profile pic

Precise Asteroid

Welcome to Amir's Blog 👋

The Cold Start Problem


The Cold Start Problem by Andrew Chen is a book about the unique charactaristics of network products and network effects. What is unique about these products,what makes them strive or fail.

My highlights are available here.

My Main Takeaways

Atomic network

it is not the overall size of the network that is important, but rather it's density and completness. you need to find the miniaml network set and start there. with messaging apps it is 1:1. With products like slack it is the team. That's one of the reason Google+ failed. The network size and growth was impressive, but the user who got to the network through bundliing (links on for example) did not find any of their (atomic) netowrk there. so they just make a quick tour and left.

it is important to add the right people of that atomic network at the same time. Otherwise the network will not sustain.

When launching, start with the smallest atomic network size possible, and with a minimal but most important list of features. The network itslef will take care of the rest.

But here’s the paradox: To build a massive successful network effect, I argue that you must start with a smaller, atomic network.

Diminishing Returns

At Uber, ETA is one the key value props. They started with 15 min and wanted to get to 3 min across a city. Going below 3 is a waste. On the one hand side it starting to bug the users, who like to order a head of time and get some time to go to the street and find the driver. On the other it means you are in over supply. Stop at an expereince that is good enough

Eventually, the difference between two minutes to get a car or one minute to get a car becomes diminishing returns.

Hard Side of the network

Always florish the hard side of the network. At Uber, these are the drivers. At youtube, the creators.

Make sure they are happy. without them the netowrk will fail.

Moment Zero

The worst user expereience the user could have. Going on Netflix and finding no film to watch. Going on Uber and finding no driver.

Monitor these cases and understand the reasons.

I encourage product teams to develop their own form of this metric, laid out as a dashboard of networks—whether that’s divided by geography, product category, or whatever else makes sense. Within each, it can be useful to track the percentage of consumers that are seeing zeroes.

Bundeling is NO key

MS bundled Bing. Never worked. The product should be better than the others in order to keep people. We need to remeber that when we talk about cross polynation. The user is just a click a way from the competitors. Uber Eat for example next killed DoorDash

Made by Amir 💚