Michael Seibel - Building Product | Summary and Q&A

Summary:

The video discusses building a successful product, emphasizing talking to customers, narrowing the problem, iterating quickly, measuring metrics, and having a structured development process. It provides anecdotes and advice from the speaker's experiences at startups like Justin.tv and Twitch.

• Having a strong technical founding team is critical. Justin.tv succeeded because the founders were not intimidated by technical challenges.
• Don't spend too much money early on. The Justin.tv founders lived very cheaply on $500/month, allowing them to survive early mistakes.
• Make sure your product solves a real problem for real customers. Talk to your users frequently to understand their needs.
• Build a process for rapid iteration. Have structured sprints to build/measure features quickly. Avoid long dev cycles.
• Be willing to change your solution, not your problem/customers. Pivoting to a new problem should be rare. Iterate your solution frequently.

Q: What should you focus on first when building a product?

A: You should start by clearly defining the specific problem you are trying to solve and who your target customer is. Don't get distracted by your solution - identify a real problem experienced by real users frequently and intensely. Talk directly to your target customers to understand if they have this problem and would find value in your solution.

Q: How can you set yourself up for faster iteration?

A: Build a streamlined product development process with structured specs, regular standups or meetings (e.g. weekly or biweekly), and easy/medium/hard prioritization. Use metrics to track what users actually do and tie features to moving key numbers. Write everything down instead of relying on memory and verbal debates. Execute quickly - long dev cycles kill momentum.

Q: What metrics should you focus on?

A: Pick 5-10 key stats that reflect core actions users take in your product. Use a tool like Mixpanel instead of just Google Analytics page views. Make sure your team knows how to use your metrics tool. Name metrics so others understand them. Tie features and experiments to moving specific numbers.

Q: How do you know if you have the right customers?

A: Target the most desperate, intense users first. Don't listen to your non-user friends. Identify bad customers who take advantage or complain constantly. Charge money early to validate willingnesss to pay. Talk to real users experiencing the problem frequently in their work. Iterate on the solution, not the customer/problem.

Q: How do you balance imagination and iteration?

A: Have a bold 10 year vision, but break it down into tactical 2 week sprints. Build a process to quickly test ideas and double down on what shows traction. Avoid long dev cycles focused on theoretical perfection - talk to users and adapt. Steve Jobs iterated the iPhone yearly from the crappy v1.

Q: How do you know when it's time to pivot versus iterate?

A: Pivoting means changing the problem or customer you are targeting. This should be rare - don't give up after just 2 months. Iterate means improving your solution. Before pivoting, make sure you've given your solution enough time to properly test it with users. Pivoting is for when you clearly have the wrong problem or customer. Iterating is for refining the implementation.

Q: How do you get early user feedback on your MVP?

A: Get your MVP in front of real users as early as possible, even if it is bad. Watch them use it and ask follow up questions. Don't rely on what they say - observe what they actually do. Use analytics to see how they engage. Offer to pay users for feedback if needed. Your goal is to learn, not pitch them. Be open to finding out your initial assumptions were wrong.

Q: How do you balance speed vs scalability early on?

A: Optimize first for speed and learning. Get a working prototype out for users to react to. Ship the minimum feature set needed to start getting traction. Technical debt is fine early on. However, build your architecture so it can scale modularly once you have product-market fit. Don't scale things prematurely that haven't proven value. Always be ready to throw away or rebuild parts that don't work.

Q: How do you get users to care about your product?

A: Users will care if you solve an intense problem they have frequently. Understand their desires and pain points deeply through conversation. Exceed their expectations on delivering value - build what they ask for. Have amazing customer support. Build features users request. Create an emotional connection through your branding and personality. Deliver incredible product experiences they can't get elsewhere.

Q: How do you get good at estimating if ideas are easy, medium or hard?

A: Practice - it takes time to build intuition. Learn your tech stack deeply. Break down ideas into specific components/screens/flows. Discuss with engineers on the team. Refer back to past estimates and compare to actuals. Estimating gets easier as you build more features and understand your architecture and team's velocities.

Q: How do you balance building for power users versus the mainstream?

A: Power users are helpful for traction early on so do cater to them. However, don't optimize purely for power users as they have very different needs from mainstream users. Build core mainstream experiences first before layering on advanced functionality. Talk to power users but observe mainstream usage patterns. Prioritize fixes that help mainstream over power users.

Q: How do you avoid falling in love with your original idea?

A: Talk to real users continuously - let their feedback drive development, not your own ego. Have a beginner's mindset and assume your initial ideas are wrong. Focus on solving user problems, not being attached to your solution. Be willing to throw things out and rebuild. Have a culture of empiricism and be data-driven. Kill your darlings - stay focused on what is best for customers.

Q: How do you get your product in front of users early?

A: Leverage existing channels you may have access to - your own networks, communities you're a part of. Be creative in reaching your target users - find where they already congregate online and offline. Offer considerable discounts or exclusivity to incentivize early adopters. Partner with influencers/enthusiasts who can get the word out.

Q: How do you balance perfectionism versus moving quickly?

A: Don't let perfect be the enemy of better. Set timeboxes for development work - ship the progress made rather than waiting. Trust in your ability to iterate. Focus perfectionism only on core user flows. Add polish/enhancements iteratively after getting feedback. Let non-critical bugs be technical debt to address later. Start with ugly but usable and refine over time.