The pivot that paid off: How fal found explosive growth in generative media

Key takeaways

1. Yurtseven and his co-founder ran the old data product and the new inference product side by side for two to three months, saw inference revenue growing faster, then dropped the old customers. He says running both made selling hard because the website showed something else.
2. Founders can talk themselves into believing a pivot is 'not that drastic.' Yurtseven says they lost time doing exactly that.
3. When investors doubted revenue durability, fal built a sales team early and pushed customers toward yearly commitments. Contracts often came from self-serve users who passed a daily spend threshold, which flagged them for a sales call.
4. Owning a category name helped. Yurtseven says calling fal a 'generative media platform' meant talking about the industry was indirectly talking about fal.
5. Revenue is the main metric he tracks. Request counts stopped meaning much once video models arrived, since one video request is worth far more than one image request.