Understanding incentives is how you actually change the world

Executive overview

Most founders believe a good idea plus software equals disruption. It doesn't. The real obstacle is the web of incentives holding existing systems in place — and most founders haven't mapped it.

The people who do change the world get small things right first. They study every stakeholder, design around real motivations, and research what failed before them.

Getting the small things right — incentives, research, constituent needs — is what separates world-changers from people pushing on trees with roots.

Building for what you want vs. what customers want

• Founders inspired by Jobs or Musk try to force the universe to their will; most can't.
• Convincing someone to want something they don't want is nearly impossible.
• "This thing is annoying, let's build a website to kill it" ignores why the thing exists.
• Car dealers and realtors persist because of incentive structures software can't simply remove.
• Simplistic engineering-efficiency arguments break down in the real world.

The recruiter problem: a case study in ignoring incentives

• Classic hiring startup thesis: non-technical recruiters are the problem; software empowers hiring managers directly.
• The premise is partly correct — engineers are better at evaluating engineers.
• But engineers don't enjoy doing it, and it's not the best use of their time.
• Recruiters have more institutional power than founders expect.
• The hiring manager is your "talk friend" — agrees with you, never changes behaviour.

Getting small things right: the ride-sharing example

• Ride sharing was a common, repeatedly failed startup idea before Uber.
• The leap of "get in a stranger's car" was genuinely bizarre at launch.
• Uber started with hired limo cars — a normalised behaviour — before expanding.
• Every small detail had to be designed around actual constituent motivations.
• Understanding all parties in the ecosystem is what made it work.

Doing the research

• Founders often apply to YC with ideas they haven't Googled.
• Researching prior attempts is free and fast — not doing it is willful ignorance.
• Young founders assume "I haven't heard of this" means it hasn't been tried.
• Asking friends is far weaker than searching; friend knowledge is similarly shallow.
• Psychological resistance: founders don't want to find something that kills the dream early.
• During the dot-com era and the 2000s, nearly every idea was tried — study the history.
• Instacart knew Webvan existed and intentionally designed around why it failed.

Human problems outlast platforms

• Platforms change fast; human problems stay constant.
• Grocery delivery, food delivery, job search — same problems across Web 1.0, 2.0, mobile.
• New technology is a new technique attacking an old problem, not a new problem.
• Almost guaranteed: you are not solving a new problem.

Using experts without being captured by them

• Experts can be too biased positively or negatively based on a small personal data set.
• An expert who lost money on your category will be pessimistic; one unfamiliar with it may be optimistic.
• Expertise does not equal optimism — often the opposite.
• Brex spoke to everyone who had tried challenger banking; they absorbed the useful data and ignored "don't do it."
• Get a wide variety of expert opinions; synthesize, don't blindly accept.