The Cold Start Problem: How to Start and Scale Network Effects by Andrew Chen (Book Notes) Part III
Part III — The Tipping Point
Chapter 11 — Tinder
Datings apps are hard to start and even harder to scale. Dating is a hyper-local affair. Where even if people are in the same city that’s not enough to meet up. Don’t ask two single people, one in SF and one in Oakland. Across the bridge from each other, to get together or the LA equivalent Santa Monica and Silverlake. A dating apps network needs a lot of density to be successful. Even if you gain success in one demographic for example 40-year-old Christian singles, you have to restart the network for new demographics as well. Dating networks are naturally high churn. When people find success on dating apps they’ll be more inclined to leave the platform.
USC campus 2012
When Tinder first launched they showed dating profiles but it didn’t have swiping left and right just yet. It was only a red x to pass and a green heart button to pass. It was only later that the left and right swipe would be added. The initial growth of tinder was slow. So they texted all of their friends in their address books. Bout 400 people trickled in slowly but it didn’t take off because there weren’t enough people on the app.
The product of Tinder was hard. It required you to attract multiple segments of users at the right proportions at the same time. It’s a two-sided network of men and women that must be scaled up at the right time in the right ratios. Not too many ladies and not too many guys at similar levels of interests, demographics, and attractiveness to ensure everyone has matches. Online dating is typically something that doesn't have viral growth. Many folks feel uncomfortable telling people they use dating apps. If the product is successful and matches the right people, you now have two people who essentially quit using your app. All of these dynamics make these products hard to start and harder to scale.
USC was the ideal place for Tinder. The university has 19,000 undergrad students. It had a very active social scene centered on sororities and fraternities. All in about the same age range, similar geography, and similar interests including going to parties.
The founder of Tinder coordinated with hyper-connected people on campus to throw a party and a luxurious house in LA. There was one catch with the party. First, you had to download the tinder app to get in. The party was great and was a success. The next day everyone woke up and remembered they had a new app on their phone. This one-party created the highest spike of downloads. 500 of the “right people” is was mattered here. These people were the most social, most hyper-connected people on the USC campus that were all on tinder at the same time. Tinder started to work! Matches started to happen and they started to chat.
They figured out how to build an atomic network and figured out how to build the next one by throwing more and more parties. Tinder quickly reached 15k downloads within a month. 500k just a month after that. They believed in top-down marketing so they went to highly social people and had them promote it to their friends and grew from there. They found that once they hit 200k downloads in a single market then the app would hit escape velocity to take over that region completely.
Tinder’s first stage solution was launching at a party at USC. The second stage solution was then going from campus to campus which was the tipping point.
This method allowed them to build an atomic network. These growth tactics continued to scale there were valentines tinder parties, cocktail tinder parties sorority tinder parties, and more. The team scaled by recruiting ambassadors who were highly connected on campus and executed the same strategy until they leveled up from campus to cities to international regions.
Introducing the tipping point
The key to Tinder’s start was discovering a repeatable strategy that took them to USC to other colleges. Then absorbing the metropolitan areas and then country to country. This is when the market hits the second part of our Theory. The tipping point. This is where momentum starts to go in your direction and your strategy should be oriented around tipping over entire markets rather than launching individual atomic networks one at a time.
Invite Only-LinkedIn
Why has the invite-only tactic worked? Some say so people go to social media to ask for invites and that generates hype. Others say it’s good because it limits growth which gives teams time to fix bugs and scale infrastructure before going to market fully. The most important part of the invite-only mechanic is that it works like a copy and paste feature. If you give out invites. That network will copy itself over and over again.
Reid Hoffman said on the first week of launch employees and investors of Linkedin could invite as many people as they wanted. But you couldn't sign up with just the website. We intentionally seeded the website with mid-tier professionals that wanted to take time to connect. The invites went out to this tier of the talent hierarchy. Making Linkedin invite-only helped them see explosive growth. Members started inviting their own networks on the platform. Invite-only mechanisms are useful to its initial network because they know the person that invited them. In total it was a couple of thousand individuals.
The professional network was relatively non-competitive, Linkedin tipped the market faster than competitors can emerge and ultimately won its category. When a new product carefully curates a network followed by implementing invites so it can copy and paste similar networks then it can grow to take over the market.
The Welcome Experience
Invite-only mechanics provide a better welcome experience. They’re already connected to one person as they join the network which is a giant step towards solving the cold start problem. The most connected people tend to be invited earlier and they tend to invite other highly connected people. Which creates an environment of social butterflies which is extremely helpful in starting a new network. Invite-only is a powerful strategy.
Chapter 13- Come For The Tool Stay For The Network — Instagram
Come for the tool stay for the network is one of the most famous strategies for launching and scaling networks. Start with the tool, a product experience that is useful even for one user as a utility. Then over time, pivot the users into a series of use cases that tap into a network. The part where you collaborate, share, communicate and interact with other users.
Dawn of The App Store
When the app store was first released there were only 50K or so apps were published. Within the first month, Hipstamatic made it to the top 10 apps in a few countries. Yet some of their design choices were odd and added friction. The app cost 1.99 and after you applied filters it would save to your camera roll. Hipstamatic was a tool.
In the same year, Burbon was being created. This was a browser-based app for checking in at locations, making plans with friends, and sharing photos. Their problem was that the product was getting too complex. The team stripped everything but photos and changed the name. They wanted to be really good at one thing. More importantly, Instagram was built with a network from day one. It had user profiles, a feed, friend requests, invitations, and many other features of a modern social product. The filters were more direct than Hipstamatic. Also, Instagram was free.
Instagram launched on October 6th, 2010 to the App store. At the end of the first week, it had been downloaded 100K times. Instagram is still one of the fastest-growing apps ever created. 6 months after the app's launch it was discovered that 65% of users were not yet following other people on the network. In other words, Instagram was being used first, as a tool. Instead, their engagement was mostly around photo editing noting that Instagram 2.2 million users upload 3.6 million new photos her week or 6 photos per second. Instagram was first being used as a tool. The network would come later. Instagram grew faster and faster and celebrities started showing up. They all came to create content that built network density and increased engagement. Over time the tool part of the app has waned in importance. At this point, the vast majority of photos 82% had no photos used at all. Eight years after it’s launch. Network Effects took over. It’s more network and less tool. Instagram is worth several 100 Billion.
How Great Tools Help Tip Entire Markets
A popular strategy for bootstrapping networks is called come for the tool stay for the network. The idea is to initially attract users with a single-player tool then over time get them to participate in a network.
The tool gets them to initial critical mass. The network creates the LTV for users and defensibility for the company. There are many other examples.
The google suite creates documents, spreadsheets, and presentations but also network features around collaborative editing, comments. Minecraft and street fighter can be played in single-player mode or multi-player mode. Yelp started out as a directory tool for people to look up local businesses showing addresses and phone numbers. Eventually, they built a database of photos and reviews. Linkedin started as a tool to put your resume online but encouraged you to build your professional network over time. Come for the tool stay for the network circumvents The Cold Start Problem and makes it easier to launch into an entire network. It minimizes the requirement for an atomic network in turn makes it easier to take on an entire network.
Underlying patterns for tools and Networks
4 Examples of tool and network
Tool: Create, Network: Share with others
Instagram Youtube G-suite and Linkedin
Tool: Organize, Network: Collaborate with others
Pinterest, Asana, Dropbox
Tool: System of record, Network: Keep up to date with others
OpenTable, Github
Tool: Lookup, Network: Contribute with others
Zillow, Glassdoor, Yelp
Building a tool and network combo is a powerful approach but it doesn’t always work. Pivoting users from tool to network can be hard. Sometimes only a small percentage will make the transition because it requires new behaviors and also requires them to stick. Some might get stuck on just the tool. That’s what makes this tricky to pull off. The conversion rate from tool to network might be low. Some tools and networks are highly integrated like dropbox’s folder sharing functionality which defines its network. This has a high conversion.
Flintstoning — Reddit
Manual human effort. In software, this is where missing functionally is replaced with manual human effort. Sometimes apps don’t have basic functionality but will have you contact the developer so they can complete a task on the back end. This allows for getting a product out in the market quickly, handholding users, and getting feedback fast. The goal is to manually fill in the critical parts of the network until it can stand on its own.
Reddit is a great example. When they first launched. There were just 2 users. No one wants to join a dead community. The founders, posted on Reddit initially using dozens of dummy accounts. Otherwise, the community would not have gotten off the ground. All of the dummy accounts looked and acted like real users. It held Reddit over until Reddit was able to shut the script that was posting the dummy posts down.
Flintstoning The Hard Side Of the Network
Food delivery apps signing up restaurants is difficult. DoorDash and Postmates flintstoned by showing a large selection of restaurants regardless if those restaurants were signed up or not. Flintstoning can be thought of as a spectrum. On one end there are fully manual human-powered efforts. Then there’s a hybrid where software suggests actions to take where people are in the loop. And there’s fully automated powered by algorithms. If the manual efforts work, technology can be layered on to create leverage. How far can flintstoning go?
Nintendo did this with the switch. Nintendo launched Mario and Zelda games to help the Nintendo Switch launch. They were built by internal studios within the company each with their own creative directors and game designers. Imagine 100’s of people flintstoning a new network. That’s what happened with the switch. It worked and in the first few years, the switch launched 70 million units. The gaming industry calls this first-party content.
Chapter 16 - Always Be Hustling- Uber
Travis would say to product teams can solve problems but it’s slow. Ops can do it fast. As a result, Uber says itself as an Ops-led company. The hustle within the ops team was renowned and one of the foundational elements of Uber’s success.
The Importance of Creativity
Creativity is important because there are often brief moments of opportunity that can cause a market to quickly tip if you try the right idea. This is like Twitter launching during the SXSW conference where a critical mass of users was attending. Airbnb also latched onto local events like renting your apartment to Octoberfest attendees. The most powerful way to bootstrap supply is to guarantee demand. Usually, these types of hacks a neither scalable nor repeatable but in the early days when the focus is on tipping each new additional network everything helps. The Uber ops team provided a steady stream of this creativity. Every new city launch was essentially was a cold start problem and the ops teams were structured to be autonomous and decentralized. The goal was to tip over the entire market one city at a time. The launch team would coordinate that a celebrity would be rider zero with local press coverage. They also had special promotions like Uber puppies and Uber Kittens which was a way for you to have puppies or kittens show up at your office for an hour at a time. They found that each city required different aspects to get the market going. NY was a licensed market and Uber competed with the subways and professional drivers. LA is a sprawl of a city where everyone has a car. Each new market became easier and easier as the tipping point kicked in.
Hustle as a system
The ops team culture rewarded experimentation. They had Uber ice cream, Uber puppies, Uber mariachi band, Uber health (flu-shots), Uber line dance (celebrate Chinese new year), and dozens more. The ops team would align holidays and dates with special features that promoted growth. A driver referral program “give 200 dollars and get 200 dollars” would be bumped up to 300 dollars and branded as a new years campaign. In-app notifications were used to celebrate holidays like July 4th, Thanksgiving, and others to keep messaging fresh and response rates high. They created new levers in the app so city teams could customize strategies. Uber moto, helicopter, and pitch could be launched.
Enterprise Hustle
Founders tapping their personal network. Only three sourcing strategies account for B2B company's early growth. Personal network, seek out customers where they are and get press. Thus your choices are easy yet limited. Almost every B2B business both hit’s up their network and heads to the potential places their customers were spending time. The question is not which one to pursue but how far your network with take you before you move on. It’s a huge advantage to have a strong personal network in B2B. Getting press is rarely the way to get started.
Paul Graham said that finding and convincing users is a good way to start getting people on your product. One of the most common bits of advice that YC’s give is to do things that don’t scale. You can’t wait for users to come to you. Founders don’t talk to users for mainly two reasons. Shyness and laziness. They rather be writing code and don’t feel like going out just to be rejected all the time. He advocates B2B startups take the consulting approach initially with customers. To act as if they were consulting clients. Building what they need on an ad hoc basis.
