents that analyze and compare multiple tools. Tally’s comparison pages provided the structured data necessary for these answers.

② The Power of Brand Mentions (Strategic Brand Mentions)

Tally.so drove constant mentions in communities like Reddit, Indie Hackers, and Twitter (X).
  • Why it works: LLMs (Large Language Models) learn from “real user reviews” in communities as high-trust information. When data accumulates saying “People recommend Tally,” AI learns to recommend Tally too.

③ Transparent Building in Public

Co-founder Marie Martens transparently shared revenue, roadmaps, and even failure experiences on social media.
  • Why it works: Unique, specific data acts as a Citation Magnet—the best source for AI to quote. Because Tally provided information only they could offer, AI cited them as an authoritative source.

3. From SEO to GEO: The Shift You Must Prepare For

Tally’s case is the signal that the era of Search Engine Optimization (SEO) is setting, and the era of Generative Engine Optimization (GEO) has arrived.
FeatureTraditional SEONext-Gen GEO
GoalRank on Page 1Be Cited & Recommended by AI
Core ActivityKeyword Repetition, BacklinksAuthoritative Info, Brand Mentions, Structured Data
Success MetricClick-Through Rate (CTR), VisitsShare of Citation, Brand Recommendations
You must no longer write “content that search engines like to read,” but “content that AI likes to learn and cite.”

4. Next Step: Your Brand Can Be the Next Tally

You can achieve what Tally built manually over time much faster and more efficiently. DECA is a GEO solution that systematizes Tally’s success formula.
  • Discover: Diagnose how your brand is currently perceived by AI.
  • Optimize: Optimize content into a structure that AI finds easy to cite.
  • Measure: Visualize the previously invisible AI search market share.
Note on Constraints: While GEO offers immense opportunity, tracking remains a challenge compared to traditional ads. Solutions like DECA are essential to bridge this gap by measuring “Share of Citation.” Start now. The AI search market is a winner-takes-all game for early adopters.

❓ Frequently Asked Questions (FAQ)

Q1. Is GEO only effective for B2B companies? No. GEO is essential in any field where users seek “recommendations” or “comparisons” (Travel, Electronics, Finance, etc.). However, the effect may be more immediate in B2B SaaS, where complex decision-making is involved. Q2. Does this apply to other AI search engines like Perplexity or Gemini? Yes. Most LLM-based search engines share a common preference for “authoritative sources,” “user reviews,” and “structured data.” Q3. It’s hard for us to share data transparently like Tally. What should we do? You don’t have to reveal everything. You can build authority through “proprietary insights,” “customer success stories,” or “industry reports” that only your brand can provide. Q4. How do we measure GEO performance? Standard GA4 (Google Analytics) cannot fully track AI search traffic. You need specialized solutions like DECA to monitor new metrics such as “Share of Citation.” Q5. Should we discard all existing SEO content? No. Simply reformatting high-quality existing content into an “AI-friendly structure” can yield significant results. This is the most efficient way to start GEO.

📚 References


The “Context Lock” Protocol for AI Citation
Written by Maddie Choi at DECA, a content platform focused on AI visibility.