Andrew Chen on Startup Cold Starts - a16z Insights

Understanding Startup User Engagement and Growth

Andrew Chen possesses extensive experience in analyzing how startups effectively attract and retain users, while simultaneously expanding their user base.

Currently, Chen shares his insights – gained as a founder and a former executive at Uber, where he spearheaded “rider growth” product teams for several years – with the companies in his portfolio as a general partner at Andreessen Horowitz. These include All Day Kitchens, Clubhouse, Substack, and Snackpass.

A New Book on Network Effects

To disseminate these lessons more widely, Chen, a consistently prolific writer, has recently released a new book through HarperCollins titled “The Cold Start: How to Start and Scale Network Effects.”

Within its pages, he presents interviews with 30 leading teams and founders from companies like Twitch, Airbnb, and Slack, illustrating the essential elements for transforming a startup into a prominent brand and sustaining user engagement.

We recently spoke with Chen about his book, including a discussion of how web3 projects align with, or diverge from, the patterns observed in earlier startups. The following are excerpts from our conversation, lightly edited for brevity.

The Tinder Cold Start Story

TC: Your book highlights numerous instances of overcoming “cold starts,” including a fascinating story about Tinder’s early days. Could you elaborate on what transpired?

The initial iteration of the app was well-developed, featuring swiping, profile photos, and all the expected functionalities. However, when the founders invited friends, retention rates were disappointingly low. A critical mass of profiles was lacking.

To address this, they sponsored a birthday party at USC, requiring attendees to download Tinder for entry. This event proved successful, and the following day, users discovered a pool of potential connections they hadn’t had the opportunity to meet in person.

Sean Rad credits those initial 500 users with facilitating Tinder’s expansion across the USC campus, a strategy subsequently replicated at other institutions.

Defining the “Atomic Network”

You discuss the significance of “atomic networks” in the book. In Tinder’s case, it was 500 people. How can a founder determine the appropriate size for this network? Is there a quantifiable formula?

Conceptual reasoning is key: Is the network inherently small, requiring only a few participants, or does it necessitate a larger base of over 100? Data analysis provides valuable insights.

Examining threshold points, such as Facebook’s “7 friends in 10 days” or Airbnb’s need for 300 listings with 100 reviews, reveals the point at which a product’s value truly begins to emerge.

Strategies for Building a Core Audience

What are the recommended practices for cultivating this initial core audience?

Innovation and creativity are paramount. Simply replicating past successes is unlikely to yield results, as evidenced by the declining click-through rates on banner ads.

The approach must align with the product’s value proposition. Social networks thrive on college and high school campuses, while workplace collaboration tools are better suited for startups and small teams. Web3 projects require engagement within Discord servers, subreddits, and with key Twitter influencers.

Clubhouse’s Unique Trajectory

Regarding Clubhouse, its initial growth didn’t rely on college student adoption. Is that accurate?

I was an early adopter, user number 104. When we invested in the Series A round and I joined the board, the platform had only 500 daily active users. The initial atomic network consisted of founder Paul Davison and his network in the San Francisco Bay Area.

Retention rates were remarkably high, with users spending over 90 minutes per session listening to conversations. Further analysis, including data from Sweden and Nigeria, demonstrated similar engagement patterns in other countries, reinforcing our confidence in the company’s potential.

Sustaining User Engagement Through Feature Development

How do you advise founders on adding features that encourage continued use? Could you specifically address Clubhouse’s plans?

Facebook’s initial lack of appeal underscores a crucial point: platforms become valuable only when a user’s network – friends, family, and content creators – joins. This principle applies to platforms like Clubhouse and Substack.

Substack is expanding beyond its initial appeal to media, politics, and tech, venturing into areas like cooking, comics, and graphic novels. Clubhouse is exploring sports, news, and politics, aiming to broaden its appeal through vertical-specific content.

The Impact of Web3 on Network Building

What impact do you foresee web3 having on network development?

Web3 inherently incorporates network effects, potentially educating a new generation of founders about their importance. Bitcoin’s value, for example, stems from collective belief. A forked version, even with identical code, would lack the same value without widespread adoption.

The same principle applies to NFTs like Bored Ape and CryptoPunks. However, web3 also presents a cold start problem. Launching a new NFT collection requires targeted outreach to relevant communities. Without it, the NFTs may go unnoticed.

I anticipate revisiting these case studies in five years, replacing examples like Tinder, Dropbox, and Slack with their web3 counterparts!