Brent Ryan, a Senior Developer at Zynga presented Scalability and Metrics Applications for Software Development at the 11/14 Tech Meet And Share
* Core of Systems should be built on statistics gathering
* Some of those metrics could be:
* Productivity
* Features over time compared to:
* Revenue - Top Line
* # of Consumers
* CPU
* Memory Footprint
* Be sure to gather them over number of years for up or down trends
* Metrics need to be over large timeline to see major shifts
* Metrics need to be visible to whole company to unleash full power
* Developers understanding revenue affects of feature will help the business
* Business understanding the revenue affects of poor CPU will do the same
* Book recommendation - Super Crunchers - Ian Ayres
* Software Development
* Utilize the most automation you can
* Feature to QA should be automatic and take no more than 5 minutes
* Rapid cycles and automation allow for scaling during down business cycle or business growth cycle
* Need to ensure refactoring is built into new development
* One idea is to develop features then pull them into architectures as commonized pods that can be reused for new software products.
* Don't slow down productivity at cost of customer enjoyment
* Tools to use:
* CollectD - tools for data analytics
* Rightscale - automation
* Cruise control - automated builds
* Create a home grown stats manager for things like revenue generation or new customers over release cycles.
