Blitzscaling's Big Shift: Village Stage


The Presentation inside:

Slide 0

CS183C: Blitzscaling Instructors: Allen Blue, Reid Hoffman, John Lilly & Chris Yeh October 20, 2015


Slide 1

Organizational Scale 3: Village


Slide 2

Now for the big shift.


Slide 3

Class Structure: Organizational Scale Org Scale User Scale Customer Business (employees) (B2C users) Scale (B2B) Scale (rev) OS1: Family 1s 10,000s 0 <$10M OS2: Tribe 10s 100,000s 1s $10M+ O3: Village 100s 1,000,000s 10s $100M+ OS4: City 1,000s 10,000,000s 100s $1B+ OS5: Nation 10,000s 100,000,00+ 1,000+ $5B+


Slide 4

OS1: The Household Identify a non-obvious market opportunity where you have a unique advantage and/or approach. Iteratively build a product with strong product/ market fit


Slide 5

OS2: The Tribe Execute & iteratively improve a plan which gets you to significant market share.


Slide 6

OS3: The Village Identify, plan and execute the core business you will scale up.


Slide 7

OS3: The Village Identify, plan and execute the core business you will scale up. … with a plan to scale more


Slide 8

OS3: The Village Identify, plan and execute the core business you will scale up. … with a plan to scale more … and do it fast


Slide 9

OS3: The Village 1. Articulate the core business you will scale up first 2. Identify critical steps to scale that business 3. (Hyper)grow an organization to execute it 4. Finance for hypergrowth and future growth


Slide 10

OS3: Articulating the Core Business Goals for the core business: 1. Create continued growth 2. Generate growing revenue 3. Build competitive advantage 4. Grow strategic assets for later opportunities


Slide 11

LinkedIn All-Hands February 17, 2010


Slide 12


Slide 13


Slide 14


Slide 15


Slide 16


Slide 17


Slide 18


Slide 19


Slide 20


Slide 21


Slide 22


Slide 23


Slide 24


Slide 25

OS3: Critical Components to Scale the Core Business • What must you do to scale the core business? • • • • Across growth, revenue Product and Technology Go-to-market (sales, marketing) Partnerships


Slide 26

Product Plan February, 2010


Slide 27

Strategic Stack Table stakes Game changers Value Data Global Monetization Leverage unique business model to monetize assets while adding value to members on a global basis Example products: Online Jobs, Recruiter, Premium subs, Display Ads, Self-serve Ads, Business Pages Professional Insights & Knowledge Sharing Be the essential source for professional shared knowledge and business intelligence Example products: NUS, Sharing tools, Groups V2, Inapps, Outlook & Twitter Integration Professional Identity ecosystem Establish leading identity ecosystem by building upon unique database of 60M+ members Example products: Profile improvements/game dynamics, SEO, Search, APIs, Mobile Viral Growth Engine Continue viral engine improvements to drive member growth and critical mass within professional context Example products: Reg. optimization, New User Experience, In-Box, Outlook Integration, SEO, Address Book Firm Foundation Improve and scale site resilience & reliability, development productivity, data reliability, API’s Measurements: uptime, load time, developer activity, Underlined: New focus in 2010 Targeting & analytics WVMP, MyStats Matching PYMK Warehousing Mining


Slide 28

Strategic Stack Core Strategic Security Code Modularity Productivity Tools Distributed Computing Data Reliability / Scale Feature APIs Reg Optimization / NUX PYMK SEO Address Book Profile - Data Inbox / Comm Core Search NUS / Sharing Groups Standardization A/B Testing Platform Online Jobs Subs Recruiter Payment Platforms DirectAds Display Support CS Tools Addressbook bcard Profile - Engagement APIs Rec Engine / Intelligence Engine / WVMP Mobile Business Pages International - Languages Venture Skills Apps / InApps Career Center Salary


Slide 29


Slide 30


Slide 31


Slide 32


Slide 33


Slide 34


Slide 35


Slide 36


Slide 37


Slide 38


Slide 39

Talent Solutions GTM March, 2010


Slide 40


Slide 41


Slide 42


Slide 43


Slide 44

LinkedIn 3.0 / Scaling 10x February, 2010


Slide 45


Slide 46


Slide 47


Slide 48


Slide 49


Slide 50


Slide 51

OS3: Organization • • • • • • • • Your organization has to change fundamentally The right CEO Key executives for critical areas Core mission, culture, and values to enable rapid distributed scaling An HR function that can hypergrow Necessary processes to allow large groups to work together Navigate necessary changes among founders and early employees Robust reporting to allow business not only to learn, but also to plan


Slide 52

Leadership, Culture, Hiring, Process 2008-2010


Slide 53


Slide 54


Slide 55


Slide 56


Slide 57


Slide 58


Slide 59

LinkedIn Values and Culture Members Come First - Candidates are also our members, leave them feeling great about LinkedIn regardless of our decision to hire or not. Be compassionate, offer to take a candidate for a quick walk, for a break, offer snack/beverages if they look tired. Relationships Matter - Candidates will either advocate us, say nothing about us, or speak negatively about us based on how we treat them. Glassdoor.com and other sites/blogs keep all companies on their toes in terms of reputation and sharing stories. Open, Honest and Constructive - consider this as you answer the candidate’s questions and as you write your own feedback after the interview. Demand Excellence - consider this not only of the candidate but yourself regarding how you conduct the interview... don’t focus too much on technical/analytical skills so they overshadow your assessment of the candidate’s creativity and vision. Transformation of world, company and self Integrity Collaboration Humor Results Take Intelligent Risks - If the candidate doesn’t have the experience in the talent solutions space or social media, or if you initially rate as B + rather than A+, as yourself if they have the raw materials and potential to become an A+ within a short timeframe. Act Like an Owner - Talent is our most important asset. We expect our employees to act as an owner in each decision they make, no matter how big or small. © 2009 LinkedIn All rights reserved. Confidential 60


Slide 60

Tech All-Hands December, 2009


Slide 61

People/Recruiting/Hiring ● Best people produce best outcomes ● Need to keep bar high in the hiring/recruiting process ● Need to understand our limitations and bring outside expertise where needed ● Need to be flexible in our resource allocation (it is okay/desirable to move around the organization to apply key personnel to key projects) ● Onboarding/Ongoing Training/Mentoring are key.


Slide 62

Why technologists want to work at Linkedin ● Having a positive/lasting impact on the world by developing products  that create economic opportunities for people on a global basis. ● Building systems that scale and perform is paramount (think 10X). ● Data drives our solutions: Searching, querying, analysis of this data in real-time, nearreal time, or batch creates value for our users, and our paying customers. ● The business models (Subscription, Corporations, Advertising) are a diverse source of revenue that require real value propositions, and technical excellence in delivery. ● Combining speed of execution with quality solutions, while still pushing the envelope on new/improved features requires engineering skill sets that are unheard of in traditional enterprise computing, and are hard to find even in the Internet. ● Engineers will constantly learn/use new technologies to create leading edge solutions. ● Great engineers want to work with great engineers. ● Constant improvement requires new talent/additional expertise to solve new/difficult challenges. ● When you come to work here, you will, by definition, have a large impact. ~ 150 technologists today ● Access to the senior leadership is real, actual, and actionable. CEO->VP->Director>Manager->Engineer


Slide 63

Site Up -> Company Up


Slide 64

What do we want?
 ● Scrum vs. Waterfall: not the question; want best quality in timely fashion; good outcomes depend on good engineers + good engineering practices (AND function required here) ● Prioritize the list of Infrastructure + Product features merge sorted: We apply resources against that list, and the result is 2 clear choices, either add more people or cut the line higher so that some things don’t get done (ideally there is no distinction between features and infrastructure in our shop) ● Will have multiple serving data centers running: ETA likely > = 1 year ● Architecture: we will tease apart dependencies for our 180 subsystems so that we can have SLA’s for each one that we can trust and adhere to. All calling services must be resilient to failure of called services. (Service oriented architecture) ● The Data scaling concern demands a solution: Examples: Comm, DWH, volume of members, search index. ● Reminder: We have scaled this most excellent service successfully to this point. Now we need to go to the next level. Think 10X and 100X. How would we solve for that?


Slide 65

Current Release Process



Slide 66

Release Process Gap – Changes Needed
 ● The Current Process – What we do today ● Testing earlier for features: Example: the 26 feature branches in R951. ● Stabilizing the code base for integration ● Practice on staging ● Deploy a well tested reliable solution on prod ● Concern with this model ● It’s a long cycle (see previous slide) 2.5 week duration due to merges, compatibility testing, configuration validation, heavy track team involvement etc… ● We still have large releases with many dependencies (some of which are not understood) which increases our risk ● Inconsistent tooling across environments (Do we have enough envs?) ● Many release processes are still manual ● Post release hangover that says fix on fail: Bug fix releases are an assumed part of the Release Process. Our goal should be to eliminate the need for this step. ● What do we want? ● Shorter cycle ● Smaller releases with “no” dependencies decreasing the risk. Each component released must be able to be pushed and rolled back if needed. ● Automation for push-button releases (What if we were deploying to 5,000 servers not 500?) ● Don’t need to be fixing on fail because there are no bugs introduced into Prod ● Release when ready: (What does this mean?) a) Don’t release it on schedule because it didn’t pass our tests OR b) Release any module anytime with low risk. I want B.


Slide 67

LinkedIn Series D Summer, 2008


Slide 68

Global Internet Business Platform and Applications


Slide 69

Investment Snapshot 2008E Rev $82 million / 2.6x 2007 Members 20 million, adding 1M+ per month at ~ $0 CPA Unique visitors1 6.6 million / 2.8x y-o-y Email addresses 384 million Connections 260 million Bookings ($MM) Investors 1 Subscriptions Jobs Advertising Corporate Total The next massive business model in technology ▪ ~ $0 CPA per member, 1M+ / month ▪ High margin product: all digital goods, micro cost of sale ▪ Highly scalable: digital goods, infinitely replicable 2006 $6.3 2.1 2.1 1.8 $12.4 2007 $17.3 6.3 7.8 7.9 $39.2 2008 $33.0 15.1 30.1 34.5 $112.6 Sequoia, Greylock, Bessemer, Marc Andreessen, Peter Thiel ▪ Network effects: network between users, network between business lines ▪ Huge markets: recruiting, media, services, sales, productivity software and others ➢ Google for finding professionals ➢ eBay for Labor Markets ➢ Microsoft for Internet productivity comScore January 2008 data
 LinkedIn - Confidential 2


Slide 70

Summary The Network Key Differentiation: - Business focus: features, brand, network - Viral growth: entirely by individuals’ actions - Value scales with entire network (network effects) - Growing in every industry, globally - Organic growth into every business LinkedIn - Confidential 5


Slide 71

Largest Professional Network Domestic 20 International Growth Days 0 to 1MM members 1 to 2MM members Millions of members 181 5 to 6MM members 102 9 to 10MM members 60 18 to 19MM members 15 477 28 10 199 of top 200 markets grew 70%+ in 2007; 155 grew 100%+ 5 Domestic Globally x4 larger than nearest competitor; 50x larger in the U.S. 0 5/1/03 9/1/03 1/1/04 5/1/04 9/1/04 1/1/05 5/1/05 9/1/05 1/1/06 5/1/06 9/1/06 1/1/07 5/1/07 9/1/07 2/1/08 LinkedIn - Confidential 6


Slide 72

Best and Broad Demographics 13M University Alumni Demographics Average Age 41 1.9M F500 Employees HHI >$150K College Grad Portfolio $250K+ School: 50K GSB: 8K Employees: 58K Alumni: 23K 31K 13K $109,762 Employees: 116K Alumni: 71K 32K Average HHI School: 58K HBS: 17K Employees: 13K Alumni: 9K 13K 19K 19K Employees: 15K Alumni: 12K 27% 78% 28% 1.2M Small Business Owners
 2.2M Senior Executives
 VPs at every F500 company Source: @plan Winter 2007/2008, internal data LinkedIn - Confidential 7


Slide 73

Summary Media Key Differentiation: - User generated content - Best of class demographics - Unique targeting capabilities - Organic growth in every industry, globally - Ability to scale across the web - Future possibilities with self-service, B2B lead gen LinkedIn - Confidential 11


Slide 74

LinkedIn Revenue Upside Revenue/worldwide unique visitors $100$100 LinkedIn potential revenue $1,000 $86.91 ($34.05 Ave Rev/User) x25.0MM Uniques) $800 37.5$100 Mean:
 $34.05 $600 ($34.05 Ave Rev/User) x13.4MM Uniques) 25$100 Mean:
 $21.08 $400 12.5$100 Mean: $10.52 ($34.05 Ave Rev/User) x 5.5MM Uniques) $200 0$100 Dice Monster HotJobs eBay Google Xing $0 2007 Job boards LinkedIn - Confidential Internet 
 leaders Business/ 
 social 
 networking 2008E 2009E Upside revenue opportunity Current revenue projections 26


Slide 75

Appendix


Slide 76

Articulating biz


Slide 77

The Bullseye Jeff’s articulation of LinkedIn’s bullseye: Talent Solutions Description of the business: Passive Recruiting and how it works. Main point is the clarity with which we chose this target, and were willing to subordinate almost everything but user growth to it.


Slide 78

User Growth User Growth understood as: Viral tuning SEO and Public Profiles Some known user value (network updates via email, receiving job offers) Some user value experiments (news, app platform, etc.)


Slide 79

Revenue Growth Main focus on revenue growth will be growth of field sales Reasoning for this Some online sales, but not for big-$$ accounts Other revenue sources beginning to enter the mix, but not core


Slide 80

Strategic Moats Professional Identity would be primary moat Defined roughly as size of network x quality of average profile Main contributors were user growth and user activity. Also first-mover advantage on passive recruiting would allow us to lock in customers. But we also felt we needed to respond to Facebook’s F8 announcement, which might have created a competitor built on top of FB.


Slide 81

Future strategic assets Non-Jobs brand Potential discover of additional value propositions improves brand, Drives growth and profile moats builds additional platforms for future businesses


Slide 82

Summary A growing network of professionals with valuable profiles, driven by viral tuning and value experimentation Will form the foundation for a sales-driven business focused on large enterprise accounts… And we will do some competitive blocking on App Platforms.


Slide 83

Key steps to succed


Slide 84


Slide 85


Slide 86

Changing roles for founders/early team Jean-Luc, engineering changes New engineering group leaders Turnover with new execs


Slide 87

Data Building our data function


Slide 88

Daily Executive Dashboard Signups Guest Invites 112,500 600,000 90,000 Comments Uniques 500,000 3,200,000 67,500 300,000 45,000 1,600,000 200,000 22,500 100,000 0 120% 0 300% 80% 8/ 09 1 6/ 7/ 09 6/ 2 6/ 09 09 09 4/ 5/ 7/ 1 7/ Daily Year over Year 3/ 2 7/ 09 1/ 8/ 09 1 8/ 09 0/ 09 9/ 1 8/ 7-Day Cumulative Avg 8/ 2 8/ 09 9/ 09 6/ 09 0% -40% 5/ 1 9/ 160% 8/20/2009 9:05:35 PM 100% 0% 0 200% 40% 6/ 9/9/2009 4:10:27 PM 400,000 09 09 7/ 8/ 6/ 1 6/ Week over Avg 4 Weeks 09 6/ 2 6/ 5/ 7/ 09 09 4/ 1 7/ Daily Year over Year Page Views 3/ 2 7/ 09 8/ 09 09 / 10 8/ 1/ 09 9 7-Day Cumulative Avg 9/ 2 8/ 0% -100% 09 5/ 09 6/ 0 8/ / 19 8/ 80% 09 8/ 6/ 1 9/ Week over Avg 4 Weeks 9 9 0 7/ 09 /0 26 6/ 1 6/ 5/ 7/ 9 /0 14 7/ Daily Year over Year Searches 9 /0 23 7/ 9 0 1/ 8/ 9 /0 10 8/ 09 9/ 1 8/ 7-Day Cumulative Avg 9 /0 28 8/ 09 09 -80% / 15 9/ 6/ 9/ Week over Avg 4 Weeks 8/19/2009 7:34:57 PM Premium Subs (New + Recurring) 50,000,000 5,000,000 $160,000 40,000,000 4,000,000 $120,000 30,000,000 3,000,000 20,000,000 2,000,000 8/20/2009 6:13:53 PM 10,000,000 $40,000 1,000,000 0 300% 8/19/2009 6:54:30 PM $80,000 0 320% 200% $0 150% 160% 80% 0% 0% 100% 0% 6/ 8/ 09 6/ 17 /0 9 6/ 9 /0 26 7/ 5/ 09 7/ 14 Daily Year over Year /0 9 7/ 23 /0 9 8/ 1/ 09 10 8/ 9 /0 9 /0 19 8/ 7-Day Cumulative Avg 9 /0 28 8/ 6/ 9/ 09 9 -100% /0 15 9/ 8/ 6/ 09 17 6/ Week over Avg 4 Weeks 9 /0 9 /0 26 6/ 7/ 09 5/ 7/ Daily Year over Year Members Joining Groups 9 /0 14 7/ /0 23 9 8/ 9 09 /0 10 8/ 1/ 9 /0 19 9 /0 28 8/ 7-Day Cumulative Avg 9 /0 6 9/ 8/ 09 5/ -160% Week over Avg 4 Weeks 09 1/ 8/ 9 /0 10 8/ 7-Day Cumulative Avg 9 /0 19 8/ 09 / 28 8/ 9 0 6/ 9/ 5/ 1 9/ Week over Avg 4 Weeks $20,000 500 0 09 3/ 2 7/ $30,000 1,000 320% 09 4/ 1 7/ $40,000 1,500 40,000 09 5/ 7/ $10,000 0 300% $0 150% 200% 240% 160% 80% 100% 80% 0% 0% 0% 8/ 6/ 09 09 7/ 1 6/ 6/ 2 6/ 09 Daily Year over Year 09 5/ 7/ 09 4/ 1 7/ 09 3/ 2 7/ 8/ 1/ 09 09 0/ 1 8/ 7-Day Cumulative Avg Metrics: 09 9/ 1 8/ 8/ 2 8/ 09 6/ 9/ 09 09 5/ 1 9/ Week over Avg 4 Weeks 8/11/2009 8:42:25 PM $50,000 2,000 80,000 9 /0 26 6/ 8/19/2009 6:53:46 PM Job Dollars 2,500 120,000 09 / 17 6/ Daily Year over Year Groups Created 160,000 09 8/ 6/ 1 9/ -80% 09 8/19/2009 6:54:12 PM W/W jobs increase driven by 10% more spend per buyer due to job pack purchases and 8% better conversion For Jobs on Wednesday, 8/19: Conversion decreased 9% W/W but spend per buyer increased 6% and traffic to the flow increased 3%. We’re investigating the root cause of the conversion decreases that we’ve been seeing. New subs bookings growth driven by greater number of annual plans purchased relative to monthly plans (Business YR up 24% wk/wk and Business Plus YR up 18% wk/wk) “W/W declines driven by drops in conversion and spend per buyer. All hands are on deck investigating if something in last week's release or Saturday's Oracle upgrade is causing this.” Online Job sales were down –17% Y/Y at $39.7K but are trending towards being up Y/Y in the next 3 weeks. Page Views came in at 45.8M for Tuesday, the largest day since 27th January and the second highest day ever, up 78% Y/Y. Signups came in at 108.5K for Tuesday, the largest day since 27th January. The spike in the year over year numbers towards the end of May 09 is due to a site outage in May 08 -80% 09 8/ 6/ 6/ 9 /0 15 9 /0 22 6/ Daily Year over Year 9 /0 29 6/ 7/ 09 6/ 9 /0 13 7/ 9 /0 20 7/ 09 7/ 2 7/ 9 0 3/ 8/ 9 0 0/ 1 8/ 7-Day Cumulative Avg 9 0 7/ 1 8/ 09 / 24 8/ 9 0 1/ 3 8/ 9 /0 7 9/ 9 0 4/ -100% 1 9/ Week over Avg 4 Weeks Daily, 7-Day Cumulative Avg, Week over Avg 4 Weeks, Year over Year 09 8/ 6/ 9 /0 17 6/ 9 /0 26 6/ Daily Year over Year 9 /0 5 7/ 9 /0 14 7/ 09 / 23 7/ 09 1/ 8/ 09 / 10 8/ 7-Day Cumulative Avg 9 /0 19 8/ 09 8/ 2 8/ 09 6/ 9/ 09 -80% / 15 9/ Week over Avg 4 Weeks Add a Comment


Slide 89


Slide 90


×

HTML:





Ссылка: