By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.

a16z Podcast | The Basics of Growth 2 -- Engagement & Retention

with Andrew Chen (@andrewchen), Jeff Jordan (@jeff_jordan), and Sonal Chokshi (@smc90) Once you have users, how do you keep them engaged, retain them, and even "resurrect" or re-engage them? That's the focus of this episode of the a16z Podcast, which continues our series on the basics of growth from user acquisition to engagement and retention -- covering, as always, key metrics and how to think about them. Especially as many products and platforms evolve over time, so do the users, some of whom may even use the product in different ways... so what does that mean for engagement, and how can startups analyze their users? "Show me the cohorts!" may be the new "show me the money"... Featuring a16z general partners Andrew Chen and Jeff Jordan, in conversation with Sonal Chokshi, the discussion also covers everything from how network effects come in to play (is there really a magic number or "aha" moment for a product?) to who are the power users (and the power user curve for measuring, finding, and retaining them). Because at the end of the day, you don't want a leaky bucket that you're constantly trying to fill up. That doesn't work, and definitely won't scale.

Sonal ChokshihostAndrew ChenguestJeff Jordanguest

Jan 2, 201932mWatch on YouTube ↗

Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.

CHAPTERS

0:00 – 2:39
From Acquisition to Engagement as Cohorts Change
Sonal sets up the episode on engagement and retention. Andrew Chen explains that user quality shifts as a product saturates its core demographic, and Jeff Jordan says successful companies move from chasing users to engaging the ones they have, using Pinterest as an example.
- •Early cohorts differ from later ones, as in rideshare moving from urban to suburban and rural users
- •This shift shows up in engagement and eventually in LTV
- •Successful companies run out of users and must turn to engagement
- •Pinterest is cited as a company where most US women have already downloaded the app
- •Investors like engagement because it is hard to hack
2:39 – 5:21
Cohort Analysis and the Growth Accounting Equation
Andrew Chen explains how to build cohort curves by bucketing users by signup week and comparing their activity over time. He covers the leaky bucket problem and the equation of new users minus churned plus resurrected, and how a company's focus moves from acquisition to churn to resurrection.
- •Cohort curves typically decline, flatten for good products, and rise for the best
- •Comparing cohorts shows whether usage patterns are improving or worsening
- •A curve that decays to zero means a leaky bucket that needs ever more acquisition
- •Net MAU equals new users minus churned plus resurrected users
- •Companies move from acquisition to retention to resurrection as they grow
5:21 – 7:40
Using Cohorts to Prove Network Effects and Segmenting Them
Jeff Jordan says cohort curves reveal whether a business has real network effects, since newer cohorts should behave better than earlier ones, as with OpenTable. Andrew Chen and Jeff discuss segmenting cohorts by market geography or team size, with Slack as an example.
- •Marketplace models are best built from cohort curves
- •With network effects, newer cohorts should outperform early ones
- •In network effects businesses, Jeff always asks to see the cohorts
- •Segmenting by market geography acts as a natural A/B test
- •B2B products can be segmented by team size, as with Slack
7:40 – 10:37
The Aha Moment and the Magic Number
Andrew Chen describes the aha moment where a user grasps a product's value, often requiring onboarding setup. Sonal asks whether it is real, and Jeff Jordan discusses the magic number idea and how Pinterest needs local content and quick personalization.
- •Aha moments often require setup, such as following the right people
- •On Facebook, existing friends make the aha moment more powerful
- •Jeff recalls a service that statistically linked a follow count to long-term retention
- •Pinterest needs local content in new markets
- •Entrepreneurs must engineer their way to delivering that experience
10:37 – 15:21
Shifting Leverage and Climbing the Ladder of Engagement
Sonal asks what happens after acquisition, and Jeff Jordan says leverage shifts over time from acquisition to activation and retention. Andrew Chen describes the engagement ladder, using Dropbox, and how to move users up it with education, incentives and product refinement.
- •Leverage moves to engagement once most users are acquired
- •Pinterest is still in acquisition mode internationally
- •Users should be segmented into an engagement state map
- •Dropbox users move from syncing computers to sharing folders at work
- •Tactics include content and education, incentives, and refining the product for the use case
15:21 – 19:38
Frequency Metrics: DAU/MAU, L28 and the Smile Curve
The speakers cover engagement metrics, starting with frequency. Andrew Chen discusses DAU/MAU, when it applies, and the L28 and L7 histograms, while Jeff Jordan describes the smile shape and why investors prize engagement.
- •Facebook popularized DAU/MAU, with 60%-plus daily over monthly actives
- •Low DAU/MAU can be fine for naturally infrequent products like travel
- •L28 is a histogram of days active in a month that exposes a hardcore segment
- •The smile curve rises at the end, as with Facebook, WhatsApp and Instagram
- •A big engaged audience is rare and valuable
19:38 – 21:55
Why Engagement Metrics Are Hard to Game
Andrew Chen and Jeff Jordan explain that growth is easy to game with spend or distribution hacks, while DAU/MAU and L28 are much harder to fake. Extra notifications can even lower DAU/MAU, and some apps reach millions of users who never return.
- •Engagement powers acquisition loops like viral sharing
- •More notifications can increase MAU but lower DAU/MAU
- •Acquisition is easy to game by spending money
- •Early Facebook platform apps reached a million users quickly with a mean of one visit
- •Investors need to look below seductive growth curves
21:55 – 24:31
Network Effects in the Data: OpenTable and Diminishing Returns
Sonal asks about engagement and network effects. Jeff Jordan describes his OpenTable diligence in San Francisco, where every metric improved over time, and Andrew Chen says network effects are a curve with diminishing returns rather than a yes/no.
- •Real network effects show in the data, such as improving cohort curves
- •OpenTable San Francisco metrics all improved over time
- •Restaurant density should be measured as a share of the market's restaurant universe
- •Rideshare wait times show diminishing effects
- •The same analysis applies to YouTube channels and workplace collaboration tools
24:31 – 26:11
Engagement vs. Retention
Sonal asks whether engagement and retention are the same. Andrew Chen and Jeff Jordan say they overlap but differ, using weather apps and e-books as examples, and note that retention depends on each product's natural cadence.
- •Weather is low frequency but high retention
- •Games and e-books have high engagement but low retention once finished
- •Retention asks if users are still around after some time
- •Different cadences require different models
- •Products should be strong on at least one of retention, frequency or monetization
26:11 – 28:54
Time Spent and the Battle for Attention
The conversation turns to engagement as time spent. The speakers discuss competing for limited minutes against incumbents like Facebook, Amazon and Google, why attention competition crosses categories, and how Google's goal is quick results.
- •Few businesses achieve great time-based engagement
- •Consumer products now must displace minutes from incumbents on the first screen
- •Content competes with anything that grabs attention
- •Some products have long sessions while others have many micro sessions
- •Google succeeds when users spend less time, making it a frequency and monetization play
28:54 – 30:48
Measuring Retention and Churn for Episodic Products
Sonal asks about retention metrics. Andrew Chen says churn is tricky for episodic consumer products and that active versus inactive matters more, while Jeff Jordan suggests adjusting the time period. Andrew also describes tracking upstream signals like Zillow users checking listings.
- •Subscriptions and SaaS have obvious churn through payment
- •Episodic products need active versus inactive judged against acceptable rates
- •The Halloween costume example shows an infrequent buyer who has not churned
- •Adjust the measurement window to the product's cadence
- •Upstream signals such as searches or availability checks can serve as proxies
30:48 – 32:40
Iterating with Data: Closing Takeaways
Sonal asks how to balance architecting with discovering. Jeff Jordan advises iterating through hypotheses and deep data analysis, while Andrew Chen stresses choosing metrics that fit your strategy and finding loops and networks that keep engagement improving.
- •Develop hypotheses and test them, informed by data
- •Consumer internet marketing is more science than art
- •Pick metrics that validate your strategy, since products differ in nature
- •Find loops and networks within the product to keep engagement improving
- •Growth plus engagement is described as very good