Lecture 6: Growth - Alex Schultz (Facebook VP of Growth)
Introduction: From Physics to Growth
I studied physics at Cambridge and paid for college through online marketing, starting with SEO in the 1990s. I created a paper airplane site (learning: check if your market is big enough) and later a cocktail site that became the UK's largest. When Google AdWords launched, I bought paid clicks and resold them to eBay's affiliate program for about 20% margin. That's how I transitioned from physics to marketing.
The #1 Most Important Thing for Growth: RETENTION
Q: What matters most for growth?
A: Retention. Not virality, not marketing tactics - retention.
At Facebook, we have a fantastic product. If we get people on and ramped up, they stick. This is the foundation everything else is built on.
How to Measure Retention: The Retention Curve
Look at your percent monthly active vs. number of days from acquisition.
- If your retention curve becomes asymptotic (flattens out parallel to the x-axis), you have a viable business and product-market fit
- If your curve slopes down toward zero, you don't have product-market fit yet - no matter how good your growth tactics are
Early-stage methodology: You only need 10,000 customers to see this pattern. For each cohort day (day 1, day 31, day 32, etc.), calculate what percentage of users are still monthly active. The curve will get noisy on the right side, but you'll see whether it flattens.
Real Example: Facebook's B2B Product
When I joined Facebook, we used this methodology on a product that was only 3 days old. Within 90 days of launch, we predicted the one-year value of an advertiser to 97% accuracy.
What Does "Good Retention" Look Like?
People always ask this. Here's how to figure it out yourself (dimensional reasoning from physics):
Facebook example:
- ~2.4 billion people on internet
- Minus China (banned) = ~2 billion
- Facebook has 1.3 billion active users
- Do the math: 65%+ retention
The answer varies by vertical:
- E-commerce: 20-30% monthly retention = very good
- Social media: If your first users aren't 80%+ retained, you won't have a massive social network
Look at comparable companies in your vertical. How many people use Amazon? WhatsApp? Calculate their retention as a benchmark.
Key Principle
Don't focus on growth tactics until you have retention. I've seen this as the #1 problem inside Facebook for new products and for startups - they think they have product-market fit when they don't.
Operating for Growth (Once You Have Retention)
My contrarian view: Startups shouldn't have growth teams. The whole company should be the growth team, with the CEO as head of growth.
1. Define Your North Star Metric
Every company needs ONE metric that the entire company optimizes for.
Examples:
- Facebook: Monthly Active Users (not registered users - Mark made this the internal and external metric)
- WhatsApp: Number of sends
- Airbnb: Nights booked
- eBay: Gross Merchandise Volume (not revenue)
Why this matters: Once you have more than one person, you can't control what everyone does - only influence them. When someone is working on a feature and you're not in the room, they need clarity on what matters. Is it revenue? Registrations? Active users? Pick one.
Real example from eBay (2004): We changed affiliate payments from "Confirmed Registered Users" (CRU) to "Activated Confirmed Registered Users" (ACRU - users who registered AND bid/bought/listed).
Result:
- CRUs dropped 20% overnight
- ACRUs only dropped 5%
- ACRU growth massively accelerated
Why? When optimizing for CRUs, affiliates landed people on registration pages. When optimizing for ACRUs, they landed people on search results (e.g., trampolines), so users saw what they wanted, got excited, then registered. Better magic moment.
2. Get Users to the Magic Moment
What's the magic moment? The moment users understand your value and get hooked.
Facebook's magic moment: Seeing your friend's face. That's why we focus on "10 friends in 14 days." Without friends, you have an empty newsfeed, no notifications, nothing bringing you back.
Look at any successful social product - LinkedIn, Twitter, WhatsApp - they all focus on connecting you to people as quickly as possible.
Other examples:
- eBay: Finding that unique collectible PEZ dispenser you've been searching for
- Airbnb: Finding that cool house to stay in (as a guest) or getting your first payment (as a host)
Get people to this moment as fast as possible. This can move your retention from 60% to 70%.
3. Optimize for the Marginal User, Not Yourself
Everyone in Silicon Valley gets this wrong.
Example: Notifications
Every company says "I'm getting too many notifications, we need to reduce them." But ask yourself: Are your power users leaving because of too many notifications? No? Then why optimize for that?
Focus on the marginal user - the person who ISN'T getting notifications and ISN'T coming back.
Facebook's growth accounting framework:
- New users
- Resurrected users (absent 30 days, came back)
- Churned users
Resurrected and churned users dominate new users after a few years. These users had low friend counts and hadn't found their friends. So we focused on getting THEM to their magic moment, not on making life better for power users who were already engaged.
Growth Tactics
Internationalization: Breaking Down Barriers
Facebook internationalized too late. Sheryl Sandberg has said this publicly.
We faced clones everywhere: Fakebook (literally had "Fakebook.css" in their HTML), VK, Mixi, Cyworld, Orkut. These grew when Facebook focused only on the US.
Facebook's barriers we had to break:
- College-only → High schools (company-shaking decision)
- High schools → Everyone (this got us to 50 million)
- Then we hit a wall at 50 million
At 50-100 million users, people inside Facebook questioned if ANY social network could get past 100 million. No one had ever done it. That's when the growth team formed.
What we did right with internationalization:
We moved slow to move fast. We built scalable infrastructure:
- Created FBT (translation extraction script)
- Built community translation platform
- French translated in 12 hours by community
- 104 languages total; 70 translated by community
We prioritized for the future, not the present:
Then (2007-2008): French, Italian, German, Spanish were top languages
Now: Italian isn't even on the list. In the last year we quadrupled Hindi users.
Build for where the world is going, not where it is today.
Virality: Two Models
Model 1: Sean Parker's Framework
Virality = Payload × Conversion Rate × Frequency
Hotmail example:
- Payload: Low (email one person at a time)
- Frequency: High (email same people repeatedly)
- Conversion Rate: High (people wanted free email not tied to ISP)
- Tactic: "Sent from Hotmail. Get your free email here" at bottom of every email
- Result: Extremely viral
PayPal example:
- Seller side: High conversion (someone wants to send you money!), low payload/frequency
- Buyer side: Gave away money for referrals ("Sign up and get $10")
- Result: Viral on both sides through high conversion
Facebook reality: Facebook was NOT viral through email or sharing. It was word-of-mouth viral because it was an awesome product. There's no native way to contact people not on Facebook from within Facebook.
Model 2: K-Factor (Ed's Model)
Contact Import → % who send invites → # of recipients → % who click → % who sign up → % who import
Multiply all these percentages to get your K-factor. If K < 1.0, you're not viral.
Critical point: Even if you get K > 1.0 (like Viddy did), if you don't have retention on the backend, it doesn't matter. Only focus on virality AFTER you have strong retention.
SEO: Three Key Things
1. Keyword Research (Do this FIRST)
I spent a year optimizing my cocktail site for "cocktail making" - 500 searches/month in UK. Everyone actually searches for "cocktail recipes" (UK) or "drink recipes" (US). I optimized for the wrong word.
Research: Supply (who else ranks), Demand (search volume), Value (worth to you)
Best tool: Still Google AdWords Keyword Planner
2. Links (Authority)
PageRank drives SEO. You need valuable links from high-authority websites.
Facebook SEO story (2007): We launched public profiles but got no traffic. Why? The only way Google could reach profiles was: Click footer "About" → Click blog → Click author → Spider through friends.
Google saw these pages as low-value because they were buried.
One change: Added a directory so Google could quickly reach every page.
Result: 100X'd SEO traffic.
3. Table Stakes
XML sitemaps, proper headers, etc. - all covered well online.
Email, SMS, and Push Notifications
Major point: Email is dead for people under 25. Young people use WhatsApp, SMS, SnapChat, Facebook - not email. If targeting older audiences, email still works.
Universal principles for Email/SMS/Push:
1. Deliverability First
"To finish first, first you have to finish." Your message must reach the inbox.
Avoid:
- Sending spam (dirty IPs)
- Shared servers with spammers
- Abusing inboxes
Once you're in spam folders or blocklists, it's nearly impossible to get out. Same for SMS grey routes - they always get shut down.
If users block your push notifications, you can never reach them again.
2. Focus on Notifications, Not Newsletters
Newsletters are stupid. Why send the same message to someone who signed up yesterday and someone who's used your product for 3 years?
Best approach: Notifications of real events
Facebook example: As a power user, I don't want an email for every like. But for a new user, that first like is a magic moment.
We turned on more email/SMS/push notifications, but ONLY for low-engagement users not returning to the site. Result: Increased engagement without spamming.
3. Triggered Campaigns
Context matters. At eBay, one of our best email campaigns was triggered when someone completed their first cross-border transaction. High open and click rates because it was timely and relevant.
Focus on:
- Deliverability
- Notifications (not newsletters)
- Triggered, contextual messages
Closing Principles
"A good plan, violently executed today, is better than a perfect plan tomorrow." - General Patton
Move Fast and Break Things
If you can run more experiments than the next company, if you're hungry for growth, if you'll fight for every extra user and stay up late to get the data and run the experiments over and over - you will grow faster.
Mark Zuckerberg: "I think we won because we wanted it more."
We're not crazy smart or experienced. We just worked really hard and executed fast.
Growth is optional. You have to choose to work for it.
Key Takeaways
- Retention is everything - Don't do growth tactics until your retention curve flattens
- Pick your North Star - One metric the whole company optimizes for
- Find the magic moment - Get users there as fast as possible
- Optimize for marginal users - Not yourself or power users
- Execute relentlessly - Run more experiments, work harder, move faster
