On the podcast: the month they went from $1M to $12M ARR, selling Brazil on career upgrades instead of English lessons, and why you shouldn’t tell influencers how to do their job.

Top Takeaways:

🔍 Filter creators by reach, not following
Under 5,000 followers but tens of millions of views means the algorithm has already picked them and nobody else has priced them yet.

🙅 Don't tell influencers how to do their job
The creator video the CEO didn't like did 30M views across three platforms and took ARR from $1M to $12M in a single month.

🧱 Keep the creator program and paid acquisition as separate streams
Creators keep generating angles the in-house team has run out of, and paid reproduces whichever video goes viral.

🎓 Staff the program with ex-creators, not managers
The influencer lead and the country coaches were all making videos first, and each coach now rotates a roster of 20 or 30 creators.

💼 Sell the outcome, not the activity
A $100-a-year subscription sold in a market with weak purchasing power because the claim was a better job, not better English.

🚧 Growth can be unhinged; the product can't
Cursing avatars went viral upper-funnel, then users rejected them inside the app, so product rules now live on a one-pager whilst creators keep almost unlimited freedom.

About Adam Turaev:
🗣️ Co-founder of Praktika, an immersive language learning app with generative AI Avatars, empowering the next billion learners to overcome language barriers and seize new opportunities.

👋 Linkedin

💬Adam Turaev on X

🖥️Learn more about Praktika

Follow us on X:

• David Barnard - @drbarnard
• Jacob Eiting - @jeiting
• RevenueCat - @RevenueCat
• SubClub - @SubClubHQ

Episode Highlights:
[00:00] Introducing Adam Turaev of Praktika

[01:26] Why Adam Left Enterprise Chatbots for Language Learning

[04:22] Betting on Conversational AI Before the LLM Boom

[06:45] How Early Chatbot Experience Shaped Praktika

[08:53] Competing With Duolingo by Targeting Career Growth

[10:03] Why Brazil Became Praktika’s First Target Market

[12:23] Selling Better Jobs Instead of English Lessons

[14:07] Building AI Language Tutors Before the Technology Was Ready

[18:41] Will Real-Time AI Translation Make Language Learning Obsolete?

[21:07] Why Speaking a Language Is About More Than Translation

[23:09] What Separates an AI Coach From ChatGPT Voice

[26:10] Why the Job to Be Done Should Drive Product Decisions

[29:16] How AI Could Help Human Tutors Reach More Students

[30:51] Why Praktika Uses Animated Avatars to Reduce Speaking Anxiety

[34:19] Finding Micro-Influencers With High Reach and Low Followings

[37:15] The Correction Video That Took Praktika From $1M to $12M ARR

[40:19] Replicating a Viral Format Across Countries and Creators

[41:49] Hiring Gustavo and Giving Creators Freedom

[43:58] Building an Influencer Program With Over 100 Creators

[46:41] How Praktika Pays Creators and Turns Wins Into Paid Ads

[48:47] When Unhinged AI Avatars Backfired Inside the App

[50:36] Setting Different Creative Boundaries for Growth and Product

[52:25] Reaching $50M ARR With a Team of Just Over 40

[53:14] Building Proactive AI Coaches Beyond Language Learning

[55:13] Balancing AI Inference Costs With Product Quality

[58:04] Why Switching From Flutter to Native iOS Improved Conversion

[01:00:03] Praktika’s Biggest Failed Experiment With AI Avatars

[01:00:35] Why Copycat Competitors Make Growth Harder

On the podcast: the “magic moment” test for whether creators can sell your app, the one number that caps what you can pay them, and why cloning hit apps like Cal AI rarely pays off.

Top Takeaways:

🥊 Choose creator niches by what the viewer wants, not what the creator does
UFC fighters track every calorie, but their audience is on the couch with pizza, and thousands of dollars of posts produced no download spike.

✨ Creator content only sells an app that has a magic moment
A scan-analyze-result sequence made viewers do a double take; without that, the app’s appearance in a creator’s video reads as an ad and never goes viral.

🧮 Never pay a CPM above your revenue per thousand views, and go flat rate
Break-even is set by a creator's average views, and the profit sits in the one post that goes mega viral, which a flat fee lets you keep.

🏷️ Keep the price low on purpose when the goal is to be the app everyone has
$30 a year, among the cheapest in the category, was a deliberate choice to be the easiest app to buy rather than the most profitable per user.

🧬 Sharing the playbook publicly is safe when the edge is a thousand tested details
The clones that copied a public growth story were never going to catch up with a team that was obsessed with small compounding improvements.

About Jake Castillo:
🚀Co-founder of Cal AI, the AI-powered app for easy calorie tracking. Snap a photo, scan a barcode, or describe your meal and get instant calorie and nutrient info

👋 Linkedin

💬 Jake Castillo on X

Follow us on X:

• David Barnard - @drbarnard
• Jacob Eiting - @jeiting
• RevenueCat - @RevenueCat
• SubClub - @SubClubHQ

Episode Highlights:
[00:00] Why the truth is always in the comments

[00:29] How Jake met Zach and joined Cal AI

[06:11] The creator partnership that changed Cal AI's trajectory

[09:45] Why fitness creators were the perfect channel

[13:57] Cal AI's three-step "magic moment"

[17:40] Why Cal AI wasn't afraid of copycats

[19:53] The UFC and mukbang creator bets that flopped

[25:21] How to spot creators who will convert

[28:00] Finding creators before they blow up

[30:12] Pricing creator deals around revenue per 1,000 views

[35:30] Measuring creator performance when attribution gets murky

[39:04] Building the systems to manage hundreds of creators

[43:09] The real secret to signing creators at scale

[46:50] When and how Cal AI added paid ads

[51:59] Keeping the experience consistent from ad to onboarding

[55:07] Why Cal AI kept its price at $30 a year

[58:47] The MyFitnessPal acquisition

[01:03:36] Lightning round: affiliate deals and doing the work

On the podcast: growing for four years on press and organic alone, affiliate deals as an underexplored revenue stream, and why deciding to pivot is the hardest part.

Top Takeaways:
🗞️ A PR retainer can beat paid UA for a mass-market app
A $3–5K monthly retainer produced months with half a million downloads and four years of growth with zero Meta or Google spend.

🤝 Affiliate deals are the revenue stream most subscription apps still ignore
Brands pay commission only on sales, so the pitch to Nike-sized partners is low-risk for them and the deals are reached directly, not invented.

📏 Pick the metric that is the user's outcome and A/B test every feature against it
A walking app that refuses to track time in app and measures steps instead shipped a feature that lifted steps 10% by locking social media behind a step goal.

🌍 A growth playbook that works in one country can fail completely in the next
The PR strategy that pulled 300,000 downloads from a single French TV segment produced nothing in the US, where nobody knew the founder or the app.

⭐ A celebrity works harder as a shareholder than as a spokesperson
An ambassador-plus-investor deal changed hiring, press replies, CAC and retention at once, though the impact can be difficult to precisely quantify.

⏱️ At scale, app review times are a growth constraint, not an inconvenience
The more parallel A/B tests you run, the more a 24-hour (or five-day) release cycle caps how fast you can learn.

About Yves Benchimol:
🚶 Founder of WeWard, a free mobile app backed by Venus Williams designed to make walking a more rewarding part of people's lives. WeWard’s mission is simple: to get more people walking the world over.

👋 LinkedIn

🚀WeWard

🖥️ WeWard Careers

💬WeWard on X

Follow us on X:

• David Barnard - @drbarnard
• Jacob Eiting - @jeiting
• RevenueCat - @RevenueCat
• SubClub - @SubClubHQ

Episode Highlights:
[00:00] Meet Yves Benchimol of WeWard
[01:36] Why Yves left B2B retail analytics
[05:04] Shutting down a business to start over
[07:31] Building an app that rewards walking
[09:29] Why WeWard ditched blockchain
[12:24] Choosing user value over technical complexity
[15:55] Loyalty, gamification, and motivational layers
[17:41] Why WeWard measures steps, not screen time
[18:37] Building revenue beyond subscriptions
[21:46] How affiliate marketing fits WeWard
[23:54] Using PR as the first acquisition channel
[25:18] 300,000 downloads in under five minutes
[28:32] Four years without paid UA, and why US press fell flat
[31:12] Why US growth required local credibility
[33:11] Pitching Venus Williams at lunch
[37:47] What Venus Williams brings as ambassador and investor
[42:01] What's next for WeWard's global growth
[43:14] Staying ahead of copycat apps
[46:25] Locking social apps until users walk
[48:27] Motivational layers and staying focused on walking
[50:11] Localizing rewards across 29 countries
[52:35] Why growth strategies change by market
[54:10] Moving to the US to scale WeWard
[57:13] Why big growth requires bigger bets
[1:00:06] Biggest win: rebuilding the onboarding
[1:00:46] Biggest fail: returning to B2B
[1:01:46] Why mobile app iteration is still too slow