Running a startup is tremendously hard. When I started out with my first company, I felt that though there was a huge amount of high level advice, such as “focus on building a great product” given to me, there was little tangible advice on how to manage the subtle nitty gritty details of day to day operation. While I failed with my own startup, I would like to pass on some of the learning’s I gathered, so you hopefully can avoid these pitfalls. For those of you familiar with YC, some of my advice overlaps with PG’s advice on how to run a startup.

How to run a tech startup:

• Communicate firm, clear, concise (no long emails, no fluffy words or long speeches/meetings, just get to the point and start WORKING on it, rather then talking about it.)
   
• Have clear roles/responsibilities/expectations within your team and give people the freedom to choose how they make it happen. Example: If you have a UI guy on your team how codes the frontend it is not your job as a CEO to tell him what color the button is supposed to have. Let him decide but measure him by clearly communicated goals (front end conversion, conversion of page view to $, render time etc.). To collaborate not cooperate should be your goal. Have your team leads show their work of the last week for 20 minutes each. Then have the entire team give feedback to them. But remember, the decision on what to change is theirs.
   
• No meetings, but the daily standup meeting (each person gives a 1 minutes update on what they work on today, what they are stuck with and who they need help from)
   
• Everyone must help with marketing and distribution! Getting the word out about your new product or company is every bit as hard as building it. Have everyone in the company pitch in. Practice guerrilla marketing through small bloggers, social media, inexpensive ads, guest blogging etc. Avoid the expensive PPC campaigns.
   
• Work lean and test your hypothesis as early as possible (even if you do not have a complete product - you can test market demand by creating a slide show, a video, a blog post etc.). Avoid building stuff that is not necessary!
   
• Launch fast/often. If you decide to build a feature (you should say no to most features and only focus on the true hypothesis you are trying to prove), speed is key. Iterate fast! Speed comes from simplicity. Do not over-architect, do not worry about too many edge cases. Do not worry about scalability. You can worry about all these cases once you have proven that your product is desirable.
   
• Understand your users. Have a fanatical customer support. Email/call/meet/survey them to understand what they want. After asking 100 people you will get 110 opinions, but in this sea of opinions you will see a pattern form. The pattern will lead you to the 2 or 3 core value added propositions that the users want. Implement those and only those.
   
• Silicon Valley is binary. You are nobody until you are somebody! Fund raising is useless without traction/leverage. So avoid it until you have traction. Get Ramen profitable first.
   
• Prioritization is saying ‘no’. Have 1-2 goals max. Make what you measure. Plot a chart every day with your key metrics to stay on track.
   
• Spend little (be resourceful/frugal). You don’t need fancy lunches, a nice office or a launch party. Cut out everything. All you need is good equipment to code with and your EC2 instance to run it on.
   
• Embrace the uncomfortable. Starting a company is hard. Nothing gets handed to you. You have to fight for every user, convince every investor, hustle for every BD deal. Accept that these things are hard, and that you will get rejected 99% of the time. Do NOT loose perseverance over it.
   
• Deals fall through. Assume it’s not a deal until the money is in the bank. Do not get demoralized if stuff does not happen.
   
• Exercise (for better sleep and stress management)

Last but not least, good luck!