Chaos Graph

Notes on startups, web3, philosophy, psychology or whatever else I'm obsessing over.

How to think about building consumer products

(the below is adapted from an essay I wrote during my time at Facebook outlining how consumer PMs think.  I've removed some of the FB focus, and I'm sharing it here since it might be useful for folks outside, and I'd like to be able to reference it in conversations.  Not all of the concepts are as relevant in a web3/crypto world, or even a SAAS world, but some might be…)

Sometimes the frameworks that consumer product leaders use to make decisions can seem mystifying, or arbitrary.  “Why would we focus on this new feature when we’re getting lots of DMs from users asking us about X?”, “Why build that stupid invite functionality when this new feature could be super fun?”

But the reality is that consumer leaders usually (push them to verify!) have a very clear structure for how they prioritize and it all falls out of a framework that we use to think about what a successful product achieves in each stage of its lifecycle. My goal here is to introduce everyone to this framework so that each of us has it in their toolkit.

An important caveat: this is not always the right way to think about building product, and we should not be afraid to deviate from it. But decisions to deviate from this framework need to be intentional and well-considered (if for no other reason than because our stakeholders are going to be asking us to defend those decisions).

0 : Are we focusing on a problem that matters?

While we should never be a slave to metric optimization, metrics are the best way for us to understand how good of a job we are doing at serving people. Choosing the right metrics serves two purposes: (1) It helps us quantify our empathy — are we doing a good job at solving a real problem for people? (2) It helps focus us on the things that will help us to achieve our actual mission. We could choose to build anything, but to keep us focused, we should always be able to trace its impact back to our mission - whether personal or organizational. Given this framing, there are a few metrics that are consistently treated as strategically important for consumer products. If we are focused on another metric, there is usually a clear reason why it will ultimately translate into wins in one of these key areas:

(1) Visitation -

What it is: How many times a week/day do people choose to visit our app/product?

Actual Value: “Do I find this valuable enough to keep visiting to connect with friends?”

How We Measure It: L7 (number of days out of the last 7 a user is active), Sessions Per Day

(2) User Growth - How many people are using our product?

What it is: How many people are using the product?

Actual Value: Have we built something that is valuable to a large number of people?

How We Measure It: MAU, WAU, DAU

(3) Core Action - “Are people sending a message, or sending money, or making a purchase?”

What it is: We usually create an app not just for people to open it, but for them to do something with it.  Are they taking that action?  If it’s a messaging app - are they sending a message?  If it’s a camera, are they taking photos?  If it’s a store, are they making purchases?

Actual Value:  Have we built something that people are using to achieve their goals.

How we measure it:  Actions per user, Retention on user-actions (e.g. if I completed this action yesterday, how likely am I to do it again tomorrow, or a week from now?)

When you think that you have identified a project that could be awesome, you should have a clear story about how it will help us to move one of these metrics so that we can understand how effective the project will be in solving people’s problems and helping us to achieve our mission.

0→ 1 (Product market fit): Can you show that this feature can get to Product Market Fit? (Flat retention curves)

Okay, so you think you have an awesome idea that's going to totally change the game for our users and in doing so, move a key metric. How will we evaluate if it's really a great product? To answer that, you need to understand the type of products that we want to build.

As I write this I'm listening to an awesome song on my shiny headphones. Both the song and the headphones are genius products of engineering. But while I will listen to the song maybe 10 times before moving onto other music (okay, 20 if it's T.Swift), I will use these headphones for years to come. Our first choice is almost only ever to build the headphones because that is the thing that we think of as having enduring value to our users. We want to partner with creators to generate the songs that make people excited to use our beautifully designed headphones.

This is why the term “Product Market Fit” means only one thing at consumer companies — user retention. Retention measures the likelihood that a given user who has tried a feature on Day 0 will continue to be using that product on Day 30. Often we use j-curve graphs (seen above) to capture this value. We want the line to flatten somewhere above 0, telling us that we have retained some segment of users. This tells us that we have built something that people are finding value in and will continue to use. When evaluating an idea, we need to have a clear story about why people will use a feature repeatedly. This is all any PM means when they say Product Market Fit.

Product Market Fit (Retention) is also important because it's the critical component that enables us to grow a product to 1B+ people. Imagine pouring water into a cup... if there's a hole in the cup, the cup will not retain the water and no matter how much you pour, you will not be able to fill the cup. This is how we think about whether to grow a feature. Before we start trying to have billions of people experience a product, we want to make sure that the product will retain them.

1 → 1,000,000,000: Can we grow the cohort of engaged users to 1B+ people?

Cool, so we have an awesome feature that can move a key metric and that is retaining users, what now? Our next step is to prove that we can grow the product from something in which a small set of users are finding value to something that can appeal to 1B people. Facebook used to have a general rule that if a feature cannot be valuable to >100M people within two years, it's likely not worth investing in. This is a really high bar. They choose it because there are so many investments they can make that will touch the lives of this many people, and th need to be laser focused on doing the greatest good for the greatest number.

But how do we grow a product? We have a couple key strategies, and it's useful to think of how we can plug them in when designing any new feature.

Entry Points - How will people discover our new feature? Are we focused on search engines?  Do we have a partnership with another app? Do we have a core viral loop where people will tell their friends?

Education - How will we teach people that our feature is awesome and they should try it? Think ads and NUXes.

1B → $1B: Can we make money off of it?

Only now, after we have built something that millions of people are using regularly do we begin to focus on monetizing. Before we hit that bar, we do not even consider how we might make money from our products because the revenue opportunities are likely going to be too small to be interesting for a large, venture scale company. At this stage, we begin to think about how much revenue we can generate for each person that is using our product. We do not need to be too worried about monetization opportunities early in the product lifecycle. If we focus on making a product that people want to use and spend time with, monetization will fall out naturally.