By Michelle Sun
Here are the notes for the Roadmapping and Execution Lightning Talks and Panel at Startup Product Summit SF. Omissions and errors are mine (please let me know if you find any, thank you!), credit for the wisdom is entirely the speakers’.
- Product manager’s role is to capture, communicate and distill product ideas, and mediate between business stakeholders and makers.
- When building a product, pick two out of the three: quickly, correctly, cheaply. Joe later mentioned on Twitter that he would pick quickly and correctly, as paying for quality is no brainer.
- “Want to increase innovation? Lower the cost of failure” – Joi Ito
- Empower every developer to commit things to the product through non-blocking development (NBD).
- Advocate the move to 100% asynchronous communication because current approach is broken (needs human input to track reality) and remote teams are becoming more common.
“Raw Agile: Eating Your Own Dog Food” – Nick Muldoon, Agile Program Manager, Twitter
- Twitter does dog-fooding by allowing developers deploy to internal server. Dog-fooding allows:
- gathering real data from real (though internal) users.
- increases incentive to produce quality shipped code.
- better feedback. He found that feedback in dog-fooding environment is generally more constructive.
- keeps momentum through a positive reinforcing loop of continuous deployment and feedback. The team gets 50-100 feedback from internal users each day.
- How to decipher and sift through the volume of feedback. Look at only the “love” feedback, then all the “hate”, then discard the middle, categorize and show to the whole team.
- Other important aspects in dog-fooding:
- Automation. Allow deploy more frequently especially internally. ”On any commit, deploy internally.” Avoid accumulating technical debt.
- Visibility. Record progress and share on a wiki.
- Speed. Minimize cycle time (from to do to in progress, to done).
Best Practices for Architecting Your App to Ship Fast and Scale Rapidly” – Solomon Hykes, Founder & CEO, dotCloud
- 3 things to aim for in architecting your app: speed (continuous deployment), scale, future-proofing (be prepared for things moving very fast, avoid bottleneck and need to refactor when adding every new feature).
- What are the patterns/strategies in getting to these three goals?
- Be aware of trade-offs. There is no silver bullet; always trade-offs and prioritization.
- Trade-offs evolve over time. Priorities change. Be aware of assumptions and revisit them from time to time.
- Trade-offs differ from team to team. Be aware of bias in different teams. Always keep ownership of key decisions.
- Put yourself in a position where you are embarrassed, and things are going to happen faster.
“Rocket Powered Bicycles: Avoiding Over and Under Engineering your Product” – Chris Burnor, Co-Founder & CTO, GroupTie & Curator, StartupDigest
- A product connects business priorities with user experience.
- Proposes that instead of Minimum Viable Product (MVP), think about Product: Viable Minimum (PVM). Focus on viability.
- A scientific method to approaching product roadmapping.
- Idea: think about business priorities, user experience. Do not let technical decisions drive your product. Let product drive your technical decisions.
- Test: Viability of the solution is whether it solves the problem it’s setting out to solve. Determine what level of viability is suitable in different stages: GroupTie’s first viable minimum was a keynote presentation that was sent to potential customers.
Scale of tests will vary. Lack of big tests means the lack of breakout growth/ideas, lack of small tests means the team is doing too much.
- Conclusion: Debriefing phase is vital, share test results with the team and learn what it means to the idea. Testing without debriefing is like “talking without listening” in a conversation.
- An unusual example of a PVM is Apple. Product first: cares about user experience and business priorities. Viability second: it just works. Minimalism third: wait till a technology is ripe before adding to the product (no LTE for a long time, no RFID).
Notes on other panels:
About The Author
Michelle Sun is a product enthusiast and python developer. She worked at Bump Technologies as a Product Data Analyst and graduated from the inaugural class of Hackbright Academy. Prior to Hackbright, she founded a mobile loyalty startup. She began her career as an investment research analyst. When she is not busy coding away these days, she enjoys blogging, practicing vinyasa yoga and reading about behavioral psychology. Follow her on Twitter at @michellelsun and her blog.