Lecture 6 - Growth (Alex Schultz)

Alex Schultz, Facebook's VP of Growth, teaches that retention is the foundation of all sustainable startup growth.

Retention First

• Without a retention curve that flattens, virality and paid acquisition are wasted (0:00-10:03).
• Product-market fit exists only when enough users keep coming back.

North Star Metric

• The CEO must set one company-wide metric (e.g., Facebook's MAUs, Airbnb's nights booked) so every team optimizes the same goal (10:02-20:04).
• Focus on ACRU (activated confirmed registered users) over mere sign-ups.

Magic Moment

• Identify the instant a user feels core value and engineer the product to deliver it in the first session (Facebook: 10 friends in 14 days).

Growth Accounting & Channels

• Track new, resurrected, and churned users; resurrected users often dominate later growth (20:03-30:06).
• Apply virality models (payload, conversion, frequency) and SEO basics only after retention is solid (30:05-47:15).

0:00Introduction and Speaker Background

• Speaker: Alex Schultz (VP of Growth at Facebook) introduces himself, noting he's been watching the course lectures.
• 0:26 He contrasts his background with Paul Graham's, who suggested physics for college. Alex actually studied physics at Cambridge.
• 0:42 He views physics as providing valuable transferable skills.
• 0:50 Alex paid for college through online marketing, starting with SEO in the 1990s.
• 1:00 He created a paper airplane site, achieving a "monopoly" in a small niche. This taught him early SEO tactics, like using white text on a white background with keywords repeated many times to rank on AltaVista.
• 1:30 He then created a cocktail site, which became the largest in the UK, teaching him programming and more advanced SEO with Google (PageRank, backlinks).
• 2:00 When Google launched AdWords, he learned paid marketing, buying clicks and reselling them to eBay for a 20% margin via their affiliate program.
• 2:20 This experience was his entry into what is now called "growth," "growth hacking," or "growth marketing," which he simply calls "internet marketing" – using any channel to achieve desired output.

2:40The Most Important Factor for Growth: Retention

• Alex asks the audience what matters most for growth. The answer is a "great product."
• 2:50 A great product leads to customers, and those customers need to "spread the word" and "stay on your site."
• 3:00 ⭐ HIGH PRIORITY: Retention is the single most important thing for growth.
• 3:10 Facebook's growth team benefits from a fantastic product that people genuinely want to use.
• 3:30 Alex advises many startups (e.g., Airbnb, Coursera) and consistently sees the same pattern:

- 3:45 Visual Description: A chart showing "Active Users vs. Days from Acquisition." - If the retention curve asymptotes (flattens out) to a line parallel to the x-axis, the business is viable and has product-market fit for some market segment. - Most companies' retention curves slope down towards the x-axis, eventually intercepting it, meaning users churn completely.

• 4:35 This methodology was used for Facebook's B2B growth (self-service advertisers) even with a new product.
• 4:50 For Facebook's B2B product, which was only three days old when he joined, they used this technique to predict the one-year value of an advertiser to 97% accuracy within 90 days.
• 5:10 How to analyze your retention curve:

- 5:20 Visual Description: A graph with "Number of Users" on the y-axis and "Days from Acquisition" on the x-axis. A red line shows the number of users who have been on the product for a certain number of days. - 5:30 For a product that's been around a year, there will be zero users who have been on it for 366 days. - 5:50 Look at all users who have been on your product for one day. What percentage are monthly active? (100% for the first 30 days). - 6:10 Then, for each subsequent day (e.g., 31st, 32nd day after registration), what percentage of those users were monthly active? - 6:30 This allows you to understand the curve's shape with a relatively small number of customers (e.g., 10,000). - 6:45 You can determine if the curve flattens out (asymptotes) or continues to decline.

• 7:00 If the curve doesn't flatten out, do not focus on growth tactics, virality, or hiring a "growth hacker." Instead, focus on achieving product-market fit.
• 7:20 As Sam Altman said, "idea, product, team, execution." Without a great product, executing well on growth is pointless.
• 7:35 The number one problem Alex has seen in new Facebook products and startups he's advised is a lack of true product-market fit, even when they believe they have it.

7:50What Does Good Retention Look Like?

• People often ask what a "good" retention rate is, assuming Facebook's is exceptionally high.
• 8:05 Alex gets "pissed off" by this question because he believes it can be figured out.
• 8:15 Analogy: Geoffrey Taylor and the Atomic Bomb (Dimensional Reasoning)

- Story: In 1950, Life magazine published a picture of a Trinity nuclear bomb test. - Geoffrey Taylor, a British physicist and Nobel laureate, was able to calculate the power of the US atomic bomb (and compare it to Russian bombs) just by looking at the picture, using dimensional reasoning. - 8:50 Definition: Dimensional Reasoning: Analyzing a problem by looking at the fundamental dimensions (e.g., mass, length, time) involved and how they relate to the desired outcome (e.g., energy). - Taylor used the volume of the sphere (meters cubed) and time to deduce the energy (kilograms meters^2 seconds^-2). - This allowed him to reveal a top-secret capability.

• 9:40 Figuring out Facebook's retention rate is not as hard as that.
• 9:50 Example: Facebook's Retention Calculation (Ballpark)

- Internet users: ~2 billion (excluding China). - Facebook active users: ~1.3 billion. - Dividing these numbers gives a ballpark figure for Facebook's retention relative to the total addressable market.

• 10:10 Example: WhatsApp

- 600 million active users compared to the number of smartphone users globally.

• 10:20 Example: Amazon

- Number of online users in the US compared to Amazon's customer base.

• 10:30 Different verticals require different terminal retention rates for success.

- E-commerce: 20-30% monthly active retention might be good. - Social media: If initial users aren't ~80% retained, it's unlikely to become a massive social media site.

• 11:00 You need tools to compare your product to others in your vertical and assess if you're close to "real success."
• 11:15 Retention comes from a great idea and a great product that achieves product-market fit.
• 11:30 The best way to assess retention is by normalizing on a cohort basis – looking at how groups of users acquired at the same time are retained over time.

11:50Operating for Growth: The Company as the Growth Team

• Assuming you have product-market fit (e.g., an e-commerce site with 60% monthly retention).
• 12:10 Now it's time to scale and execute.
• 12:20 Contrarian Viewpoint: "If you're a start-up you shouldn't have a growth team."
• 12:30 The entire company should be the growth team.
• 12:40 The CEO should be the head of growth.

12:50The North Star Metric

• You need someone to set a North Star for the company's direction.
• 13:00 Example: Mark Zuckerberg at Facebook

- While other companies (MySpace, Friendster) focused on registered user numbers, Mark focused on monthly active users (MAU). - He held everyone internally accountable to MAU and published it externally. - This meant everyone focused on getting people active, not just signed up.

• 13:50 Example: Jan Koum at WhatsApp

- Published send numbers as the key metric, not just active users, because for a messaging app, message volume indicates primary usage.

• 14:10 Example: Airbnb

- Focuses on nights booked, benchmarking themselves against large hotel chains.

• 14:30 Example: eBay (Alex's experience)

- The North Star was Gross Merchandised Volume (GMV) – the total value of goods sold through the platform. - Externally, eBay was often judged by revenue, but GMV was the internal focus, which was much larger than revenue.

• 15:10 Every company needs a different North Star metric.
• 15:20 It's critical for leaders to define this North Star.
• 15:30 Why it matters: Once you have more than one person, you can't control everyone's actions; you can only influence them.
• 15:50 A clear North Star ensures that even when leaders aren't in the room, employees make decisions aligned with the company's most important goal (e.g., GMV over revenue for eBay).
• 16:30 Registrations don't matter unless they become long-term active users.

16:40The Magic Moment

• Example: eBay Affiliate Program (2004)

- Initially, eBay paid affiliates for confirmed registered users (CRUs). - They changed the model to pay for activated confirmed registered users (ACRUs) – users who registered and then bid, bought, or listed an item. - Overnight, CRUs dropped by 20%, but ACRUs only dropped by 5%. - The ratio of CRU to ACRU improved, and ACRU growth massively accelerated. - Reason: To drive CRUs, affiliates would land users on a registration page (e.g., for a trampoline search). To drive ACRUs, they landed users on the search results page for trampolines, allowing them to see the product and get excited before registering to buy. - The latter approach created a "magic moment" for the user.

• 18:00 ⭐ HIGH PRIORITY: How do you drive towards the magic moment that gets people hooked on your service?
• 18:20 Resources: Links to Danny Ferrante (retention curves), Chamath Palihapitiya (Facebook growth team founder), and Naomi Gleit (Facebook growth) on the lecture page discuss the magic moment.
• 18:50 Example: Facebook's Magic Moment

- "See your friends." It's often very simple. - Mark Zuckerberg focused on getting users to "ten friends in 14 days." - Without friends, the news feed is empty, notifications don't happen, and users don't return.

• 19:40 Other services like LinkedIn, Twitter, and WhatsApp also prioritize connecting users to people they know as quickly as possible.
• 20:00 Example: eBay's Magic Moment

- Finding that unique, desired item (e.g., a PEZ dispenser or broken laser pointer). - The first time you get paid for listing an item.

• 20:30 Example: Airbnb's Magic Moment

- Finding a cool house to stay in. - The first time you get paid for listing your house.

• 21:00 Brian Chesky (Airbnb co-founder) is excellent at describing this journey.
• 21:20 Identify your product's magic moment and connect users to it as fast as possible. This can significantly improve retention (e.g., from 50% to 60-70%).

21:40Optimize for the Marginal User, Not Yourself

• "Everyone in the Valley gets wrong that we optimize when we think about growth for ourselves."
• 21:50 Example: Notifications

- Companies often think, "I'm getting too many notifications, so we need to optimize for that." - However, power users (who might feel overwhelmed) are not the ones you need to worry about for growth. - 22:20 Building a great product optimizes for power users, but driving growth focuses on the marginal user – the one who isn't getting any notifications or isn't engaged.

• 22:50 Resource: Danny Ferrante's video also discusses Facebook's growth accounting framework.

- This framework tracks new users, resurrected users (came back after 30+ days), and churned users. - 23:10 For mature products, resurrected and churned users often dominate new user acquisition numbers. - 23:30 These churned/resurrected users often had low friend counts and weren't connected to Facebook's core value. - So, the focus was on getting them to "ten friends" or whatever number was needed.

• 23:50 ⭐ HIGH PRIORITY: Think about the user on the margin, not yourself, when considering growth.

24:10Summary of Operating for Growth

• Define your company's North Star metric.
• Ensure everyone in the company understands and drives towards this metric.
• Pick a metric that aligns with your mission and values (many metrics are correlated).
• Know the magic moment that hooks users and drives the North Star metric.
• Think about the marginal user, not yourself.
• "Everything has to come from the top."

25:10Growth Tactics

• Speaker: Alex Schultz introduces Tom Fishburne, a cartoonist known for cynical marketing cartoons.
• 25:40 Analogy: Silenced Mousetrap

- Imagine a niche market for a silenced mousetrap for under beds. - Even if your product is superior for that niche, the "build it and they will come" mantra is false.

• 26:10 Marketers are often seen as useless, especially by engineers.
• 26:20 Resource: An article interviewing Ben Silbermann (Pinterest CEO) on how Pinterest's growth was driven by marketing.

26:40Tactic 1: Internationalization

• 26:45 Facebook internationalized too late, leading to the rise of local clones (e.g., StudiVZ, Vkontakte, Mixi, Cyworld, Orkut).
• 27:15 Knocking down barriers is crucial for growth.

- Facebook started college-only, then expanded to high schools (a "company shaking moment"), then to everyone. - 27:50 When Facebook hit a "brick wall" at 50 million users, existential questions arose about whether any social network could exceed 100 million users. - This led to the formation of the growth team.

• 28:20 The two key drivers of initial growth were:

1. Focusing on "ten friends in 14 days" (the magic moment). 2. Internationalization (knocking down another barrier).

• 28:50 How Facebook approached internationalization:

- Even though they were late, they built it scalably ("moved slow to move fast"). - Resource: Naomi Gleit's video linked on the lecture page provides the full story. - They wrapped all strings in the site with a translation script (FPT). - They created a community translation platform, allowing users to translate the site. - French was translated in 12 hours. Today, Facebook supports 104 languages, with 80+ translated by the community. - This scalable infrastructure allowed them to adapt to future language needs.

• 30:00 They prioritized the right languages at the time (French, Italian, German, Spanish – FIGS).

- However, today, Italian is no longer a top language, and French/German are declining. - Hindi quadrupled its Facebook users in the last year. - 30:40 Building for "where the world is today" is a common mistake; instead, build for "where the future is going to be." - Facebook's scalable translation system allows them to attack all languages.

31:10Tactic 2: Virality

• Resource: Viral Loop by Adam Penenberg (case studies of viral growth).
• Resource: Ogilvy on Advertising by David Ogilvy (creative tips, e.g., "stick a car to a billboard with super glue").
• 32:00 Sean Park's Virality Model (3 factors):

1. Payload: How many people can you hit with one viral blast? 2. Frequency: How many times can you hit them per blast? 3. Conversion Rate: What percentage will convert?

• 32:40 Example: Hotmail (Canonical Viral Marketing)

- In the early days, free email clients were tied to ISPs. - Hotmail offered web-based email access from anywhere. - Traditional advertising was expensive. - Hotmail's viral mechanism: "Sent from Hotmail, get your free email here" link at the bottom of every email. - Payload: Low (one email to one person at a time). - Frequency: High (people email the same contacts repeatedly, multiple times a day). - Conversion Rate: Very high (people disliked ISP-tied email). - Result: Extremely viral due to high frequency and high conversion.

• 34:20 Example: PayPal

- Two-sided marketplace (buyers and sellers). - Viral mechanism: eBay (integrated payment option). - If a buyer says "I will send you money via PayPal," the seller's conversion rate is extremely high. - On the consumer side, PayPal offered money for inviting friends (e.g., "get 10 bucks"). This also had a very high conversion rate. - PayPal's virality was driven by extremely high conversion rates on both sides.

• 35:40 Clarification: Alex corrects himself, stating PayPal's virality was due to high conversion rate, not necessarily high payload and frequency in the traditional sense.
• 36:00 Facebook's Virality: Not via email sharing, but purely word-of-mouth.

- Unlike Hotmail or PayPal, Facebook had no native way to contact non-users. - It grew because it was an "awesome product you wanted to tell your friends about."

36:50 Audience Q&A on Virality:

• 37:00 Q: Doesn't payload grow in later rounds as more people use the product?

- A: Yes, as more people use Hotmail, they send more emails, increasing the overall payload. However, per-email payload remains small compared to modern viral engines (e.g., contact book imports).

• 38:00 Q: Doesn't the point of conversion matter (e.g., clicking a link vs. typing a URL from a billboard)?

- A: Absolutely. Removing friction from the flow (e.g., direct click vs. remembering a URL) significantly impacts conversion. Offline ads have much more friction.

• 38:50 Q: Aren't frequency and conversion rate related?

- A: Yes. Hitting someone with the same message repeatedly (e.g., banner ads, Facebook ads, news feed stories, invite emails) leads to creative exhaustion and diminishing conversion rates. This is fundamental across all online marketing channels.

39:50Ed Baker's Virality Model (K-Factor)

• Speaker: Alex Schultz introduces Ed Baker (formerly Facebook growth, now Uber growth).
• 40:10 Visual Description: A flow chart showing steps in a viral loop:

- User signs up -> Imports contacts -> Sends invites (to X people) -> X people click -> X people sign up -> X people import contacts.

• 40:20 This model breaks down the viral loop into conversion steps:

- How many people import contacts? - How many invites are sent per import? - How many clicks per invite? - How many sign-ups per click? - How many of those sign-ups then import contacts?

• 41:00 Multiplying these percentages (and the number of invites sent) gives you the K-factor.

- Example: 100 invites sent per import -> 10% click (10 people) -> 50% sign up (5 people) -> 10-20% import (0.5 to 1 K-factor). - A K-factor below 1 means the product is not viral.

• 41:50 While it's possible to get a K-factor over 1 (e.g., Viddy), it doesn't matter without high retention.
• 42:10 Analyze your invite flow to understand your K-factor, but prioritize retention first.

42:30Tactic 3: SEO (Search Engine Optimization)

• Three key things to consider:

1. Keyword Research: - Example: Cocktail Site: Alex spent a year optimizing for "cocktail making" in the UK, only to find 500 searches/month. "Cocktail recipes" was the real term in the UK, and "drink recipes" in the US. - Research involves: what people search for, search volume, competition, and value. - Tool: Google AdWords Keyword Planner. 2. Links: - PageRank (Google's algorithm) is based on authority, primarily from valuable links from high-authority websites. - "White on a white background" SEO tactics no longer work. 3. Internal Linking & Site Structure: - Example: Facebook Public User Profiles (2007): Initially, these pages received no SEO traffic because they were buried (only accessible via a long click path from the footer). - By adding a directory that Google could easily crawl, Facebook 100x'd its SEO traffic. - This distributed "link love" internally.

• 45:00 Table Stakes: HTML sitemaps, correct headers, etc. (well-documented online).

45:20Tactic 4: Email, SMS, and Push Notifications

• 45:30 Email is "dead for people under 25" (they use WhatsApp, SMS, Snapchat, Facebook).
• 45:40 Email still works for older audiences and for distribution.
• 46:00 These three channels (email, SMS, push) behave similarly.

1. Deliverability: - ⭐ HIGH PRIORITY: Your message must reach the inbox/device. - Avoid spamming, dirty IPs, shared servers with spammers. - Respect bounce messages (hard bounces mean stop trying). - Being a "high-class citizen" ensures long-term deliverability. - Push Notifications: Spamming power users or making it hard to opt-out leads to blocks, and it's difficult to re-engage them once they've opted out. 2. Open Rate & Click-Through Rate: - Compelling subject lines for email. - Engaging content for clicks. 3. Notifications (vs. Newsletters): - "Newsletters are stupid" because they send the same message to all users (new vs. long-term). - The most effective use is notifications – timely, relevant messages. - Example: Facebook Likes Notification: - Alex, as a power user, doesn't want an email for every like. - But for a new Facebook user, the first like is a magic moment. - Turning on this notification only for low-engaged users increased click-through rates across email, SMS, and push without spamming power users. 4. Triggered Marketing Campaigns: - Example: eBay Cross-Border Trade: An email campaign triggered by a user's first cross-border transaction had very high click-through rates because it was timely and relevant.

50:00Conclusion: Move Fast and Don't Be Afraid to Break Stuff

• Quote: "A good plan, violently executed today is better than a perfect plan tomorrow." - General Patton.
• 50:20 Facebook's ethos: "move fast and don't be afraid to break stuff."
• 50:40 If you can run more experiments, be hungry for growth, fight for every user, and execute quickly, you will grow faster.
• 51:00 "Marcus said he thinks we won 'cause we wanted it more." (Referring to Mark Zuckerberg).
• 51:10 "We just worked really, really hard, and we executed fast."
• 51:20 "Growth is optional." (Implies it requires deliberate effort).

Resources Mentioned

• 18:20 Danny Ferrante (on retention curves)
• 18:20 Chamath Palihapitiya (on Facebook growth team and magic moment)
• 18:20 Naomi Gleit (on Facebook growth and internationalization)
• 26:20 Ben Silbermann (Pinterest CEO, on marketing-driven growth)
• 31:20 Viral Loop by Adam Penenberg
• 31:40 Ogilvy on Advertising by David Ogilvy
• 43:30 Google AdWords Keyword Planner (for SEO keyword research)

Visuals Described

• 3:45 Chart: Active Users vs. Days from Acquisition (showing retention curves that asymptote or decline).
• 5:20 Graph: Number of Users vs. Days from Acquisition (illustrating cohort retention analysis).
• 40:10 Flow Chart: Viral loop (User signs up -> Imports contacts -> Sends invites -> Clicks -> Signs up -> Imports contacts).