The Basics of Growth Marketing: Engagement & Retention

Episode Summary

Executive Summary: The episode focuses on how growth shifts after acquisition into engagement and retention, and why cohorts, frequency, and network-effect analysis are essential for understanding product health. Andrew Chen and Jeff Jordan argue that the best companies measure whether users keep returning, improve over time, and become part of a durable core, rather than relying on raw growth alone.

Main Topics: From acquisition to engagement and retention (Priority: 5/5): The discussion frames growth as a progression: first acquire users, then optimize for engagement, retention, and eventually resurrection of churned users as the user base matures. Cohort analysis as the core diagnostic tool (Priority: 5/5): Users are bucketed by signup period or segment so teams can compare how behavior changes over time, spot decay or plateauing, and identify whether product value is improving or degrading. Network effects revealed through improving cohorts (Priority: 5/5): If a product truly has network effects, newer cohorts should perform better than older ones; improved curves, lower churn, and stronger conversion are evidence that the network is becoming more valuable. Metrics for frequency and retention (Priority: 4/5): The speakers discuss DAU/MAU, L7/L28 histograms, and the 'smile' curve as ways to measure frequency and identify a hardcore core of users, while warning that metrics must fit the product’s natural cadence. Power users and ladders of engagement (Priority: 4/5): Products often have a ladder from casual use to high-value, repeated, collaborative behavior; companies should identify power users and move more users toward those deeper behaviors. Aha moments, onboarding, and activation (Priority: 4/5): The 'aha' moment is not just signup but a sequence of setup actions that let users experience the product’s core value; onboarding must be engineered to get users there quickly. Product-specific retention and monetization logic (Priority: 4/5): Retention must be judged in context: episodic products like travel or costume retail can be healthy with low frequency, while daily-use products need very different patterns; upstream signals may be more useful than final conversions.

Key Arguments: Growth alone is insufficient; engagement and retention determine whether acquired users create lasting value. Cohort curves are one of the best ways to detect product health because they show whether usage decays, plateaus, or improves over time. A real network effect should show up in the data as newer cohorts outperforming earlier cohorts, not just in company claims. Engagement is harder to fake than acquisition; adding notifications may inflate monthly actives without improving true daily engagement. Metrics like DAU/MAU are useful only when they match the product’s cadence and monetization model; low DAU/MAU can be fine for low-frequency products. Products should be analyzed by segments such as geography, team size, or market density to isolate where network effects and engagement differ. The best companies use data to iteratively refine onboarding, lifecycle messaging, incentives, and product design to move users up the engagement ladder. Retention is not always binary churn vs. no churn; for episodic businesses, repeat behavior on the expected schedule may be the correct success measure.

Data Points: DAU/MAU: Facebook historically had 60%+ daily actives over monthly actives - Used as an example of a high-frequency product where DAU/MAU is a meaningful engagement measure. OpenTable median use: 2 times per year - Illustrates an episodic product where low frequency does not imply poor retention. L28 histogram window: 28 days - Used to reduce seasonality when measuring frequency across a month. Growth accounting equation: net MAU = new users - churned users + resurrected users - Explains how monthly active users evolve as acquisition, churn, and re-engagement interact. Product usage example: 5 people vs. operating-system-level adoption - Slack example showing how organizational scale changes the engagement curve. Engagement threshold: sub-15% DAU/MAU - Mentioned as a sign that a product unlikely to work if it is supposed to be a daily-use product monetized over time. OpenTable market analysis: San Francisco was their first market - Used as a diligence example where multiple metrics improved over time, suggesting strong network effects. Conversion example: 5 minutes versus 2 minutes wait time - Illustrates diminishing returns in network effects, where improvements matter more at larger waits than already short ones.

Pivotal Quotes: "They don't lie." — Jeff Jordan: Referring to cohort curves as the definitive test of whether a business truly has network effects. "Growth is good. Growth and engagement is really, really, really good." — Jeff Jordan: Summarizes the central thesis that acquisition must be paired with durable engagement. "If you spend a lot of time on Google.com, you know, refining your searches and clicking around, that means actually the service is doing poorly." — Andrew Chen: Explains that for some products, success means users get what they need quickly rather than spending more time in-product.

Implications: Teams should measure product health by cadence-appropriate retention, cohort quality, and core-user behavior—not raw growth alone. The strongest products engineer value loops that improve cohorts over time and make engagement hard to fake.