Continuous Innovation

VP Engineering, Prosper at Prosper
Sep. 15, 2013
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
Continuous Innovation
1 of 35

More Related Content

What's hot

Applying the Lean Startup Model to the EnterpriseApplying the Lean Startup Model to the Enterprise
Applying the Lean Startup Model to the EnterpriseJez Humble
Developer NirvanaDeveloper Nirvana
Developer NirvanaVMware Tanzu
DevOps & Security from an Enterprise Toolsmith's PerspectiveDevOps & Security from an Enterprise Toolsmith's Perspective
DevOps & Security from an Enterprise Toolsmith's Perspectivedev2ops
Agile for EveryoneAgile for Everyone
Agile for EveryoneAlexander Postnikov
2009 06 01 The Lean Startup Texas Edition2009 06 01 The Lean Startup Texas Edition
2009 06 01 The Lean Startup Texas EditionEric Ries
How BDD enables True CI/CDHow BDD enables True CI/CD
How BDD enables True CI/CDRoger Turnau

Similar to Continuous Innovation

SDLC SmashupSDLC Smashup
SDLC SmashupLester Martin
Walk, Don't Run: Incremental Change in Enterprise UXWalk, Don't Run: Incremental Change in Enterprise UX
Walk, Don't Run: Incremental Change in Enterprise UXuxpin
How To (Not) Open Source - Javazone, Oslo 2014How To (Not) Open Source - Javazone, Oslo 2014
How To (Not) Open Source - Javazone, Oslo 2014gdusbabek
Applying Innovation in Software DevelopmentApplying Innovation in Software Development
Applying Innovation in Software DevelopmentAmish Gandhi
Customer Presentation: Digital Globe's road to Continuous DeliveryCustomer Presentation: Digital Globe's road to Continuous Delivery
Customer Presentation: Digital Globe's road to Continuous DeliveryXebiaLabs
DevOps - Understanding Core ConceptsDevOps - Understanding Core Concepts
DevOps - Understanding Core ConceptsNitin Bhide

Recently uploaded

Orchestration, Automation and Virtualisation Maturity ModelOrchestration, Automation and Virtualisation Maturity Model
Orchestration, Automation and Virtualisation Maturity ModelCSUC - Consorci de Serveis Universitaris de Catalunya
Easy Salesforce CI/CD with Open Source Only - Dreamforce 23Easy Salesforce CI/CD with Open Source Only - Dreamforce 23
Easy Salesforce CI/CD with Open Source Only - Dreamforce 23NicolasVuillamy1
ISO Survey 2022: ISO 27001 certificates (ISMS)ISO Survey 2022: ISO 27001 certificates (ISMS)
ISO Survey 2022: ISO 27001 certificates (ISMS)Andrey Prozorov, CISM, CIPP/E, CDPSE. LA 27001
Announcing InfluxDB ClusteredAnnouncing InfluxDB Clustered
Announcing InfluxDB ClusteredInfluxData
Unleashing Innovation: IoT Project with MicroPythonUnleashing Innovation: IoT Project with MicroPython
Unleashing Innovation: IoT Project with MicroPythonVubon Roy
Privacy in the era of quantum computersPrivacy in the era of quantum computers
Privacy in the era of quantum computersSpeck&Tech

Recently uploaded(20)

Continuous Innovation

Editor's Notes

  1. One of my interview questions for an Engineering Manager: What is the one thing you should be religious about when exercising “agile” software development?The answer I’m looking for: The one mantra to be religious about is to not be “religious” about anything – that’s why it’s called “agile”.No business plan, not product vision, no strategy, no process is perfect in it’s initial conception. Likewise, none is successful without adjusting and course correcting along the way.There is no silver bullet: Judgment trumps process, process serves the organization and not vice versa.
  2. +7? At Softlab, as architect, I architected three key technology pivots that didn’t leave behind the existing technology or customers though the business had initially said that would be ok.This technology Pivot (actually two) that Enabled other pivots as wellAs VP, Eng at BroadVision, Inc, before, during and after the pivot… Also a Technology Pivot that Enabled other pivots as well (segment pivot, …) ultimately from 1 to 7 product linesBroadVision also did an early market and technology pivot in 1996 from BroadBand to Internet – same basic premise, radically different technology and market.At Cooperative Solutions, Inc when PowerSoft entered the space and did the pivot “for” us… -> PowerSoft: someone else ran with the pivot… Industry's largest acquisition to date…“PowerBuilder was originally developed by PowerSoft in 1991. PowerSoft went public in 1993 and was acquired by Sybase for $904 million in Sybase stock in 1995.”
  3. +1 In order to enable continuous innovation, you need a team that will embrace it. Start with the people.Hire for potential not current state knowledge: your employees need to be able and eager to adapt to changing markets and changing processes.Hiring a structural engineer for a software engineering position: At IMVU we hire for passion and ability to collaborate and learn, and we manage to inspire.When we interview at IMVU, we don’t probe for knowledge so much; we test for problem solving, intelligence and ability to ask the right questions to understand the problem.Manage:What makes for the most productive employee?Happiness, Passion, Engagement – Consider also what Google does for it’s employee and how one of the first things Marissa did when she came to Yahoo was to institute the same sort of benefits. Everybody is a Genius. But if you judge a fish by its ability to climb a tree, it will spend its whole life believing that it is stupid. ~ Albert Einstein If you want to climb trees, hire monkeys; if you want to innovate, hire people with innovative mindsets… Company performance based bonuses
  4. We continually improve the spin up guide with each new hireMore on this later
  5. +2 Engineering manager interview question: What would you say is the single most important contributing factor in making an engineer productive?The answer I’m looking for in one form or another: A happy engineer. Happy engineers are motivated, passionate, engaged and productive. Quoted on 1-on-1 question: are you happy?5 whys of a one-on-one: What are you working on? Why are you working on that? Why does that matter? Which company goal does that map to? Why does that company goal matter?Engineers need to believe in and care about the final outcome for product and company.Transparency
  6. Run Experiments: see above In order to continuously deploy, it is vital to have good test coverage. If you don’t start with it, how do you get there – where to start adding coverage? Use an immune system to catch abnormalities in any changes that you roll out in production Start with MVP’s to learn fast, fail fast, and start iterating and continuous innovation asap How to learn from what went wrong in production release, processes, … find the teachable moment, the take-away from any situation or outcome.
  7. We hire smart people to come up with good hypothesis on what will make our customers happy. However, we also validate those hypotheses (not every can be a Steve Jobs)In this case, looking at the data, it appears the impact of the change was minor.Label y axis
  8. +3 We hire smart people to come up with good hypothesis on what will make our customers happy. However, we also validate those hypotheses (not everyone can be a Steve Jobs)In this case, looking at the data, it appears the impact of the change was minor.Also compare click-through rates to new-paying-user conversions to LTVsWhich data to look at - Photo Stream example of engagement vs buy from picture vs downstream buy vs buy from same creator vs buy to generate better photos and collect accolades
  9. As the company grows, the mission and vision evolves, and new technologies and tools appear, IMVU too will continue to evolve rather than stagnate.Each team chooses the processes and tools that work best for it as discovered through experimentation followed by retrospectives, and that knowledge is shared across teams.1. Existing experience: Millions of monthly hits on landing page = easy to quickly see the impact of a change; however, no longer solving for early adopters = higher bar2. Infra-Maint: Balanced Priorities, Graphs, Interrupt Engineers & Lanes/Stories3. New Paradigm: Design>User-Test>Project-Brief>Develop MVP>Iterate4. A New Game: Start with no users, Story Board and user test, prototype and beta test, iterate, validate, release, iterate5. New Infrastructure: TDD
  10. +4 What it means to release an MVP differs fairly significantly with what you are introducing, to whom, where and what you hope to learn…e.g forgiving early adopters vs existing/mainstream, free vs paying users, funnel conversion vs LTVKnow your audience: eBay PM at IMVU – why people come to your site influences what they are tolerant of – they don’t just come to IMVU to buy/sell (credits/vg), primary focus is social interaction In a rapidly changing and increasingly diverse global market, it’s hard to know what will flyGet products in front of customers asapTakes RisksIterate based on feedback and dataOr, shoot the dog(aka Fail-Fast) & LearnSee also:http://www.rise.net/blog/minimum-viable-product
  11. Intuit (from 6-12 months to 2-3 weeks), Yahoo (from 2-3 weeks to 1-2x daily), IMVU (every 40 minutes to every 7)Benefits of small changes for product and for engineering – see also Immune system and continuous integration
  12. If you’re starting with a brand new company that believes in TDD, that’s one thing…If you’re an existing company and want to add tests in hindsight, where do you start?
  13. IMVU’s Immune System allows a new change to be rolled out to one or two servers first. The system monitors memory usage, cpu usage, customer usage, … If any threshold of change is exceeded, the system automatically rolls back the change and alerts the developer(s) that contributed to that change.With continuous deployment, this change-set is usually very small and easy to debug.We have have extensive monitoring of production systems to alert us when something goes awry.If something does go wrong in production, there is a post-mortem with follow-ups to be addressed within the next sprint.
  14. +5 Use this as an example of 5 (4-6) whys, lessons learned, good experience Mentoring Testing Monitoring and Immune System
  15. +6?Immune SystemHypothesis BuildsRollbacks and Finding the Issue
  16. MalcolmGladwell’s The Tipping PointIMVU Forum Moderators and other Mavens provide good insights into what will resonate with existing user base.Stipso, Steven Drost – Lunch conversation about the value and relevance of traditional survey results.
  17. http://planningpoker.com/
  18. Root Cause Analysis – Titanic (Cause Mapping) http://www.youtube.com/watch?v=GOVeO5_0qD05 Whys Fun Skit http://www.youtube.com/watch?v=5od0T_Fhk9oSocial Intelligence:http://itunes.apple.com/gb/book/social-intelligence/id436993236?mt=11 http://www.amazon.com/Social-Intelligence-Science-Human-Relationships/dp/0553803522
  19. As the company grows, the mission and vision evolves, and new technologies and tools appear, IMVU too will continue to evolve rather than stagnate.Each team and individual may innovate and evolve separately as needed
  20. Yahoo Media’s approach vFlickr’s even though Yahoo acquired Flickr…
  21. We often rip out or rewrite entire systems that have grown organically over time. We learn from refactoring existing code and take learnings from that into our review process for new code (often written by the same engineer just years later)
  22. “Issues with Traditional Methods:Artificial gameplay sessions- Many potential biases- Distorted data (interpreted behavior)- Lack of empiricism- Missing elements of objectivity- Sometimes difficult to establish emotions, baselines, and independence
  23. Imagine going to the movies, buying tickets and 10 minutes deciding it’s a really crappy movie; would you still stay to the end just because you’re already there and paid?Impermanence (split this slide into 2)