Lecture 8 - How to Get Started, Doing Things that Don't Scale, Press
Executive Summary
Stanley Tang, Walker Williams and Justin Kan explain how DoorDash, Teespring and Twitch approached early customers and publicity. Their historical Stanford lecture moves from a simple delivery experiment to hands-on customer support, fast product tests and pitching news to the right audience. Notes cover the first roughly 49,700 characters of public auto-captions for the 52-minute lecture, with speaker identities checked against the official Stanford CS183B archive. The date is the October 16, 2014 video upload. The final questions are cut off. Revenue, fundraising and customer figures describe 2014; founder anecdotes and press tactics are not guarantees. Personal-distress passages and a closing reference to deceptive press tactics are not summarized.
Key Takeaways
Test a real customer problem
Stanley Tang describes discovering unmet delivery demand through conversations with a shop owner and other small businesses. DoorDash tested that demand with a simple landing page, menus and a phone number before building a full dispatch system.
Do customer work yourself
Tang says the founders delivered orders, dispatched drivers and handled support. They also emailed new customers each evening to ask about their first delivery. This manual work taught them what software would eventually need to do.
Expect difficult first sales
Walker Williams describes days of meetings and design work to sell 50 shirts for a local nonprofit. He argues that initial customers require personal effort and help founders learn the product's actual selling points.
Speed over clean code
He says the launch product will almost certainly not be the product that scales, so optimize for speed. His CTO duplicated the codebase and database to build enterprise features in a separate product rather than take a month.
Aim press at the right audience
The Twitch co-founder says early press targeted gaming industry trade sites and game developer blogs, because they wanted the gaming industry and advertisers to see Twitch as important.
Find a story people would read
Kan describes launches, fundraising and meaningful milestones as possible stories. He advises stepping outside the founder's enthusiasm to ask whether an unfamiliar reader would care about the proposed news.
Prepare and follow up
Prepare bullet points before a call, send a follow-up with screenshots, spellings and collateral a day or two before launch, and do it yourself before hiring a PR firm.
The Conversation Breakdown
Test a real customer problem
⏱ Untimed topicStanley Tang describes discovering unmet delivery demand through conversations with a shop owner and other small businesses. DoorDash tested that demand with a simple landing page, menus and a phone number before building a full dispatch system.
Do customer work yourself
⏱ Untimed topicTang says the founders delivered orders, dispatched drivers and handled support. They also emailed new customers each evening to ask about their first delivery. This manual work taught them what software would eventually need to do.
Expect difficult first sales
⏱ Untimed topicWalker Williams describes days of meetings and design work to sell 50 shirts for a local nonprofit. He argues that initial customers require personal effort and help founders learn the product's actual selling points.
Speed over clean code
⏱ Untimed topicHe says the launch product will almost certainly not be the product that scales, so optimize for speed. His CTO duplicated the codebase and database to build enterprise features in a separate product rather than take a month.
Aim press at the right audience
⏱ Untimed topicThe Twitch co-founder says early press targeted gaming industry trade sites and game developer blogs, because they wanted the gaming industry and advertisers to see Twitch as important.
Find a story people would read
⏱ Untimed topicKan describes launches, fundraising and meaningful milestones as possible stories. He advises stepping outside the founder's enthusiasm to ask whether an unfamiliar reader would care about the proposed news.
Prepare and follow up
⏱ Untimed topicPrepare bullet points before a call, send a follow-up with screenshots, spellings and collateral a day or two before launch, and do it yourself before hiring a PR firm.
Core Frameworks & Ideas
Experiment before infrastructure
Tang separates proving demand from building a scalable operation. Start with the simplest workable test, learn from real orders and automate only when demand justifies it.
Customer-to-champion loop
Williams links customer support, outreach to people who leave and fixing bad experiences with users becoming advocates. Use direct contact to learn what the product is missing.
Audience-first press
Kan starts with a business goal and the readers who could help achieve it. Choose relevant outlets, prepare a clear story and provide accurate names, screenshots and launch timing.
Actionable Next Steps
- ✓Write personal thank-you or feedback emails to your first 100 customers.
- ✓Choose speed over perfect code for the first version and revisit later.
- ✓Pitch your own launch story to trade press in your niche.
- ✓Write the business goal and intended audience before seeking press coverage.
Notable Quotes
“Those first users are always going to be the hardest”
“you need to optimize for speed over scalability and clean code”
“you want to do it yourself before you hire someone else to do it”
The One Thing to Remember
Those first users are always going to be the hardest