Source: [How to get startup ideas](http://paulgraham.com/startupideas.html)" by Paul Graham # Part I: How to get startup ideas organically Ideal idea should satisfy all three criteria: 1. What I want to exist 2. What I can build 3. Few others realize is worth doing ## 1. What I want to exist Ensures problem really exists. Otherwise you can build what no-one wants. If you think of "startup ideas" there is a danger of a bad idea that you convince yourself is good. "Made-up" start up ideas are like those made up in a comedy show (little value but sounds plausible) e.g. "social network for pets": - Dangerous because friends will ok it - A lot of people think other people will use it ## 2. Build what few people really need instead of what many people kind of need ![[build.png|600]] ## 3. What users are willing to use no matter how crappy it is Microsoft built Altair basic for Altair users only. Facebook: only for Harvard students. Activation energy is the energy needed for user to switch to your software, from the one they are currently using. Activation energy is provided by the user's need. The more the user needs something the more energy it will give them to try your software. Activation energy is higher for corporate users, since it takes much more energy to make someone switch from one corporate software to another. Activation energy for a new search engine is low. ## 4. Make sure you can expand idea from one few people really need to one many people really need. ![[Screen Shot 2021-06-09 at 1.21.40 AM.png|600]] ## 5. Live in the future and build what's missing If you are on the leading edge of something (on either side - user or creator) then you'll know what the future holds: ![[Screen Shot 2021-06-09 at 1.23.20 AM.png|600]] Zuckerberg was a *user* of computers, realized was missing a forum where people post social. ## 6. Prepared mind subjected to external stimulus results in an idea ![[Screen Shot 2021-06-09 at 1.24.51 AM.png|600]] ## 7. You have to notice an idea not think it up Don't sit there and think about a startup idea, instead just kind of notice it bubble up. ![[Screen Shot 2021-06-09 at 1.26.16 AM.png|600]] ## 8. Startup will take 3-5 years of time So there will be less opportunity as you get older. Get on the edge of some rapidly-changing domain. ## 9. Don't focus on business of the idea Focus on what's *missing* here, and not whether this will make a good business opportunity. ## 10. Question everything What rubs you the wrong way (because you're living in the future)? Right sort of problem should be described as obvious to you. See things that are obvious to you but you can't see them. ![[Screen Shot 2021-06-09 at 1.30.16 AM.png|600]] ## 11. How to spot obvious? 1. Work on hard problems driven by *curiosity* 2. Watch yourself over the shoulder (take note of anomalies, gaps) ![[Screen Shot 2021-06-09 at 1.31.54 AM.png|600]] Give yourself time - prepare your mind for ideas. Less time & you risk starting with a bad idea. ## 12. Work on *cool* projects to you ![[Screen Shot 2021-06-09 at 1.32.30 AM.png|600]] Andriy: This is a "trick" since what is cool to you, automatically means you made a judgment about what is missing. (e.g. self driving cars are "cool" because they don't exist). ## 13. Work on toy things not think of ideas "Think of an idea" $\to$ bad ideas Working on "toy" things $\to$ good ideas come out ![[Screen Shot 2021-06-09 at 1.35.28 AM.png|600]] Toy = "everything an idea needs except importance". ## 14. Live in the future, build what seems interesting ## 15. Prefer school over entrepreneurship classes School - allows time travel into the future. $\leftarrow$ hard part of the business idea Entepreneurship - not learned, more "learn-by-doing" type of thing. $\leftarrow$ easy part of the business idea Why would you want to study the easy part (enterpreneurship)? ## 16. Clash of domains is good for you If I know software (programming) and I learn about another field I'll be better than people in that field because: 1. They don't know what programming is capable of and so never tried it 2. I am completely ignorant of the status quo in that field - so I won't be held back ![[Screen Shot 2021-06-09 at 1.38.19 AM.png|600]] ## 17. Get summer jobs in an unrelated field Computer Scientists should get summer jobs in an unrelated field (e.g. genetics research) and not in software companies (as they usually do). ## 18. Instead of taking classes build things ![[Screen Shot 2021-06-09 at 1.42.39 AM.png|600]] Instead of taking on additional classes build things. That's how both Microsoft and Facebook started. ## 19. Build things not startups Build things with other students who will then become your founding members. Don't build startups. ![[Screen Shot 2021-06-09 at 1.43.51 AM.png|600]] ## 20. Beware of research Very narrow set of ideas count as "research". So ideas that make good startups may not count as good research. Work on side projects - side projects free you. ![[Screen Shot 2021-06-09 at 1.45.22 AM.png|600]] ## 21. Worried your idea might be late? This is one of the signs of a *good* idea! Also, if someone urgently needs it, this is more important than if you're late. Focus on *users* not competitors. ## 22. Good idea + competitors = ok, bad idea + no competitors = bad "Err on the side of doing things where you'll face competitors". ## 23. Do what competitors would have done if they had the Courage to follow through on own ideas E.g. search engines were afraid of doing a good job since that meant users would leave their site sooner. ## 24. Crowded market is a good sign that none of the solutions are good enough But you *must* have a secret weapon. E.g. Google $\to$ crowded market. ## 25. Turn off these filters 1. Taking world for granted (10) 2. UnSexy filter (28) 3. Schlep filter (26, 27) Schlep definition: tedious, difficult ## 26. Schlep filter What you think software startup will look like: 1. Brilliant code 2. `git push` 3. Collect \$\$\$ Doesn't happen! Tedious, mess - real world ideas is what is valuable. ## 27. This was a schlep for them Payments was a schlep for Stripe. None wanted/afraid of dealing with payments. But after Stripe solved that issue they have very little other issues. ## 28. UnSexy filter ![[Screen Shot 2021-06-09 at 1.55.39 AM.png|600]] You despise something so you don't work on it. ## 29. Schlep >>> Unsexy Schlep is much more important to turn off than unsexy. ![[Screen Shot 2021-06-09 at 1.56.27 AM.png|600]] Schlep is an illusion, and if not - it's self-indulgence. Any *good* idea will have lots of schleps anyways! ## 30. Unsexy filter sometimes result of your knowledge But - business means you won't work on what you like - but what makes you \$\$\$. # Part II: Recipes for coming up with startup ideas This section is much weaker way of coming up with ideas, and you should use the Part I if at all possible first. ## 31. Things you really need Make sure it is a thing you really need and not just a "nice to have" app. ## 32. Is there something unusual about you? If there is, then look at your need. If the population norm, will, in the future shift in your direction, then by satisfying your own need you'll satisfy their need in the future. ![[Screen Shot 2021-06-09 at 2.00.06 AM.png|600]] ## 33. Changing ideas is an opportunity to find a need When changing ideas look back on your old idea, and identify any needs you had while you were working on it. ![[Screen Shot 2021-06-09 at 2.00.34 AM.png|600]] ## 34. Try to sell your idea before building it You will either: - Realize nobody wants it - Find a real idea in the process ## 35. Young is a type of different Young = different, unusual If you are young it allows you to view a problem your peers have. Only a young person could start Facebook. ## 36. Unmet need of someone else ![[Screen Shot 2021-06-09 at 2.02.57 AM.png|600]] Talk to them in general terms, not about specific startup idea. ## 37. Treat other person as your customer If you implement something another person needs, even if it is very specific to their case, it is likely there there will be a lot of stuff in that solution you can reuse for many others with similar needs. ## 38. Make their problem your own Work in their field. Work in restaurant as a server if you want to make a server app. ## 39. Look for schlep & unsexy ideas ## 40. What would you pay for right now is someone else made it? Asking yourself this question allows you to bypass the schlep filter. ## 41. Look for broken companies Think from the future - about how to replace. You can replace them on different axis. Andriy: For example, Uber could (and was) replaced on axis of Cost, Cleanliness and Speed of taxi arrival.