Brian Armstrong just bought one of the most expensive homes in the state of California. For $133M, his modern Bel Air mansion is now one of the priciest single-family home transactions ever.
The self-confessed “on the spectrum” (of Autism) CEO has managed to become one of the world’s richest entrepreneurs, as well as one of the most powerful men in web3. With the purchase of his Bel Air mansion, Brian has coronated his position as one of the richest and powerful men in the world, period.
With that position, both Brian and Coinbase have been the subject of several amazing profiles. Wired, The Generalist, Inc Magazine – it seems like everyone has written about Coinbase.
So, we’re not going to cover the things you’ve already read everywhere. Instead, today I want to focus on: what lessons can product builders take from Coinbase? Like we did for PayPal, Stripe, Nubank, and PayTM, we want to break down the new FinTech giant, so we can replicate some of the learnings at our workplaces.
The twist with Coinbase is that it is one of the most interesting companies in the crypto and blockchain worlds. One of the ongoing strands we are exploring in Product Growth is how to build for web3 specifically. Every PM is being asked to evaluate the impact of web3 on their business.
This makes Coinbase one of the most interesting companies to explore. It is the highest public market cap pure-play at $50B, but it also is a very centralized exchange. Unlike Uniswap, for instance, its decentralized alternative, there is still an element of trust in the centralized player. Many PMs will also need to take a centralized entry into the web3 world.
I’ve been closely following Coinbase since Kevin Rose’s interview with Brian in 2013. I have been using the product since 2016. My research note for Coinbase had 150 articles when I started this piece. After that base, to write this piece, I combed through every post on the Coinbase blog. I watched YouTube videos and read Twitter threads. All in all, this piece took well over 40 hours. That’s a whole work week of value for free.
You’ll notice my first-hand research leads to several different interpretations of history. So, strap in as we explore 10 product lessons from the growth of Coinbase.
2010
Lesson 1: It’s Easy to Be Skeptical About the Future
What were you doing on Christmas Day in 2010? I was sleeping in and overeating.
Brian Armstrong wasn’t. As a fraud systems engineer at Airbnb, Brian was reading cypherpunk literature. It was then that he came across Satoshi Nakomoto’s Bitcoin whitepaper. Released two years prior, Bitcoin was already a toddler by that time.
Plenty of smart people were skeptical of Bitcoin then. Plenty of smart people are still skeptical about Bitcoin now. But, Brian was built differently. Brian was an engineer looking for a startup to build.
Having grown up in San Jose, California, at a “brutal” high school where he “got made fun of pretty hardcore,” Brian had the backbone to withstand the criticism and stay a hungry nerd. He found his tribe, eventually. As a high schooler, he and his friends tried new companies every six months.
That backbone would prove useful. It helped Brian to stay contrarian. So, when he read the Bitcoin whitepaper, he knew he was on to something. His mom told him he should probably come downstairs to spend time with the relatives out of town.
As Patrick Collison of Stripe points out, “Pessimists sound smart. Optimists make money.” Instead of being the smart skeptic, Brian was the optimist.
Lesson 2: Get Out There and Talk to Users to Understand Opportunities
Brian was a man possessed for the next weeks and months. He kept re-reading the paper. The trust problem, and the importance of decentralization kept seeming more obvious the more he read it.
He was worried he was actually too late to the whole thing:
It’s funny because I actually remember feeling like I was late to the game…. So actually the first thing I did was kind of trying to talk myself out of doing something in Bitcoin.
At the time, over 70% of all Bitcoin transactions were being handled by the exchange Mt. Gox. It felt like the world of NFTs and OpenSea, with its dominant market share scaring off competitors, today. Could anyone really compete?
2011
But Brian persisted. He went to a couple of Bitcoin meetups in 2011. These were the types of affairs with 10 attendees. 90% of people see the size of the crowd and turn around, to the more comfortable surrounds of their known friend group at the bar next door. But not Brian. He kept going, meeting the community, and learning their pain points.
Actually going out and doing that crucial user discovery work helped Brian have the insight that the main Bitcoin client was the only way to interact with Bitcoin. But the client was desktop only.
There was no Bitcoin wallet on the internet. There were a few things like a javascript library, but, generally, using and interacting with Bitcoin was both highly technical and difficult.
Lesson 3: Rapidly Release an MVP
So, Brian and a friend experimented with an open-source Android wallet. It was a quick open-source throwaway project. They released it without any marketing. It received a ton of traction: 15K downloads in two weeks, HackerNews #1, a Wired writeup. It’s fair to say things went much better than the duo had originally thought they might.
Because all the data was stored on the phone, Brian and his friend weren’t taking any risk with people’s personal data. So it wasn’t a complete solution. If something went wrong and the data was corrupted on the phone, the bitcoins were lost. The choice was intentional: they didn’t want to create a real company storing people’s data, because they both had full-time jobs.
However, the traction was signal enough that they should work on the complete solution. The idea was validated that people were really interested in Bitcoin, and if it became easier to use, people would flock to that solution. As Brian describes it:
The day we released it, I realized that somebody is going to have to make a really rock-solid cloud service for this. The future is going to be a thin client. While the blockchain was 300MB then, it was growing exponentially.
2012
With the proof in the pudding from the open-source throwaway project, Brian was almost ready to work on Coinbase full-time. Before fully making the leap, he used nights and weekends to build bitcoin products. One of Brian’s early insights was it had to be dead simple to send and receive Bitcoin, if it were to be used for payments.
He had cold-emailed YC partner Garry Tan about how to find a co-founder. Since Garry replied, Brian followed up by sending Garry Tan 0.05 BTC (today worth $2.1K) over email.
It was a remarkable example of a single engineer building fast. From there, Brian teamed up with the founder of Blockchain.info, Ben Reeves, to build a fully functional online wallet prototype. The two worked out kinks in the security and shipped it.
With the prototype in hand – though not launched – Brian was ready to apply to Y Combinator (YC). The Coinbase website and blog were set up.
Even at the time, competition to get into YC was fierce. The duo pitched themselves as having the radical idea that anyone, anywhere, should be able to easily and securely send and receive Bitcoin. As a result of the engineering work and clever pitching, Brian and Ben were accepted.
Getting into Y Combinator would be the validation needed for Brian to make the leap. In June 2012, Brian left his cushy Airbnb gig to work on things full time. Coinbase was born.
A day before YC began, Brian decided that he and Ben did not work well together. With a quick email, the duo became one.
Brian would make it through YC as one of the few solo founders. During YC, Brian soft-launched the product, signed up 15,000 people, and crushed Demo Day. This led to a $600K seed round to round out 2012.
Amazingly, Brian did all of the work in the second half of 2012, more or less, on his own. In addition to the consumer user interface, Coinbase released an API by September.
Lesson 4: Iterate to Make the Tech More Accessible
Brian would work days, nights, weekends, and vacations. Midday December 31st, he wrote “Sending Bitcoin to People You Know With Ease,” on the Coinbase blog. Brian was solving problems in the user experience of Bitcoin.
This time, it was about those pesky wallet addresses. Anyone who has used crypto knows the feeling: you have to triple check a 30+ character address. You don’t want to send the money somewhere you are not getting it back. As he said:
One of the more difficult things about bitcoin for the common user is dealing with the 30+ character addresses that are referenced to send or receive money. Rarely would one want to see these addresses — it’s much easier and more intuitive to refer to someone by their email address.
It was a genius solution, one precipitated by Brian’s work earlier in the year on Bitbank. The easy onramp would further the growth of Coinbase.
Of course, working alone would not be sustainable forever. By the end of the year, Brian would find his co-founder: Fred Ehrasm. A former Goldman Sachs trader, Fred would prove integral to Coinbase’s early success.
2013
Another important team member at the time was a security engineer in Germany who Brian met on Github. A contractor, he was an engineering savant and able to validate all the security holes and edge cases needed to operate an online wallet.
Together with the contractor, the team released several security features. A popular one was allowing you to export your bitcoin to a paper wallet. A paper wallet allows you to store your cryptocurrency offline. It is the dominant method still used to this day for large holders.
In addition, the team made the decision to keep the vast majority of customer funds offline. Instead of pushing security to the user, where if someone got your password all your money could be gone, the company took a more active responsibility. Brian was already thinking like a banker:
If you were to go to somebody like my mother say, “well, here’s your online banking, and, by the way, if you forget your password, all your money’s gone in your bank account,” that’s not really a good way to build a relationship with a mass-market audience.
This insight would be critical, and the foundation of the company’s future success. Brian identified that what was missing in the Bitcoin ecosystem was a mass-market, easy-to-use product. And, a big part of such a product is security. So Coinbase provided security-as-a-service to customers, instead of pushing the responsibility onto them.
The team takes the vast majority of customer funds offline, then does key-splitting into various parts. Those keys are then distributed all around the world geographically. Then, to actually use the keys, two need to turn at the same time. So, it creates redundancy where you need to hack two sets of keys. In addition, there is redundancy in the key splits so there is no system failure of one key split is lost.
The remaining funds are left in a so-called “hot wallet” so that users can take their crypto in and out of fiat currencies like US dollars. It’s usually around 5% of the total funds, but as volatility increases, the amount automatically increases to accommodate additional user activity.
In addition to security features themselves, Coinbase also built a massive test suite. Brian remained paranoid about people trying to attack the online wallet. By the end of 2013, the Coinbase engineering team had a suite of over 200 tests running.
Many people are puzzled by how Robinhood, for instance, can offer commission-free Bitcoin trades, while Coinbase charges 1%. This security combined with access is what customers are paying for.
Robinhood has had crypto for years but still does not actually make the currency available to users. So if you buy Ethereum on Robinhood, you can’t use it to buy an NFT on OpenSea. It is just an investment.
At Coinbase, the currency is actually yours, you can take it out and use it. Users are hiring Coinbase for the job-to-be-done of secure Bitcoin access and storage.
In addition to security-as-a-service, in November of 2013, Coinbase began insuring people’s deposits. The company worked with Aon, the world’s largest insurance broker, to cover losses due to breaches in physical security, cyber security, accidental loss, or employee theft. Although it did not cover crypto loss if a user was negligent in maintaining their private keys, but it was a huge step up from other exchanges.
This was very different from the Mt. Gox model. While Mt. Gox was trying to be the New York Stock Exchange and Fidelity, Coinbase was only trying to be Fidelity. It wasn’t trying to make any money on the spread. It was just sourcing for users. And then providing them security, and peace of mind, that those assets would not be hacked.
This set up Coinbase to focus on a different persona than most other platforms. Instead of focusing on the retail trader, Coinbase focused on developing more of a professional trader tool and B2B service. By addressing security and insurance, it made the product more palatable for those types of users.
The team’s wise decisions would lead Coinbase to dramatic growth over the course of 2013. At the start of the year, transaction volume was in the $15M a month range. By the end of the year, it was two orders of magnitude higher than that. Coinbase was amongst the hottest companies in the Bitcoin space. Along the way, it bagged a $6M Series A and $25M Series B.
2014
Lesson 5: New Tech Wins with Integrations
2014 would be the year of scaling integrations for Coinbase. Early in the year, it managed to surpass 1 million wallets. With all those wallets, the company wanted to make the new currency useful for payments.
Then, in February, Mt. Gox was hacked. $65 million in Bitcoin was stolen. It was the hack that rocked the crypto world. Bears were out in full force. But, for believers, more security-oriented Coinbase suddenly became more attractive.
Using the added horsepower from the acquisition of Kippt and Blockr.io, Coinbase would leverage that momentum to integrate with major consumer sites like Expedia, Overstock, Wikipedia, and Dell.
The dream was to enable anyone to buy and sell with Bitcoin, to build the new currency for the web. At the time, high fees and network congestion were not the problems they are today. Throughput was enough to handle those transactions.
In addition to going direct to the major consumer sites, Coinbase also went to their payments platforms. In doing its discovery work, as Brian said, Coinbase found developers preferred those services to have the Bitcoin option:
We’ve had a lot of developers tell us they’d love to add Bitcoin. But Braintree handles all of their payments, and they don’t want to add another SDK. They said that if Braintree added it, they would add it.
So, Coinbase integrated with Stripe and Braintree, allowing merchants using those services for payments could accept Bitcoin as a payment method. This would serve to massively increase the total addressable market (TAM) for Bitcoin payments, as Braintree at the time was the payment provider for Uber, Airbnb, and Dropbox – the three hottest startups of the era.
2015
The efforts to make Bitcoin a more accepted payment option would pay off. Even when an integration did not drive new users of the payments, it drove new users of the currency acquisition. Early adopters like Fred Wilson wanted to see where the technology could go.
Coinbase kicked off 2015 with a bang, becoming the most valuable Bitcoin company. It raised what was an, at the time, massive $75M series B.
Lesson 6: Ship Growth Product Features Fast
Brian would use that to staff up the team to 24 engineers, 3 designers, and 3 product managers by the end of the year. That team would ship an insane volume of products. The top 4 growth products to mention are:
Referrals is a tried-and-true fintech growth lever. By going straight to $25, Coinbase started out aggressive. The requirement was buying $100 of bitcoin. For a volatile asset, the bar was high, but a 25% near-instant return for those who sold soon. These days, that reward has been reduced to $10. But in the early days, it was a shot in the arm for growth.
The Index is more of a top of the funnel lever, that had long-term effects. The index is actually pretty significant. The NYSE administers many of the world’s most impactful indices, like gold and LIBOR. In fact, many TV tickers still refer to this bitcoin index when displaying the bitcoin price in 2022. So, what it did was make the currency more legitimate. CNBC would even add it to the bottom of the screen.
Eventually, all of that top of the funnel awareness flowed into Coinbase. Coinbase, unlike some exchanges, had a more professional trader in mind. So building awareness with this group is something that is still paying for Coinbase to this day.
International growth is the most straightforward growth strategy. But it is also worth double clicking on, for a second. Coinbase chose to expand to a place where it could use its English-language product. The main changes would be regulatory compliance, things like Know Your Customer (KYC) and Anti-Money Laundering (AML). So the company chose a natural second step. Most tech companies do this.
Recurring buys and sells was the final important growth product of the year. While the feature had existed on the platform, the team placed it much more prominently, directly onto the buy and sell page for non-recurrent purchases. This was a great example of the growth tactic to move the next transaction call to action (CTA) on the job to be done’s surface area.
These 4 growth products are a selection of the hundreds of features the team shipped in 2015. How does a product team go about shipping features so fast? Thankfully, in an extraordinary move of transparency, Brian wrote up their product organization’s process that year, in “How Coinbase Ships Product.”
The team stayed humble. Every year, each team was asked to brainstorm 3-10 key results. They were asked to think holistically about product market fit, instead of assuming it was already achieved. Once decided, these became the basis for the team’s OKRs.
Then, on a quarterly basis, the team performed an in-depth user study. This kept the team close to the problems Coinbase was solving. As Brian articulated:
The team needs to feel the customer’s pain and have this be a constant reminder that normal people don’t use the product like they do.
In addition, the team reserved one two-week sprint each quarter exclusively tech debt. This is one of the more standardized and interesting processes for dealing with tech debt as a PM.
On a monthly basis, the team talked to at least three customers one-on-one. The question to ask was if the team was solving the right problems. From there, the PM would send out an email of the results and write a blog post about one or more features that have shipped.
The focus on having blog posts ready every month is a particularly great discipline. It forces PMs to build user-facing features. It’s a bit of an echo of Amazon’s, “write the press release before you do the feature,” but for a tech-first software company.
All these processes the team put in place worked. The team would end the year at 3M wallets.
2016
Lesson 7: Move to Consumers
If 2015 was the year of growth product to grow the core, 2016 was the year of building out the product base. The company rebranded itself over the year, capping it off with Fortune profile.
The first big move was no longer being a Bitcoin company. Coinbase added Ethereum to the platform. Born three years earlier, the crypto meant to solve the problem of being hard to develop atop bitcoin would transform Coinbase into to a blockchain and cryptocurrency company.
The second move was taking a survey of where the industry was headed. In The Coinbase Secret Master Plan, Brian laid out the future of crypto. In his eye
The second phase is where Crypto was headed, and Coinbase wanted to be the one to carry Crypto to 10M. Having already reached 3M people, Brian realized that Coinbase’s positioning as a dual wallet and exchange was confusing to users. As he wrote in Coinbase is Not a Wallet:
We set out to build a bitcoin wallet, but it turns out we were building an exchange.
The thing is, in a retail exchange, users want an easy way to convert government currency to crypto-currencies. When a customer thinks they are using a wallet, but they are using a regulated exchange, frustration abounds. The team was receiving about things they were required to do by law: transaction monitoring, identity verification, and turning information over to the IRS.
The writing was on the wall that there needed to be an unbundling of wallets and exchanges. Wallets are necessary to actually do interesting things with Bitcoin and Ethereum. However, the demand for crypto-currency is was in buying and selling, not using. So, the company went to where the market was and rebranded itself as an exchange.
Lesson 8: Double Down on Growth Levers
2017
Ethereum would prove to be a boon for Coinbase. Users flocked to the secure wallet. Ethereum was a cryptocurrency built for you to do things with it, unlike Bitcoin. Coinbase’s wallet and security foundations would provide a sturdy competitive advantage for the now-exchange when users decided where to convert their fiat into crypto.
It also played with investors to become a holistic crypto company. So, in 2017, when Bitcoin announced its hard fork, the team announced it would support Bitcoin Cash. Then, the platform added Litecoin. Each crypto brought more enthusiasts of that currency to the platform and made Coinbase more attractive as a holistic platform for professional traders.
As crypto skyrocketed in value and cultural relevance at the end of 2017, so, too, did Coinbase.
2018
2018 would continue the “add currencies” strategy. Coinbase explored and added currencies like Cardano, Stellar Lumens, Zcash, and 0x.
Lesson 9: Don’t Panic – Build for the Future
However, the market does not always behave in rational ways. In early 2018, the price of crypto-currencies collapsed. Even the number of computers keeping Bitcoin running fell towards the end of the year. Years later, academic papers are still being written deciphering what happened in, “The Great Bitcoin Crash.”
While Coinbase experienced significant employee churn, Brian and the core team endured. Fred and Brian had existential conversations about their goals for economic freedom, made the necessary layoffs, and kept building. They even deployed capital strategically. In April, Coinbase acqui-hired crypto thought leader Balaji Srinivasan for $100M.
The company also continued its prognostications and thought leadership on the state of crypto. Years before the web3 fad of 2021, the company published the 22-minute read Understanding Web3 – A User Controlled Internet on its blog.
The outlines proved very prescient. On top of the state layer created by Bitcoin and Ethereum, the company envisioned the Ethereum Virtual Machine to be a computation layer. This is the processing that underlies everything else.
With that computational power sits the component layer. These are crypto goods like NFTs, and stable coins like Tether, USDT, and USDC. In Q4 of 2018, Coinbase added USDC to the platform.
Those components enable the protocol layer. These are decentralized smart contracts, like Uniswap. On top of that are scalability, user control, and, finally, user-facing applications like DeFi, e-commerce, and fundraising. This vision for the future set off a wave of product development in Coinbase across all layers of the stack.
2019
2019 would be the year that support across layers would truly materialize. Coinbase began, as it says in its S-1, truly “powering the cryptoeconomy.”
At the bottom of the pyarmid, the state layer, Coinbase continued the tried-and-true currency growth strategy of adding more cryptos like Dogecoin. Coinbase adding a new crypto was quicly turning into a cultural phenomena known as the “Coinbase Effect,” where a crypto’s value would surge as soon as the market got a hing of Coinbase’s endorsement.
At the top of the pyramid, the application layer, Coinbase continued its position as the exchange that gives you crypto you can use instead of just trade. The company added staking rewards. This is a taken-for-granted core crypto functionality today, but in 2019, Coinbase was early to helping users earn staking rewards for holding a currency.
2020
2020 would see Coinbase double down on support for the application layer, and in particular DeFi. The team added features to the wallet to continue to make it easier to earn interest through DeFi apps through staking.
It also released services such as Save, Borrow & Lend. These services had obvious benefit to the user, providing a more diverse and encompassing set of financial products in the cryptoeconomy. Coinbase aimed to become a “one-stop shop” for institutions to participate in the cryptoeconomy. But, they also had a sneaky benefit for Coinbase as well: more predictable, subscription revenue.
Lesson 10: Do What Everyone Else Does to Get Their Results
Product features were not the only thing on the mind of public as it relates to Coinbase. 2020 would be the year Coinbase made all the business talkshows for things not related to crypto.
The first news was Coinbase’s early move to remote-first. Mere months into the Covid pandemic, in May, Brian Announced, Post Covid-19, Coinbase will be a remote-first company. Being remote wasn’t slowing down the company.
This was a remarkable announcement at the time, that helped catalyze several other technology companies like Quora, Dropbox, Upward, and Affirm to do the same.
Going remote, in fact, could help the company. In March 2020, 7 out of 10 Coinbase employees were located in the Bay Area. The shift would help the company access top talent around the world. Now, 3 out of 10 are. Recruiting globally also helps build a more diverse workforce. Nowadays, Coinbase regularly makes best remote-first companies to work for lists.
The other big news for Coinbase in 2020 was catalyzed by a relatively innocoulsy titled blog post, Coinbase is a mission focused company. But while most of the principles in the blog posts were hard to disagree with, one stuck out given the broader global context. Brain argued Coinbase should not engage in social activism.
Post the Black Lives Matter protests rocking the nation, most companies were publicly taking a stand on the issue. The argument that Coinbase should not, and defended in a blog post with a picture of NBA players, disturbed more than a few employees. The decision was national news.
Coinbase chose not to back down. Instead, it offered a generous exit package to those who disagreed. And it moved on, shipping products.
With all the controversy surrounding Coinbase in 2020, it might have been easy to miss that the price of Bitcoin nearly 5x’d.
Moreover, Coinbase’s share of crypto market capitalization increased from 8.3% in 2019 to 11.1% in 2020. Not only was the market growing, Coinbase was taking a larger part.
2021
Investment bankers surely didn’t miss the revenue that the price increase of Bitcoin drove. In 2020, Coinbase generated $1.3B in revenue, 2.4x the amount of revenue it generated in 2019. The stage was set for the company’s IPO.
On February 25th, we got the first intimate look at Coinbase, with its S-1 hitting the sec.gov website for everyone to analyze. Then on April 14th, the company completed its direct listing IPO. Shares spiked 70% within minutes of trading open, valuing the company at $100B.
It was a coronation for a little over decade’s work by Coinbase founder Brian Armstrong. In a day his, $10B paper net worth became $20B in liquid, tradable shares.
Since then, the company has continued to execute. In the most recent quarter, Coinbase’s Q3, net revenue more than quadrupled. It is executing on its latest strategy to support crypto as: an investment, new financial system, and an app platform.
And it has continued to pontificate about the future of crypto. The day before new year’s Coinbase’s new Chief Product Officer dropped 10 Predictions for Web3 and the Cryptoeconomy for 2022. It might be Coinbase’s best take on the future yet. The things Coinbase predicts don’t so much come true as Coinbase makes them come true.
2022 and Beyond
At the time of its S-1 in April 2021, Coinbase supported 90 crypto assets for trading. Today, that number is over 140. Coinbase has developed a virtuous flywheel of compounding craziness.
Nevertheless, it is hard for the decentralized idealists of cryptopia to get a grip on Coinbase. On the one end, no company has done more to onboard fiat currency into the cryptoeconomy. On the other end, Coinbase represents a centralized onramp into that economy.
Public markets are not quite so sure to do with the stock, either. At a $50B valuation today, investors at the peak of the IPO have lost half their money. Those dollars traded away are hard to stomach.
As such, Coinbase has not quite generated the goodwill of an Ethereum, Solana or Uniswap. These judgments are perhaps rightly so. But it is hard to deny: as the largest dollar onramp to Crypto in the US, Coinbase has changed the crypto world forever.
