ElevenLabs · AI voice lab
Hire engineers who build side projects, judged by their public GitHub
Unknownteam-process
What they did
Screen engineers for side projects and user-pain obsession by looking at public GitHub profiles; target people from companies with strong product-building cultures; recruit many former founders; referrals from those hires form a flywheel.
- Stage
- growth
“We take a good, hard look at public GitHub profiles to see who's engaged in what.”
“And then on the engineering side, we look and we screen for those that build side projects on their own, that can identify a user pain point and really build and be obsessed with that.”
Related topics
Get tactics like this every Monday
New tactics from the week's founder interviews, each linked to where it was said.
More from ElevenLabs
All 16 ElevenLabs tactics- Drop the translation agency; have an LLM translate app text on every code changeMuch cheaper and faster than the agency and paid tool it replaced
- Test ad formats with agencies, then hire motion designers to make variants in-houseMotion-design ads are the bulk of ads and work really well; no numbers
- Treat each ad channel separately and grow its spend 10-20% a weekGoogle search is still the biggest channel; no spend or return numbers
- Translate ads into other languages with AI dubbing and prompted copyGuest says it gets fantastic results; no numbers
- Feed your best and worst ad copy into a custom GPT that writes new adsNewly built; guest calls it really cool, no results yet
- Run holdout tests in each ad platform and discount reported sales by the resultsExample: 500 reported conversions were only 200 incremental
More from this episode
- ElevenReader: Sell hours of listening instead of AI credits, and drop a planCalled the app's biggest win of the year; no numbers shared
- ElevenLabs app: Offer one price in the mobile app and test offers around itKeeps pricing simple and the team accountable; no numbers shared
- ElevenReader: Keep tuning the onboarding quiz without testing a no-onboarding versionBiggest fail of the year; too many cycles spent on onboarding experiments
- ElevenReader: Cut onboarding down so people upload a book and pick a voice fastWorked better than quizzes and goal setting; no numbers shared
- ElevenLabs: Launch about once a month to earn press, then buy the brand searches it createsBelieved to lower CPA/CPI and compound across launches; never measured
- ElevenLabs: Write the launch Twitter thread first, then turn it into a video
- ElevenLabs app: Test new ads without needing week-one payback, reviewing each channel weekly
- ElevenLabs: Hire a creator on staff to film short videos for launches
- ElevenLabs app: Share UGC creators, tools and tracking setup across sibling apps
- ElevenLabs: Build separate apps for separate uses, each with its own brandWorked twice so far, per the speaker
- ElevenLabs app: Test paywall ideas in two sibling apps at once and swap notes
- ElevenLabs: Run the company as 10-12 small teams that each own design, build and growthSpeakers credit it for shipping speed; no numbers
- ElevenLabs: Tell users how many hours their voice agents have run, so they share it
- ElevenReader: Watch Reddit, Discord and X for what people share, then push itUsers post screenshots of surprising voices; still iterating
- RevenueCat: Launch a mobile app with a home screen people want to screenshotAbout 100x more sightings on X overnight; mobile usage near dashboard usage
- Canva: Spin out a separate app for one use case of the main productWon Apple awards but failed on acquisition and retention; folded back in
- Canva: Plan a launch 12 months out, with 3 months on messaging
- Lower the price to keep customers who are churningFailed; maybe one extra month retained if they don't care about the product