startup dna
DESCRIPTION
Startup DNA: the formula behind successful startups in Silicon Valley. These are my personal observations on a few traits that make startups successful. TRANSCRIPT
Startup DNA The formula behind successful startups in Silicon Valley
Startup DNA The formula behind successful startups in Silicon Valley
About me
Who I'm not...
● I'm not Paul Graham, Reid Hoffman, Dave McClure, Marc Andreessen (sorry)
● I don't own a company
● I'm a "rank amateur"
What I offer is a different perspective: "from the trenches"
Who I am
● Had front row seats at very successful startups: LinkedIn, TripAdvisor
● Wrote a lot of code:Monetization, Platform, Infrastructure, Apps
● Created a few programs:Hackdays, [in]cubator, engineering blog, open source
These are my personal observations on a few traits that make startups successful.
Outline
1. Make excellent mistakes
2. Speed wins
3. Boyd's law
4. If you can't measure it, you can't fix it
5. Distribution
6. Sharing
Outline
1. Make excellent mistakes
2. Speed wins
3. Boyd's law
4. If you can't measure it, you can't fix it
5. Distribution
6. Sharing
"Make excellent mistakes"
Dan Pink [1]
The road to startup success is paved with failure
Failed projects
● Answers● Events● Tweets● Github● Countless other experiments that never
made it past a "dark launch"
● Wave● Buzz● Labs● Health● Knol● Catalog● Video● Answers
The Pivot
● Started as a way to share images in Game Neverending, a massively multiplayer online game.
● Rewritten to focus on photo sharing
● Acquired by Yahoo for $35m in 2005
● Started out as burbn, an HTML5 mobile app for location-based social networking with photo sharing as one of many features
● Completely rewritten as a photo sharing focused native app
● Sold to facebook for $1bn in 2012
● Started out as odeo, a site to create and share podcasts
● Struggling to stay alive, they held a hackathon. Jack Dorsey proposed the microblogging concept (originally, text message only).
● 500m users and $10b valuation in 2012
"Timing, perseverance, and ten years of trying will eventually make you look like an overnight
success."Biz Stone [2]
The Failure
Richard Branson
● Virgin Clothes failed● Virgin Cola failed● Virgin Vision failed● Virgin Vodka failed● Virgin Wine failed● Virgin Jeans failed● Virgin Cars failed● ... and many others failed● ... but several hundred others succeeded
"I have not failed. I have just found 10,000 ways that won’t work."
Thomas Edison
"Screw it! Just do it!"
Richard Branson
Reid Hoffman
● His first startup was Socialnet.com: a social network for dating.
● Never heard of it? Exactly.
● Went on to become COO of Paypal, co-founder of LinkedIn, and one of the most successful angel investors of the last decade.
"If you're not failing every now and again, it's a sign you're not doing anything very innovative."
Woody Allen
Outline
1. Make excellent mistakes
2. Speed wins
3. Boyd's law
4. If you can't measure it, you can't fix it
5. Distribution
6. Sharing
"Speed wins"
Stephen Kaufer
You will not get it right the first time. Or the 10th.
Time
Maturity
Mature, stable
Proof of concept
Product development (assumption)
Idea
Time
Maturity
Mature, stable
Proof of concept
Product development (reality)
Idea
Time
Maturity
Mature, stable
Proof of concept
Product development (reality)
Idea
A very large portion of development time is trial and error
In a trial and error world, getting to errors faster is the key to success.
Time
Maturity
Mature, stable
Proof of concept
Product development (technology comparison)
Dynamic/interpreted languages Static languages
Idea
Time
Maturity
Mature, stable
Proof of concept
Dynamic/interpreted languages Static languages
Innovation Advantage
Product development (technology comparison)
Idea
Time
Maturity
Mature, stable
Proof of concept
RoR, PlayServlets, Spring
Innovation Advantage
Product development (framework comparison)
Idea
Time
Maturity
Mature, stable
Proof of concept
Agile, ScrumWaterfall
Innovation Advantage
Product development (development method comparison)
Idea
"If you're not embarrassed by your first release, you've launched too late."
Reid Hoffman
Outline
1. Make excellent mistakes
2. Speed wins
3. Boyd's law
4. If you can't measure it, you can't fix it
5. Distribution
6. Sharing
Boyd's Law of Iteration: "Speed of iteration beats quality of iteration."
Jeff Atwood [3]
Brikman corollary: "Speed of iteration improves quality of iteration."
Change the code See the result
Should be as close to 0 as possible
Rapid Prototyping
Demo
● jsfiddle● Firebug● Developer tools● apigee● JavaScript API Console● CodeKit
"A canvas or sketchbook serves as an external imagination."
Bret Victor [4]
"A programming language is for thinking of programs, not for expressing programs you've
already thought of."Paul Graham [5]
Leverage
● Always, always, always google first
● "There's an open source library for that"
● Many minds are better than one
● Bonus: learn from the code
Demo
● node.js modules● ruby toolbox● github● google fonts● lorem ipsum● placekitten
"You'll be lucky if you have a single original idea your entire life."
Mr. Patch [6]
Outline
1. Make excellent mistakes
2. Speed wins
3. Boyd's law
4. If you can't measure it, you can't fix it
5. Distribution
6. Sharing
"If you can't measure it, you can't fix it."
DJ Patil
You will not get it right the first time. Or the 10th.
Or ever, if you don't know what is right and what is wrong.
inGraphs
Server metrics
inGraphs
Database metrics
inGraphs
Client-side metrics
Profile metrics
Activity metrics
Bug metrics
Build metrics
Test metrics
Measure everything.
LiXLinkedIn Experimentation Platform
LiX
● Group members into buckets
● Control feature visibility by bucket
● A/B testing
● Feature ramps: dark launch, 1%, 10%, 100%
Subscriptions
Bucket A
Subscriptions
Bucket B
Subscriptions
Bucket C
Subscriptions
Bucket D
Picking the right image for a given user significantly increased CTR
"Web 3.0 will be about data."Reid Hoffman
Outline
1. Make excellent mistakes
2. Speed wins
3. Boyd's law
4. If you can't measure it, you can't fix it
5. Distribution
6. Sharing
Products do NOT sell themselves.
"Poor distribution - not product - is the number one cause of failure."
Peter Thiel [7]
All credit goes to Peter Thiel [7] and Adam Nash [8]
Common distribution channels
Sales
● Dedicated sales team: LinkedIn
● Employees as sales team: Square
● CEO as sales team: Palantir [7]
● Competitions, grants: SpaceX [7]
● No one likes spam, but email can work: personalized, universal, measurable.
● Direct marketing emails: very low CTR
● Product emails: built into to product interactions. Far better CTRs.
Year in Review Email
CTRs so high, it was "clicks-per-email"! [8]
Facebook photo tagging email
My personal CTR on these is 100%.
SEO
● Free, measurable, effective
● LinkedIn: public profiles, public groups, LinkedIn Today, LinkedIn influencers
● TripAdvisor: the vast majority of TripAdvisor traffic comes from Google
TripAdvisor SEO
One of the top search results for almost any hotel
LinkedIn SEO
"Professional profile of record"
Social
● Put it in the stream
● Social gestures: comments, likes, votes
● Just like email spam, the feed is becoming flooded
Media
● Blogs, news sites: techcrunch, mashable, huffpost, wired, and many others
● Newspapers, magazines, TV
● Huge reach and a great way to bring in many users in a short time
Build relationships with the media
Marketing
● Traditional ads: TV, print ads, billboards. Expensive, hard to measure, and rarely used by startups.
● Online ads: CTR, CPM, SEM. Measure everything!
Viral Marketing
● Your users attract more users
● Game theory, psychology, sociology
● Free, often measurable (tracking codes)
Viral loops must be built into the product.
Key Viral Marketing Concepts
● Viral loops: in my product, where do users interact with non-users?
● Viral factor: how many non-users do they reach per viral interaction?
● Cycle time: how often do these viral interactions occur?
People You May Know (PYMK)
Invite notification shows up via email, web UI, and mobile. When the other member accepts the invite, we show them PYMK too!
Twitter @mentions
Notification shows up via email, web UI, and mobile
Google Analytics & Webmaster Tools
Essential tools for measuring the performance of your distribution channels (measure everything!)
The best product doesn't always win.
The best distribution wins
● Betamax vs. VHS [11]
● Apple vs. Microsoft (early 90's) [12]
● TiVo vs. DVR [13]
● Netscape vs. IE [22]
Outline
1. Make excellent mistakes
2. Speed wins
3. Boyd's law
4. If you can't measure it, you can't fix it
5. Distribution
6. Sharing
Silicon Valley companies share (almost) everything.
Share Everything
● Papers: MapReduce [15], Dynamo [16]
● Open Source: Hadoop [17], Voldemort [18]
● Hardware: OpenCompute [19]
● Companies: Red Hat [20], Mozilla [21]
Open Source
SpeakingWriting
Why?
Mastery: the best way to learn, is to teach.
Better code & products: the house is cleanest just before guests arrive
Free labor: free QA, free documentation, free bug fixes, ...
Publicity: recognition for the company and its employees
Happier employees: something they keep forever
"If I have seen further it is by standing on the shoulders of giants."
Isaac Newton
This is why I'm here today.
Please share your feedback and experiences with me!
Recap
1. Make excellent mistakes
2. Speed wins
3. Boyd's law
4. If you can't measure it, you can't fix it
5. Distribution
6. Sharing
Questions?
References1. http://www.amazon.com/Adventures-Johnny-Bunko-Career-Guide/dp/1594482918?ie=UTF8&s=books&qid=1206925356&sr=8-1
2. http://www.bizstone.com/2010/09/timing-lessons.html
3. http://www.codinghorror.com/blog/2007/02/boyds-law-of-iteration.html
4. http://worrydream.com/LearnableProgramming/
5. http://www.paulgraham.com/hp.html
6. Mr. Patch was my 11th grade physics teacher
7. http://blakemasters.com/post/22405055017/peter-thiels-cs183-startup-class-9-notes-essay
8. http://blog.adamnash.com/2012/03/28/user-acquisition-five-sources-of-traffic/
9. http://blog.adamnash.com/2012/04/04/user-acquisition-viral-factor-basics/
10. http://www.linkedin.com/today/post/article/20121002124206-18876785-how-to-model-viral-growth-the-hybrid-model
11. http://en.wikipedia.org/wiki/Videotape_format_war
12. http://www.roughlydrafted.com/RD/Q4.06/9FD12E37-8DC7-4AD1-872F-2021BEDE6D96.html
13. http://www.quora.com/TiVo/Why-did-TiVo-fail
14. http://www.paulgraham.com/growth.html
15. http://research.google.com/archive/mapreduce.html
16. http://www.read.seas.harvard.edu/~kohler/class/cs239-w08/decandia07dynamo.pdf
17. http://hadoop.apache.org/
18. http://www.project-voldemort.com/voldemort/
19. http://www.opencompute.org/
20. http://www.redhat.com/
21. http://www.mozilla.org/en-US/
22. http://en.wikipedia.org/wiki/Browser_wars