Databricks' Ali Ghodsi Never Wanted to Be CEO. Now He's Among the Best

Key takeaways

1. Ghodsi says companies usually fail on one giant bottleneck, so put orders of magnitude more attention on it and expect it to take one to three years, not weeks.
2. To avoid a bad executive hire, start searching before you need the person, be extremely picky, and expect a good search to take at least six months. Ghodsi says to do a lot of back-door references, since he finds only 10 to 20% of front-door references are truthful.
3. When Databricks moved from product-led growth to enterprise sales, Ghodsi stopped screening sellers for technical skill and looked for professional aggression, strong emotional intelligence and command of the room. He also looked for someone who had built a sales machine, not just run one.
4. Study a rival's weaknesses and aim your strengths at them, but don't copy their features. Databricks pressed on Snowflake's proprietary format, weak AI support and high cost.
5. If sales resists something you believe in, Ghodsi says to put it in the comp plan instead of arguing, and the debate stops.