How I Built It: $400K/Month Mobile App (Gravl)

Julian Gargicevich

• Is a developer from Argentina and one of the co-founders of Gravel.
• Grew up around fitness because his dad owned a fitness center.
• Spent much of his time outside school playing sports and going to the gym.
• Studied software engineering and moved to Australia.
• Worked at small startups, big tech companies such as Atlassian, TV channels, and an investment fund.

Gravel

• Is an AI fitness app that provides smart workouts for the gym.
• Launched around two years before the episode.
• Has over 70,000 subscribers.
• Made over $440,000 in the prior month.
• Focuses on strength training and gym workouts.

Origin Story

• Julian and his partners first worked on an influencer marketing platform for mobile games.
• The business gradually became more like a marketing agency, which was not what they wanted to build.
• The agency taught them about user acquisition, marketing, and app revenue numbers.
• Julian started with a workout tracker idea, then realized it did not add enough value compared with existing trackers.
• A friend showed him Fitbod, and Julian liked the UI and concept of workouts generated on the spot.
• After using Fitbod, he felt the workouts were weird and sometimes dangerous.
• Decided to build a product with Fitbod-style UI and UX plus a stronger workout engine.

Building The MVP

• Built the initial MVP in around two to three months.
• Started with a tracker-style MVP before shifting toward the Fitbod-inspired concept.
• Found the custom workout engine to be the hardest part.
• Had to account for equipment, weekly goals, gym frequency, consistency, gender, weight, age, and many other combinations.

Launch And Validation

• Used Reddit as the first major distribution channel.
• Posted a thread explaining how he built Gravel, which was then called Games AI.
• Shared technical details in the post.
• Got a couple hundred likes within the first few hours and over 300,000 impressions.
• Got the first couple thousand users from Reddit.
• Saw developers using the app because they liked tech and some were interested in the gym but found it intimidating.
• Received bug reports and feature requests that convinced the team they were onto something.
• Added a subscription model after the app had early traction.

Growth Strategy

• Used paid ads as the main distribution channel after adding subscriptions.
• Got a subscription within the first 10 minutes of activating the first ad.
• Translated the app into Spanish and started running ads in South America.
• Spent less than $50 a day on ads in the early days.
• Continued relying on paid ads as the channel that worked from the start through the episode.

Advice For Consumer App Builders

• Validates before spending money on ads.
• Treats validation as proof that people are willing to pay, not only proof that the product works.
• Looks outside the United States when ads are expensive and competitive.
• Suggests translating into other languages when founders have the ability to do so.
• Starts small with influencers, creators, AI videos, and CapCut-edited creatives.
• Emphasizes creating creative volume because making one ad is not the hard part.
• Says UGC is king for ads.
• Uses public ad libraries to study competitors and copy what is already working.

Product Experience

• Starts users with onboarding questions about name, goals, experience level, training frequency, workout split, excluded muscles, focus muscles, and where they train.
• Shows users that the app is generating a custom workout plan.
• Uses a hard paywall before users sign in for the first time.
• Centers the main app experience around the workout card and active workout flow.
• Includes exercise lists, videos, descriptions, and instructions.
• Adjusts weights based on exercise order.
• Tracks muscle recovery based on previous workouts and external activity from sources such as Apple Health and Strava.
• Offers 24/7 in-app support with articles, messages, and a real person replying instead of AI.

Tech Stack And Costs

• Uses React Native and Expo for the app.
• Uses .NET for most core backend functionality.
• Uses Next.js and React for internal admin dashboards.
• Uses Cursor while being specific about which files AI should touch.
• Spends about one third of revenue on ads across Meta, TikTok, Google, and Apple Search.
• Grew from three founders to a team of about 13 to 14 people, including part-time people and contractors.
• Spends around $50,000 to $80,000 on salaries.
• Pays Apple’s 15% revenue cut.
• Pays about 1% of revenue to RevenueCat.
• Pays about $1,000 a month for mobile measurement partners.
• Pays about $1,000 a month for servers, AI bills, and other app infrastructure tools.

Lessons And Advice

• Encourages founders to be proud of what they build.
• Says it is easier to work for zero dollars when the founder loves what they are doing.
• Warns that building is not even half the work because founders still need distribution, ads, feedback, and persistence.
• Says founders should keep going through failed ads and setbacks.
• Also says founders need to know when to quit a product that is not working.
• Believes his previous startup probably lasted an extra year longer than it should have.