How to Launch a Small Software Project (RIA Unleashed: Boston 2010)
How to Launch a Small Software Product by Jesse Warden
Why do it?
- You might want to change the world, do good, and make money.
- Think about it as the next stage of your software career.
For the Money. Think Zynga. There is more money in products than consulting.
Lifestyle. Get to work when you want.
People that have made the switch:
Nick Bradbury: HomeSite, TopStyle and FeedDemon.
Dan Florio: RunPee.com
Randy Troppmann: Runningmap.com
Skill Validation vs. Unapplied Academia
- Consumers (more fickle) vs. Clients
- Vision you need to show the world
Service vs. Product
- Service work = working for another business
- low risk lots of opportunity/jobs
- low return: company takes a cut of your pay
- Product work = building a product consumers buy
- high risk, no guarantee it'll sell
- high return you own all profits: scale is higher
What doesn't work
- not releasing a working build
- coding first, designing second
- non-standard architecture
- no clearly defined goals
- misaligned goals
- unsustainable cost structure
- scalability: ignore it
- no competency
- not getting along with partners (someone that undoes your changes in SVN)
- not releasing early
- not releasing often
- no commitment
- mythical man month
- waterfall
What works
- designing to 70% finished
- clear goal(s)
- clear finish line
- competency: should you build, or learn
- something simple, done really well
- "Red ocean versus blue ocean"
- why people buy (Sandler Sales System): fear, pleasurable experiences, get out of trouble
- improving an existing technology
- first to market
- small agile teams: UIKit of iOS = 1 person vs. XCode = 16 people
- market research
- iterative development
- motivation is sacred
2 methodologies that work
Plan A: Product process
1) Research and documentation
- who my market is, most viable product, tipping point
- list your goals (list features)
- research market or define personna
2) Design and iterate
- sketch interface on paper, if fail, refine #1
- wireframe through, not fidelity
- build design only 70%
3) Develop
- commit to 1 feature per build
- release bi-weekly or monthly basis
4) Launch
- alpha: on right track or wrong users
- beta: need a website (download, screen shot, feedback, easy to access you)
- get beta out the door
Plan B: Reasoning
Kind of like throwing soldiers into a position where there is no escape.
Only one rule: release a build on 1st and 15th of ever month.
Marketing ideas
- gather quotes from forums for text
- use website email as primary email
- get friends to RT on Twitter
- use all social networks
- invest in metrics
- discounts
- contests
Motivation and focus
- no motivation, no products
- no focus == slow velocity
- set aside time
- oasis/tortoise environment
- iterator and user feedback
- feedback versus validation
- blocks of your time
- consulting or freelance allows you to build time to devote
Created Powerz as an example
- People 26 years old don't read book
- D&D isn't like wow
- make a digital solution
- 1) launch
- 2) super simple thing
- 3) get bought or get shutdown (get attention)
Lessons learned
- do something simple, get it work, provides validation
- motivation is sacred, focus is essential