The website and app behind a YC startup at $9.4M ARR

The challenge

What we did

1. 1Ship the website before the app

   The app was the eventual product, but the offer itself had not been tested on anyone. We designed and built the marketing website first, in Webflow, and shipped it in June 2024, one month after we started. That put a live signup flow in front of real women while the app was still in Figma. It also meant the argument on the page could change as fast as the offer did, without waiting on a release cycle. Everything we designed after June 2024 was informed by people who had already signed up rather than by an assumption about who would.

2. 2One design system across two surfaces

   Around 100 screens went through Figma, split roughly evenly between the website and the mobile app. Both ran on one medium-sized design system. Sharing it mattered here because the website and the app carry the same promise at different moments. The website has to make a stranger trust a group of women she has not met. The app has to keep her turning up for six weeks after she has already paid. Holding one set of components across both meant the product did not change character at the point where a visitor became a member.

3. 3Build the app, and the integrations under it

   We wrote the production code, not only the designs. The app was built in FlutterFlow, which produces the iOS and Android builds from one codebase, and shipped to both stores. RealRoots ran their own backend. What we built alongside it were the Firebase functions on Google Cloud that keep Airtable in step with the app and the website, so the operational data the team worked from matched what a member saw. Stripe and PayPal took the payments, which carry more weight here than usual: the money changes hands before the first meetup, so checkout is where the promise gets tested. The app holds 4.7 stars across 1,200 App Store ratings as of August 2026.

4. 4Through the batch, then hand it over

   RealRoots went through YC Summer 2025 on the product we had built. By Demo Day the company reported $9.4M in annual recurring revenue, and TechCrunch named it one of the nine most sought-after startups of the batch. We were still on the project through all of it, and worked for another three months after that. We stopped in December 2025, nineteen months after we started, when the company raised again and hired an in-house team. That is the outcome we want from this kind of engagement. We covered the design and the build for as long as RealRoots did not have its own team, and stepped back once it did.