How Gamma Grew 8× from ChatGPT Referrals: A Breakdown (Podcast Analysis)

What this is (and isn't). This article is an independent analysis of public statements Ravish Agrawal of Gamma made on the Distribution Podcast. It is not a customer success story. Gamma is not an Asva AI customer, and Asva played no part in Gamma's growth. Every figure below is attributed to the guest or host as spoken on the recording. Several are unverified claims flagged inline; treat them as what was said on a podcast, not as established fact.

TL;DR

On the Distribution Podcast, Gamma's Ravish Agrawal described spotting a fast-growing stream of ChatGPT-referred traffic in GA4 — reportedly around 8× growth — even while that channel's click-through share stayed under 1%. His response was less a hack and more a discipline: measure the channel like Facebook rather than Google, add AI engines to "How did you hear about us?", and lean into Reddit specifically because ChatGPT weights it heavily. Below we break down what he said, then translate it into five AEO moves any brand can run — with Asva tooling flagged only as optional instrumentation, never as the cause of Gamma's results.

The context

Gamma is an AI-native presentation and document tool with broad prosumer reach — individuals and small teams who discover software themselves rather than through procurement. The host framed Gamma as having reached 100M users and $100M ARR, profitably (host claims — VERIFY BEFORE PUBLISHING; confirm wording, currency, date, and profitability against a Gamma primary source). We cite these only to set scale; the analysis does not depend on them.

The relevant point for AEO practitioners is the audience type. Agrawal estimated that north of 90% of Gamma's AI/LLM-driven acquisition came from ChatGPT (guest estimate — attribute; confirm the exact metric definition and date range). For a prosumer product, that concentration makes ChatGPT the channel worth understanding first.

What they did

Agrawal's account, as told on the podcast, breaks into a few moves:

1. They noticed the channel in GA4. ChatGPT-referred traffic reportedly grew roughly 8× over the period (guest claim — attribute to Ravish Agrawal/Gamma; verify the date range and baseline). The growth showed up as a referral source, not as a line item anyone had planned for.

2. They resisted over-reading the click number. Despite the growth, click-through from ChatGPT stayed under 1% (guest claim; the denominator was left unclear on the recording — do not rewrite this as "% of users" or "% of revenue"). Rather than dismiss the channel as small, they reframed how to measure it.

3. They measured it like Facebook, not Google. Agrawal's organizing line: "As a channel, it behaves like Google. But as measurement, it behaves like Facebook." His point — the influence is largely view-through. People read an AI answer that mentions Gamma, then arrive later via a branded search or a direct visit, so last-click attribution undercounts it.

4. They added AI engines to "How did you hear about us?" To catch that view-through, the team put ChatGPT and Perplexity as explicit options in the self-reported attribution survey — a low-tech way to see influence that GA4's last-click view misses.

5. They leaned into Reddit — for ChatGPT specifically. Agrawal singled out Reddit as unusually influential for ChatGPT, not for every engine (he tied this to Reddit's data relationship with OpenAI). He described new Reddit content taking 30–45 days to surface as a ChatGPT citation (observed range, not a guarantee), while Perplexity picked up a new phrase as fast as the next day and, per a host anecdote, Google's AI Overview cited a niche Reddit query in one day.

He also offered a working split for AEO effort — roughly 80/20, agent/crawler-first versus human-first (present as one team's heuristic, not a benchmark) — and noted that citation selection didn't track view counts: he cited a case where a 400K-view and an 18K-view YouTube video were treated comparably as sources (guest observation; this doesn't mean views never matter).

The numbers (attributed)

Every figure here is as stated on the podcast. None is independently verified.

• ~8× growth in ChatGPT-referred traffic — guest claim; verify range + baseline.
• Under 1% click-through share for that channel — guest claim; denominator unclear.
• North of 90% of AI/LLM acquisition from ChatGPT — guest estimate.
• 30–45 days for new Reddit content to appear as a ChatGPT citation — guest observation.
• Next day Perplexity pickup of a new phrase; one day Google AI Overview citation of a Reddit query — anecdotal.
• 80/20 agent-first vs human-first effort — guest heuristic.
• 400K vs 18K YouTube views cited comparably — guest observation.
• 100M users / $100M ARR, profitable — host claims; VERIFY BEFORE PUBLISHING.
• Only 30–40% of the playbook overlaps across the four major AI engines — practitioner estimate.

Two other figures came up that we deliberately exclude from any headline claim: a recalled third-party stat that Claude cites "~5× less" than ChatGPT, and a hypothetical "~3-month citation half-life." Both are unverified/hypothetical on the recording and should never be presented as data.

5 lessons any brand can copy

• AI referrals hide in view-through. A sub-1% click rate can still sit on top of real, growing influence. Judge the channel on assisted and self-reported impact, not last click alone.
• Instrument self-reported attribution. Adding "ChatGPT" and "Perplexity" to "How did you hear about us?" is the cheapest way to see influence your analytics platform buries.
• Know your audience's engine. Gamma's prosumer base skewed hard to ChatGPT. Yours may not. Identify which engine your buyers actually use before spreading effort thin.
• Match the source to the engine. Reddit reportedly moves ChatGPT; Perplexity leans on third-party authority; engines overlap only ~30–40%. One universal playbook underperforms.
• Treat citations as perishable. Because decay isn't well understood, AEO behaves like an always-on channel, not a one-time project. Publish, monitor, refresh.

How to run this play

Each lesson maps to a general AEO action first. Asva features appear only as optional tooling — helpful for doing the work at scale, not responsible for anyone's results.

• Lesson 1 — See the view-through channel. Action: segment AI-referrer traffic and watch branded-search and direct lift after AI mentions. Optional tooling: Asva's AI Traffic Decoder is built to separate and quantify AI-referred sessions.
• Lesson 2 — Capture self-reported source. Action: add AI engines to your signup survey and reconcile it against analytics monthly. Optional tooling: pair the survey with traffic data in the same decoder view.
• Lesson 3 — Confirm your engine mix. Action: check which engines actually mention and cite your brand before choosing where to invest. Optional tooling: Asva's Brand Visibility Tracker monitors mentions across ChatGPT, Perplexity, Gemini, Google AI Mode, Copilot, and more.
• Lesson 4 — Build engine-specific source coverage. Action: prioritize the source clusters each engine favors (e.g., community for ChatGPT, third-party authority for Perplexity). Optional tooling: start from an Answer Engine Optimization framework, then track which sources get cited.
• Lesson 5 — Keep it always-on. Action: re-audit citations on a cadence and refresh decaying content. Optional tooling: recurring visibility tracking surfaces when a citation drops.

Want a fast read on where you stand today? Run a free AI Visibility Report, explore the platform in the Asva app, or book 30 minutes to talk through an AEO plan.

FAQs

Is Gamma an Asva AI customer?

No. This is an independent analysis of public statements from the Distribution Podcast. Gamma is not an Asva customer, and Asva had no involvement in Gamma's growth.

Did Asva produce Gamma's 8× ChatGPT growth?

No. The ~8× figure is a claim Ravish Agrawal made on the podcast about Gamma's own results. Asva is not connected to it and has not verified it; the number is attributed to the guest and pending verification.

How can a channel grow 8× but convert under 1%?

On the recording, Agrawal framed AI referrals as largely view-through: people see a brand mentioned in an AI answer, then convert later via branded search or direct traffic. Last-click attribution undercounts that, which is why he measures it "like Facebook."

Why Reddit, and does it work for every AI engine?

Agrawal said Reddit is especially influential for ChatGPT specifically — he linked this to Reddit's data relationship with OpenAI — and reported roughly 30–45 days for new Reddit content to appear as a ChatGPT citation. He did not claim Reddit works uniformly across engines; other engines weight different sources.

Are the numbers in this article verified?

No. Every figure is attributed as spoken on the podcast and none is independently verified. Host claims such as 100M users and $100M ARR, plus recalled third-party stats, are flagged for verification and should not be treated as established fact.

Close

The most transferable idea here isn't the 8× — it's the reframe. Agrawal's team treated AI mentions as a view-through channel worth measuring on its own terms, then matched their content to the specific engine their audience used. That discipline is the AEO play any brand can copy, whatever the exact numbers turn out to be. If you want to see your own AI-referral picture instead of guessing at someone else's, start with a free AI Visibility Report.

Analysis by Viren Inaniyan. Source: public Distribution Podcast episode featuring Ravish Agrawal (Gamma). Figures pending independent fact verification before publishing.