General
Notes from Eric Ries’ Lean Startup Event in Boston
Thursday, January 27, 2011
Microsoft NERD
- “Most of you are entrepreneurs, which means most of you are going to fail.”
- we’re wasting time by pouring smart people into ventures set to fail.
- remember when Web 2.0 was still cool?
- we have to get better about talking candidly about failure
- waterfall methodology/traditional product management
- all entrepreneurship is difficult to manage, so we manage it by a metaphor
- why doesn’t waterfall work?
- we successfully can execute the plan, but we don’t know if the plan is good
- metaphor: drive at the right speed, on schedule, but we drive off a cliff
- three “shadow beliefs†that we don’t really know
- #1: “we know what customers wantâ€
- #2: “we can accurately predict the futureâ€
- #3: “advancing the plan means progressâ€
- if we’re building something nobody wants, who cares if it’s high-quality? is it progress to advance the plan, or do your job, if the plan is failure?
- main idea against waterfall/traditional product management: “If we’re building something nobody wants, what does it matter if we accomplish it?â€
- Lean Revolution
- in Japan, started the lean manufacturing mantra
- old mantra is wrong
- if we go lean, we can drive waste out of the process
- Ries picked up Agile Product Development
- despite the name, agile still has its roots inside larger companies
- Agile works when the problem is known, solution unknown
- (this doesn’t work in startups)
- in Japan, started the lean manufacturing mantra
- we need a iterative, waste-free customer-based way
- Lean Startups’ unit of progress: validated learning
- Eric Ries at IMVU: his really good idea (interoperability with instant messaging + 3d avatar system), used 6 months to build the product, which actually turned out to not be a good idea
- had to throw it out and pivot – all of the code was thrown away
- he built something that nobody wanted, so it was of no use
- If my goal in the last 6 months was to learn this important bit about customers, why did we need 6 months? we would have learned the same from less code. or… even any code?
- we could have basically learned the same with a fake landing code with a spec. we would have known people didn’t want the product because we didn’t get any clickthroughs on a landing page
- job #1 in startup: do whatever it makes to make the company successful period.
- need new vocabulary for what entrepreneurs “doâ€
- entrepreneurship is management
- institution-building: the coordinated building of something
- we’re prejudiced about “management†– we need something called Entrepreneurial Management
- design and build institutions under situations of high uncertainty
- THE PIVOT
- Eric Ries apologies for the overhypedness of this word.
- What do successful startups have in common?
- they know the process on how to tell good function from bad function
- systematically re-evaluation of your startup until you achieve product-market fit (PMF)
- what is runway? it’s how many pivots you have left
- pivots powered by learning
- lean startups are all about speed.
- if you have fast cycle times to tweak until you achieve product-market fit, you can get to PMF faster than others – speed
- startup: catalyst for ideas -> code
- customers interacting with code: qualitative and qualitative data
- Apple found a balance: don’t overinvest into something that doesn’t matter, but put in just enough to get to the next pivot (more learning)
- Lean startup principles
- 1. entrepreneurs are EVERYWHERE
- EVERYWHERE
- “we can find entrepreneurs everywhere we seek out uncertaintyâ€
- no “safe industries†– “waves of destruction†went through industries
- innovation accounting
- the most boring topic in entrepreneurship
- today: accounting -> accounting where money goes
- not what it was made for – accounting was invented to drive accountability for managers
- it’s a mathematical model for management
- at GM, they had mathematical models – accounting was meant to separate what managers did versus the random environmental factors that happened
- (if a manager exceeded their sales estimate by 10%, but it was a 10% better year for cars all year, they’re at parity, not exceeded)
- we need a new accounting paradigm for entrepreneurs
- actionable metrics, not vanity metrics
- not what happened in the past but the future
- test of value in the startup: customers engage in a voluntary exchange with the startup
- time, attention, referrals,
- Facebook: viral growth, value component, because there was customer engagement and interaction
- need an accounting paradigm for entrepreneurs to demonstrate quantitatively that we’ve done validated learning – real experiment
- we can affect customer actions through our own actions
- actionable metrics: our own actions affect customer actions (mark’s notes: this is repeatable)
- vanity metrics: random environmental events, press spikes, etc. (mark’s notes: this isn’t repeatable)
- EVERYWHERE
- how do we know when to pivot?
- we know when to pivot when our product experiment efforts don’t really move the needle
- we can make excellent decisions 3 years ago, but if our current experiment efforts aren’t doing anything, there is a problem
- the automobile revolution: 500 other companies before Henry Ford that failed because they didn’t have the right process that Ford had
- when we pivot, we DON’T pivot the VISION, we PIVOT the STRATEGY
- 1. entrepreneurs are EVERYWHERE
- should a feature be in your MVP?
- good rule of thumb: no
- http://lean.st/LeanStartupBos
QUESTIONTIME!
- #1: Jason Evanish: how do you keep your engineers having faith in the process when so much is done without them initially for mvps? #leanstartups cc @msmamet
- keep them calm
- abandon the crutch of the reality distortion field to keep them calm
- dishonest to do that ^ — accept uncertainty
- #2: questioning the usage of the word “movementâ€
- it’s not religion, it’s science
- ……. still going
- used to think VCs are the way to evangelize to entrepreneurs
- (though they want to advance their portfolio)
- #3: How do you test a product people aren’t actively searching for?
- when does it work?
- solves an acute problem
- … that customers know they have
- … that customers know the name of the problem
- what if it doesn’t work?
- need another way
- outbound advertising to inbound learning
- IMVU: adjacent markets
- “world of warcraft chatâ€, etc.
- didn’t want to know anything but: what are the differences in conversion rates across channels?
- they found out who the customer wants
- dropbox’s video: an MVP on Digg and etc
- perfectly well-targeted to early adopters
- a huge invite list
- case study at the startups lessons learned conference – http://www.justin.tv/startuplessonslearned/b/262672510
- Ries works with MBAs. why?
- process/discipline oriented
- predisposed to learn a new management paradigm, since they’ve already done that
- entrepreneurs are a bit sloppy about it
- we need to know what exactly is making things go up or down – reject vanity metrics
- when does it work?
- #4: innovation accounting: know what to measure? consistent across all startups?
- the right things to measure are clear and consistent across all startups (but only at a high level of abstraction)
- measure two things:
- value test: does it make value?
- if yes, does it have an engine of growth that is working?
- driving metaphor
- engine: constant revolutions
- driver: build-measure-learn
- you can’t build a list of step-by-step actions of how to drive a car to a destination (with every action, turn on steering wheel, etc.), because you respond to the environment
- however, that’s how we do product development, right now — build a plan, execute on that plan, without responding to the environment
- compulsive behavior with Facebook, like photo tags (viral)
- sticky growth: engine of growth that is addictive
- measure two things:
- engines are all feedback loops: they have compunding effects
- we can have hockey-stick improvements but we don’t know what is driving it
- that is what we’re trying to prevent – control over the actions that really make change and results
- the right things to measure are clear and consistent across all startups (but only at a high level of abstraction)
- #5: advice for getting to market first if your product is easily replicated?
- exercise
- take your second best idea.
- try to get somebody to steal it.
- contact a product manager and try to get them to implement it in their next version.
- <over-repeated>ideas have no value</over-repeated>
- you have to go through the process of REFINING a good idea
- first-mover advantage: extremely rare in reality
- exercise
- #6: how to get this into a company that isn’t really built on lean startups?
- work is a system, and systems are perfect at what they do today
- what is the organizational change process we want to do?
- implementing cold-turkey lean startup methodologies won’t work
- find one areas where there is a stress/pain point in the company
- there are a community of people that might be trying to solve the problem
- try an experiment – lean startups are a science
- can’t have high expectations – have to measure the actions with validated learning
- work is a system, and systems are perfect at what they do today
- #7: how does lean startups affect the world of sales force?
- four steps to the epiphany – Steve Blank
- sales people aren’t really paid at listening – they don’t take no for an answer
- entrepreneurs: no is powerful: we want to figure out why
- Mark’s notes: to sales people, no is a dead end. to entrepreneurs, no isn’t a roadblock or a dead end, it’s an opportunity for validated learning to take place – why did they say no?
- entrepreneurs should start out doing their own sales for validated learning