Granola Crunched: Going Viral Through Customers

This case study was created from an episode of the Product Marketing Adventure Podcast. It has been adapted into Dusted's case study format to preserve and showcase the guest's real-world experience. Written to preserve the voice of the true author, Jack Cully.

Granola is the AI notepad for people in back-to-back meetings. It lives on your computer or phone, lets you take rough notes while it transcribes in the background, then turns those notes into readable summaries so you can stay focused on the conversation. We’re based in London with a team of about 50 that’s still growing. The big shift for us has been moving from a pure notes app toward one that makes the most of meeting context—we had just launched MCP so Granola can talk to ChatGPT, Claude, and other AI apps.

We already had strong product-market fit and genuine word-of-mouth. Users swore by it, feedback forms were full of praise, and people were talking about us on AI Twitter. But most of that love stayed behind closed doors. We wanted more visible, public conversation about Granola online without forcing anything gimmicky.

Every marketer dreams of their own Spotify Wrapped, but most recaps have become noise. We needed something that felt genuinely valuable rather than a shallow feature-count gimmick. Granola knows a surprising amount about users—the phrases they use, who they spend time with, decisions they make, projects they advance, achievements they celebrate. The opportunity was to turn that context into something people would actually want to share.

At the same time we had hard constraints. We couldn’t roast people (early tests felt mean and didn’t work the way they do on Spotify). We had to stay accurate, respectful, and safe around confidential meeting content—especially for people in HR, recruiting, or anyone discussing roadmaps or sensitive topics. Nothing could ship unless it made users feel truly seen, gave them an acceptable way to talk about themselves, and stayed positive enough that they’d proudly put their name on it.

It started as a fun end-of-internship project. Engineering interns Luke and Malika built a basic prototype using our Recipes feature (custom saved prompts) during the team’s week at Casa Granola in Majorca. Even with a rough prompt the results blew us away, so I forced “launch Granola Wrapped” into our Q4 plan.

We set clear criteria: every insight had to make someone feel seen, reveal something meaningful, and be something they’d actually want to share. We optimized for three sharing vectors—public posts (LinkedIn/X achievements or consultant testimonials), internal Slack/Teams (catchphrases, work bestie/partner-in-crime, company in-jokes), and friends (WhatsApp-worthy details that travel outside work).

Joe, our chief of staff, spent well over 24 hours iterating the AI prompts for accuracy, humor, empathy, and safety. We killed roasting entirely and landed on a celebratory-but-sharp tone. We tested relentlessly inside the company, then with ~10 external users across different backgrounds and riskier industries. Early versions that were just one or two slides failed because different people resonated with different moments, so we shipped eight or nine screens. James (design engineer) and I spent roughly six weeks end-to-end; the core build team grew to five people covering backend, frontend, AI, product design, and brand.

We launched at 8 am UK time so London users would seed conversation, New York would wake up to momentum, and San Francisco would amplify it. A banner inside the product (“Your 2025 Crunched”), a staggered email, and early organic social did the rest. I personally replied to every single post about Crunched—tagged or not—from morning until 8–9 pm to keep the energy personal and human.

We relied on our existing Recipes system (custom prompts) as the foundation for the prototype and final experience. Heavy AI prompt engineering by my colleague, Joe, was the single most time-intensive piece. James built the front-end experience quickly so we could iterate in days, not weeks. We added product instrumentation on the share button so we could see true sharing intent versus public posts, plus standard engineering readiness checks for scale. Share flows were optimized for LinkedIn, X, Slack, and easy image export. Privacy messaging was added at the last minute so users knew only they could see their data unless they chose to share it.

We hit millions of organic impressions in the first few days, with a long tail that still has people asking for their Crunched four months later. Sharing rates hit around 25%—far above the low-single-digit click-through rates typical for this kind of campaign. The week after launch delivered our highest acquisition to date.

Beyond the numbers, two bigger shifts happened. Granola’s name was suddenly everywhere in a positive, “best thing since Spotify Wrapped” way that positioned us as a real necessity rather than just another notes tool. And an unintended but powerful outcome: people started seeing Granola as more than meeting notes—as a product that actually understands them at work. The campaign gave users a socially acceptable mirror to talk about themselves, and they used it.