Building a massive company requires a dual focus: investing in new innovations and constantly grinding to improve the core business. The latter is often unglamorous but is critical because the natural state of technology is decay, and the core business funds future bets.
To build a complex real-world business, the founding team did every job themselves. This hands-on experience provided critical insights that algorithms or data analysis alone could never uncover, such as knowing not to assign a driver if food isn't ready.
At DoorDash, disagreements between smart people are not resolved by who writes the best document or has the most seniority. Instead, their "bias for action" value means they ship something—even a hacked-together prototype—to get real-world data and let the market settle the debate.
DoorDash's CEO frames the market as two battles: for digital attention (bits) and for facilitating the physical world (atoms). DoorDash focuses on moving atoms (goods) to complement the digital ecosystem, which clearly defines its strategic focus against other tech giants.
Focus on what customers value (e.g., delivery speed, order accuracy) rather than internal business metrics like ARR or user growth. This approach naturally leads to a better product roadmap and a more defensible business by solving real user problems.
When facing constant rejection from investors, the ultimate test of whether a founder's vision is ambitious or delusional is customer behavior. Despite being a non-consensus bet for years, DoorDash persevered because metrics like customer retention proved people genuinely wanted the product.
DoorDash uses the value "One Team, One Fight" to define everyone's job as "helping the customer win," irrespective of job title. This fosters a culture of high accountability for the end result while simultaneously ensuring low blame, as everyone shares responsibility when problems arise.
To fight complacency and find product flaws, DoorDash's CEO advises using the service in concentric circles 15 minutes further out from city centers. The product experience often degrades quickly in these less-optimized suburban areas, viscerally highlighting customer struggles and revealing improvement opportunities.
DoorDash is creating a unique data moat by digitizing physical-world information unavailable on the internet, like hyper-local parking data or real-time store inventory. This proprietary dataset, which LLMs cannot currently access, becomes a key strategic asset for building specialized AI models.
Instead of focusing on the 'how' (chat vs. voice), DoorDash's AI strategy starts with the 'what': the customer's complete, end-to-end job. For DoorDash, that's getting a physical item delivered. This grounds AI development in solving a real problem, preventing teams from chasing shiny tech without purpose.