How Gumlet Gets 20% of Inbound Revenue From ChatGPT Mentions

What does the case study claim?

The case study also reports 127 new top-ten Google rankings over the same period.

What did Gumlet do?

1. Defined how it should be described. The agency mapped a small set of statements to repeat consistently: a secure video hosting platform, an alternative to Vimeo, Wistia and Mux, used by SaaS companies, course creators and enterprises.
2. Rewrote key pages around buyers' questions. Pages such as video hosting, pricing and the AI subtitle generator were reworked for phrasing like “secure video hosting platform without Vimeo limits” or “how to prevent downloads in hosted videos”.
3. Published supporting guides matching questions seen in AI answers, such as the best video hosting for SaaS teams, Vimeo alternatives and protecting paid course videos. Its top guides included a best video hosting platforms list, a Vimeo cost page and a guide to preventing screen recording, all linked to the main product pages.
4. Measured self-reported discovery. Onboarding asked new users where they had heard of Gumlet, and those answers were joined to signups and upgrades in its analytics.

The case study says ChatGPT often mentioned Gumlet's pricing transparency, security and subtitle tools, the same points its pages stressed.

Why did cost and comparison pages matter?

Look at the guides that performed best: a list of video hosting platforms, a page on what Vimeo costs, and one asking whether Vimeo is worth it. These match the moment a buyer is weighing a well-known tool against alternatives, which is exactly when people ask an assistant for options. Answers to “Vimeo alternatives” or “best video hosting for courses” draw on pages that list options with prices and limits, and a page that explains a competitor's pricing fairly gives the answer something concrete to use. For a challenger brand, these pages do two jobs at once: they rank in Google for the same comparison searches, and they give AI answers a reason to include the challenger in the list.

How far can you trust the 20%?

Take the direction seriously and the precise number cautiously:

• It is marketing. The agency published it to win clients, and no independent audit is available.
• It relies on self-reporting. People remember where they first heard of a product imperfectly, and surveys can nudge answers.
• Three assistants are counted together: ChatGPT, Perplexity and Bing Copilot.
• Definitions are loose. “Inbound revenue” and the time window are not specified, and the smaller May to July figures, seven paid upgrades, sit awkwardly beside a fifth of revenue.
• Google gains happened at the same time. Better rankings may have helped convert AI-aware users, so the effects cannot be separated.

The finding that holds up best is qualitative: people who hear about a product from an AI assistant often go to Google or type the address, so referral data alone undercounts AI's influence. How such referrals behave is discussed in whether ChatGPT referrals convert better.

Why do AI-discovered buyers arrive through Google?

The case study describes a path that other companies report too: an AI answer names the product, the buyer opens Google and searches the brand, then visits and signs up. Some visits also lose their referrer through privacy-focused browsers or app handoffs, so they appear as direct. In analytics, that influence shows up as branded search growth and direct traffic, not as chatgpt.com referrals.

How can you measure the same thing?

1. Ask how people heard of you at signup or in the first conversation, with explicit options: ChatGPT, Perplexity, Gemini, Copilot, Google's AI Overview, Google search, a friend, and other.
2. Store the answer with the account in your CRM or product analytics, so you can follow revenue, not only signups.
3. Track branded search impressions in Search Console alongside your AI visibility tests; rising branded demand while AI mentions grow is supporting evidence.
4. Keep referral tracking too. Tag and report chatgpt.com, perplexity.ai and copilot visits, as covered in how to track ChatGPT referrals in GA4.
5. Compare cohorts by activation, conversion and revenue over several months before quoting a share.

For lead-based businesses, forms and call tracking can capture the same signal, as covered in how to track leads from AI answers.

What can other SaaS companies copy?

• A short, consistent description of what you are, who you are for, and what you replace.
• Fair comparison and alternative pages for the tools buyers already know.
• Pages that answer specific buying questions, with prices and limits visible.
• Guides linked to product pages, so answers that cite a guide lead to the page that converts.
• A discovery question in onboarding, which costs almost nothing and gives you evidence.

Be sceptical of claims that content “teaches” AI models. Pages mostly help by being found and cited when assistants search; training takes far longer and cannot be steered page by page.

We have no independent data on Gumlet's results. This page is based on the published case study and our reading of its figures.

What might this look like in practice?

Consider an illustrative scheduling tool for clinics, Ruddby, that added a “Where did you first hear about us?” question to signup with AI assistants as options. Over three months, 14% of new trials picked ChatGPT or Perplexity, though chatgpt.com referrals showed only 3%. Those trials converted to paid at a higher rate than average. The team rewrote its pricing and “Ruddby vs” pages, tracked branded searches in Search Console, and reported AI-influenced revenue as a range, noting that answers were self-reported.

What do people ask about the Gumlet case study?

Who published the 20% figure?

Derivate X, the agency that worked with Gumlet, in an October 2025 case study.

Did the revenue come from clicks in ChatGPT?

Mostly not. Most AI-aware users arrived through Google search or direct visits.

Is self-reported attribution reliable?

It is imperfect but useful, especially combined with branded search trends and referral data.

Can a small SaaS company do the same?

The measurement, yes. The results depend on your category and how clearly your product answers buyers' questions.