Do Things That Don’t Scale: Crypto Edition

Qiao Wang
Alliance
Published in
16 min readOct 27, 2023

Exactly 10 years ago Paul Graham wrote an essay titled “Do Things That Don’t Scale”. In my opinion, that was the single most important piece of content YC has put out for the startup community.

The core idea is that in the early stage startups must 1) manually recruit users and 2) manually give them an extraordinary experience. These are things that almost no big companies do, therefore they are thought of as things that don’t scale.

Since starting AllianceDAO 3 years ago, I’ve been thinking about how much of this advice applies to crypto startups. After 3 years and nearly 200 startups, I have enough data points to answer this question.

But before I give you my answer, let me walk you through some of the “things that don’t scale” that AllianceDAO alumni have done to recruit and delight their users.

Synthetix (ALL1)

Synthetix is the OG derivatives trading platform and consistently one of the top 100 protocols by FDV.

Kain recently came back and gave a talk to ALL11 founders. He said something that would surprise most people: “you have to play the short-term narrative game.”

When Synthetix first started, which was during the first innings of DeFi Summer 2020, the prevalent narratives in the DeFi community were “liquidity mining”, “token staking”, and “TVL”. In 2023, these concepts are by and large no longer fashionable. In fact, it has become consensus, for example, that liquidity mining is a net negative for startups.

But in 2020, Kain embraced these things. They did liquidity mining. They did token staking. They talked about TVL all the time in their Discord and Twitter. And it worked. They amassed some of the most influential people such as DegenSpartan as their most ardent supporters.

Now, I’m not saying the likes of liquidity mining still works today. But the advice of embracing short-term narratives still stands, especially when you are building a crypto-native product. This is because the crypto-native community is still fairly small and everyone is within 1 or 2 degrees of separation from each other.

On top of that, Kain was the single most active member in their own Discord and constantly engaged with the Synthetix community. Today still, he regularly puts out blog posts about Synthetix, obviously with the goal of acquiring more users.

0x/Matcha (ALL1)

0x is one of the first DEX protocols. They also built Matcha, which is a leading DEX aggregator.

0x has always been ahead of its time. When 0x first launched many years ago, professional market makers were skeptical of DeFi and DEXes. It was nearly impossible for 0x to onboard them to provide liquidity.

So the 0x team built their own internal market-making infrastructure (which spun off into its own entity called Periscope Trading). The idea is twofold. First, to dog-food their own liquidity. This helped solve the cold-start problem whereby professional marketmakers were only interested in exchanges that already have sufficient liquidity. Second, to walk in the shoe of market markers and experience first-hand the pain points of providing liquidity. This helped improve the onboarding experience. Eventually as more organic liquidity came in they sunsetted the market-making infrastructure.

Ribbon/Aevo (ALL2)

Ribbon was the first and most active structured product marketplace. They are now building Aevo, a leading options protocol.

Julian also came back and gave a talk to ALL10. Like virtually all other successful DeFi protocols, he manually reached out to the first users. Interestingly, he noticed that there were the same 20 people in every options protocol’s Discord, so he messaged everyone of them and brought them into the Aevo community since they already had a natural interest. The DeFi derivatives community is still quite niche, so it’s possible to speak to all of them.

But he did something even more extraordinary. Because the onchain options liquidity is overall quite small, some users of Aevo can’t find enough entry or exit liquidity. Julian would manually source the liquidity for them, and post a bid/offer on their order book.

Zerion (ALL3)

Zerion is a leading browser and mobile wallet.

Zerion started off as a DeFi portfolio tracker years ago. It had some immediate signs of PMF as users needed a tool like that to track an explosion of DeFi assets. The natural next step for Evgeny to think about growth.

He met with basically every DeFi team in 2019 and told them about the product. Notably, he integrated 3 of the first DeFi protocols, namely Uniswap, Compound, and Maker, directly into the Zerion interface. The combination of these personal relationships and product integrations led to public support from these teams and thus greater distribution.

Nowadays it has become consensus for front-ends such as wallets and DeFi apps to integrate third-party protocols. But back then Zerion was one of the first front-ends to leverage permissionless composability to create distribution for themselves.

Mux (ALL3)

Mux is a top 3 perpetual swap trading protocol on Arbitrum.

Mux, at the time known as Mcdex, launched on Twitter in 2020. Then serendipitously DeFi Pulse and Bankless found out about Mcdex, liked the product, and reached out to Jean and Jie. The latter took the call with DeFi Pulse and Bankless (separately), negotiated for a deal where DeFi Pulse and Bankless would write an article about Mcdex in return for a referral fee. The product took off from there.

But more importantly, Jie doesn’t sleep. He does the typical 996 thing that every Chinese startup does. 996 means 9am-9pm, 6 days/week. I heard this has evolved into 997 now. I lurk in their Telegram group, and I would see him reply to users at 11pm local time. Mind you, he’s a technical founder. I rarely see technical founders do this.

Dodo (ALL3)

Dodo is a leading DEX with 5% market share.

Dai Dai worked at DDEX in 2018, another OG DeFi protocol. This was two years before she started Dodo. During her free time she started a blog called “DeFi Labs” where she would would produce educational content on DeFi.

She then realized she wanted to get to know her readers personally. So she pulled them into a WeChat group, called “DeFi the World”. She would also organize regular IRL meetups with the members. This became the first and largest DeFi community in China. Make no mistake, this was during the last bear market and DeFi wasn’t really a thing yet, so the “largest” meant ~100 people.

When she started Dodo in 2020, these 100 people became the first 100 users of Dodo.

Pendle (ALL4)

Pendle is the #1 yield trading protocol today.

I vividly remember TN invited me to give a talk at their internal all-hands with the aim to uplift team morale. This doesn’t scale for Pendle nor for me, but the moment I got the invitation I said to myself “TN is special”.

But he did way more than this. In the early days, he reached out to potential LPs (liquidity providers) and booked 1:1 sessions to pitch them yield opportunities on Pendle. He didn’t have any sales experience before Pendle, but with enough attempts, he got better at pitching and started seeing conversions. Indeed, most founders don’t realize that they can become great salespeople and that sales is a number’s game. He also shared yield trading ideas every now and then with the potential LPs. As they became more comfortable with the team and the protocol, they converted into real users.

Similarly to Synthetix, he also identified influencers he wanted to work with, and reached out to them via cold Twitter DM or cold email. Most decent founders do this today, but TN went the extra mile and did the homework. For example, if he wanted to work with a certain influencer, he would research the said influencer’s interests and habits and write the material that he thought the said influencer could use publicly. He did that for a while, until he started seeing more organic mentions on social media.

Charmverse (ALL7)

Charmverse is the leading web3 community operations platform. (2 years ago the competitive landscape was cutthroat but most of their competitors died in the bear market.)

The first version of Charmverse was a token-gated Notion. But building a Notion is obviously incredibly difficult. So Alex and Matt hacked the Notion internal APIs and offered the token gating feature as an admin Notion user. (At the time, Notion hadn’t released their public API, but they have now.) This was how they got their first DAO customers.

And of course, Alex went to a bunch of crypto conferences, built relationships with key decision makers with DAOs, many of whom wrote DAO proposals to adopt Charmverse that ultimately got passed by the DAO votes.

One extremely laborious thing Alex did to convert existing Notion users to Charmverse is to transition, by hand, their Notion content to Charmverse. Then he held onboarding calls with the DAO members to teach them how to use Charmverse.

Hubble (ALL7)

The Hubble team built Kamino, the #1 marketplace for yield vaults on Solana (the Yearn of Solana).

When FTX collapsed in November 2022, everyone (except for me) told Marius and Thomas to ditch Solana and move to an Ethereum L2. The thought of this crossed the mind of every Solana builder.

I remember talking to Marius and Thomas about this week in week out for a few months. Eventually, they concluded that Solana was here to stay because:

  1. Solana has a soul: there were still a number of devs who were crazy about it.
  2. The tech is actually legit. It’s fast and cheap. The downtime issues at the time were solvable engineering problems, and they were indeed solved eventually.
  3. Anatoly is the most inspiring L1/L2 leader after Vitalik.

And meanwhile, it was unclear how to compete with the incumbents on Ethereum if they were to migrate.

Marius and Thomas made a very difficult decision, and in hindsight a very good decision, of sticking to a smaller but growing market that they understood intimately (Solana), instead of a bigger more scalable market where they had no unfair advantage (Ethereum).

Synquote (ALL7)

Synquote is the #1 decentralized options trading protocol by trading volume.

When Synquote went live, Ahmed made sure that he and the team were responsive to user requests 24/7. One morning, an onchain whale wanted to do a large block trade. The team built an ad-hoc OTC product within a couple of hours, sourced liquidity for them, and arranged what was at the time a record trade of over $3M notional that afternoon.

An unexpected benefit came with being available to users 24/7. Those who saw the product rapidly improve firsthand with their involvement, and got to know the team, became champions of the product.

Tensor (ALL8)

Tensor is the #1 NFT marketplace on Solana.

Ilja and Richard used an onchain tracking website to identify all the Solana NFT whales, find their Twitter handle, DM’ed everyone of them, and got on a call with the few that responded.

But this wasn’t enough. At the time they were fighting an uphill battle against Magic Eden which had orders of magnitude more distribution. So they built an alliance with other NFT marketplaces. Notably, they struck a deal with Hadeswap, which at the time had 10x more distribution than Tensor, whereby Tensor would build a brand-new front-end for Hadeswap in exchange for a backlink to Tensor’s website. This dramatically increased distribution for Tensor.

Of course Tensor made a number of other great decisions in its short history but this partnership with Hadeswap was a pivotal moment.

Liquifi (ALL8)

Liquifi is the #1 token vesting solution for crypto protocols. Think of them as the Carta of crypto.

My office hours with Robin and Oliver are some of the most enjoyable ones I’ve had. They would come to every meeting extremely well prepared. They would give me a list of other AllianceDAO startups and ask me one at a time if I think they are a good fit as customers, and then reach out to them. By the way, a number of AllianceDAO startups acquired their first few customers through our community.

Robin and Oliver would also comb through all the public fundraising data such as Twitter, Messari, and Crunchbase, to find all the possible customers. They would ask every investor for warm intros to key decision makers. They would even jump into random Discords and literally ask everyone for intros.

Robin describes Liquifi’s style of customer support as “unhealthily high touch”. For example, the team would check the vesting schedule spreadsheets that their customers send to help reduce onboarding friction. His team has spotted tons of mistakes, and in one instance, proactively called out a scenario where someone was being underpaid for 2 years. Handling token vesting is such a sensitive topic, and this level of diligence helps customers feel comfortable with working with Liquifi.

Clique (ALL8)

Clique is the #1 offchain credentials issuer.

During the Alliance program, Jaden interviewed over 200 protocols to validate use cases for off-chain credentials. They found their first strong use with the Lens protocol, that is, sybil resistance.

Jaden got in touch with the Lens team. At the time Jaden had just moved to New York. The Lens team was also based there. After 3 weeks of back and forth the Lens team had finally agreed to an in-person meeting. However, on the day of the meeting, the team mentioned that they had flown off and would be at ETHSF instead. Without a ticket to the event, Jaden flew to SF the same day and waited outside the venue until the end of the event to make his way into the venue, eventually finding the Lens partnerships lead. He was finally able to get a pitch through and got Lens to sign the deal.

Jaden and Kevin then built a bunch of adhoc features to keep Lens happy. Jaden semi-jokingly told me “Clique is the least scalable product in crypto.” But ultimately Optimism and other L2s noticed that Lens used Clique, reached out to Clique, and became customers too. The things that don’t scale immediately could scale serendipitously.

Yakoa (ALL8)

Yakoa protects some of the household Web2 and Web3 brands from onchain intellectual property thefts.

Graham’s B2B sales strategy is “just be there in person”.

With one customer, Graham and Andrew flew across the country, twice, during the course of the negotiation. They worked out of the customer’s office for a day getting to know the whole team and the dynamics. They sat down with multiple levels of leadership that would have been almost impossible to coordinate a call with to get their buy-in. They went for dinner with the individuals that would do the actual onboarding. And throughout all of this they were able to identify multiple additional pain points they could assist with, and ultimately cross-sell.

Stride (ALL9)

Stride is the #1 liquid staking protocol on Cosmos.

When Stride first launched, Aidan, Vishal, and Riley took turns to be on call 24/7, until they were sure the protocol was secure. For the first two weeks, one of the three would wake up at 3am every night to make sure the protocol’s periodic IBC packets landed properly on Cosmos and that the protocol’s accounting lined up.

While most founders are active in their own Discord and Telegram, the Stride guys went out of their way to monitor all the other main Discord servers where Cosmos people hung out (eg, those of Osmosis and Mars) and to answer all questions. This is because people in all the other Cosmos zones could use Stride to liquid-stake their native tokens. Eventually they even set up cron jobs that crawled all these Discord servers and dumped relevant messages into their own Slack, so they could answer them without having to monitor them so actively.

Primodium (ALL9)

Primodium is the most popular fully onchain game by DAUs.

Emerson and Morris come from somewhat of a non-crypto-native gaming background. With the help of Will Robinson on our team, they spent a year building a solid network within the onchain gaming community. Similarly to Ribbon and the onchain derivatives community, the onchain gaming community is niche and it’s possible to get in touch with most members of the community that “matter”.

So Emerson and Morris simply launched the game on Twitter, and told everyone in their network about it. Then the game spread via word of mouth in private, before it started spreading organically on social media. There’s even a random Wall Street Journal article that mentions it.

The lesson here is that when the community is small, you don’t have to do anything dramatically creative to get off the ground. Spend some effort to build tightnit relationships. And those relationships will pay off on launch day.

Sleepagotchi (ALL9)

Sleepagotchi is a gamified sleep tracker, and the crypto consumer app with the single highest 30-day and 90-day retention I have seen of any crypto consumer app.

Anton identified the Axie Infinity community as potential early adopters for their app. Back then, “play2earn” was all the rage and Sleepagotchi presented themselves as “sleep2earn”. By the way, this ties back to Kain’s point about “playing the short-term narrative.” Anyway, Anton offered users on the Axie Twitter to “3D-fy” their characters for free. Twitter users would share their favorite 2D Axie, and they would create a 3D rendition on it. This brought them to the attention of Axie founder, JiHo, who retweeted their work. JiHo also ended up angel-investing in Sleepagotchi after.

Anton did something similar with the BAYC community too. He knew Eminem had an ape, and decided to 3D-fy his NFT, singing in front of an audience of Sleepagotchi characters. Luckily Eminem and Snoop Dogg released their song during NFT NYC in June 2022, and Anton released the 3D-fication the morning after (since the video had already been made). The tweet went viral, getting 30K impressions within a few hours.

Importantly, Anton converted potential users into their Discord, rather into their mobile app, ie, they focused on community first. This is because jumping into a Discord is slightly less friction than installing an app. Anton found out that many people would enter the Discord as a joke, just to check out the funky-sounding “sleep2earn” project. But to every question, serious or not, Anton’s team would reply with 2–3 paragraphs. People who were first curious about the idea realized that the team behind took the idea seriously, and eventually downloaded the app.

Refract (ALL10)

Refract is the #1 “crypto anti-virus” for consumers.

When Refract first launched, an NFT platform called Premint got hacked and Nish DM’ed everyone who said they got hacked on Twitter. This wasn’t some obnoxious “you could’ve used Refract” kind of sales but rather “let us learn about what happened”.

And then, everytime a major crypto hack occurred, Nish would write a post-mortem on Twitter. I’ve seen some of those Tweets and they got more engagement than virtually anything I’ve seen in crypto because people get emotional when they lose money. One time Nish and Justin showed me the metrics, and there was a staggering correlation between spikes in new users and occurrence of major hacks.

With existing users, they manually DM’ed 1000+ of their Twitter followers to ask the superhuman PMF questions, e.g. “how disappointed would you be if Pocket Universe disappeared”. This data point gave them strong validation that they indeed had PMF.

Caldera (ALL10)

Caldera is a leading rollup-as-a-service (RaaS) provider.

The biggest thing Matt did to sign their first few B2B customers was building small ad hoc features for them. Each feature would take 2 to 5 engineering days. One such example is support for secp256r1 Curve Support (EIP7212). And yes, it’s so esoteric that most founders wouldn’t think of it as something that would move the needle.

But the result is that work like this led to a lot of goodwill. The customer was so delighted that Caldera was willing to provide such high-touch and tailored support, that they organically told their friends about it, which led to even more deal flow. Remember, in crypto B2B, customers all know each other.

Funny enough, Jaden from Clique is Matt’s friend. And in the last 2 years the most frequent thing Jaden told Matt was “are you sure this scales?”

Guardrail (ALL10)

Guardrail is a leading security detection and investigation tool for protocols.

Their first customer happens to be another AllianceDAO alum, Pendle. Sam joined their Discord, gave them feedback as a user on their product, connected with their dev-rel lead, then asked him to connect with their internal tooling lead, who eventually became a core user of Guardrail. I know this sounds confusing. But the point is that with B2B customers it often takes time and a few hops to reach the internal champion and the decision maker.

Furthermore, this customer (Pendle) is in Singapore so most of the back & forth happened at a 12-hour time difference, so Sam did all the customer support between 12am and 2am local time. After the product demo, the customer had a few unique use cases, such as an auto-refresh so they can keep Guardrail on all day on a TV. Sam and James built it for them in the same week.

Kravata (ALL11)

Kravata is a leading fiat on-off-ramp provider in Latin America whose customers include the largest Latam exchange.

Felipe understood that behind tech businesses there are humans. And humans like to eat, to drink, and to talk to other people. I’ve found this to be particularly true in countries where the rule of law is weaker, in which case relationships matter more.

To acquire his first customers he scheduled in-person meetings with them for a coffee and a good dessert. In that first coffee he would always offer to pay the bill, and he would pay from his pocket and not from the company’s pocket. Like the Yakoa guys, sometimes he would fly to the customer’s physical location just to meet with them. He would talk about common concerns such as regulations that everyone can relate to. Sometimes common friends would arise in the conversation which is useful as you build the relationship around homophily, ie, the tendency for people to relate to others that are similar to them. He would offer business advice and helpful introductions without expecting anything in return. If you have met Felipe, you would agree with me that he’s the very definition of humility and honesty, so that helps too.

Importantly, he would do all these things to make the customer comfortable before direct sales.

Do Things That Don’t Scale

After reading all these stories, I hope you have reached a similar conclusion as me. The “do things that don’t scale” advice by and large still applies in crypto. Whether their users are consumers or businesses, whether or not they have a token, some of the most successful startups in crypto have done extremely manual and laborious things to acquire and to delight their users.

There’s one caveat though. Crypto does offer a native scalable way to acquire early users, and that’s token incentives. Token incentives are probably one of crypto’s biggest unlocks. Other big unlocks being asset ownership (Bitcoin), data ownership (crypto social), permissionless programmability (Ethereum), fundraising (ICOs), permissionless financial services (DeFi), culture (NFTs), and tokenization of everything (RWAs).

Token incentives have evolved a lot since Bitcoin. From immediate rewards, to retroactive airdrops, now to the point-not-token system. Empirically, these are extremely powerful, scalable user acquisition strategies. So they are so powerful that they deserve an deep dive of their own.

Nevertheless, token incentives are not the panacea. They are not a replacement for things that don’t scale. They are complements, rather than substitutes. Many of the aforementioned AllianceDAO alumni launched token incentives while doing the laborious manual things.

Even the crypto protocols where the token is so instrumental that it is arguably the product itself, such as DePin, stablecoins, some games like StepN (ALL7), and L1s, first got off the ground by leveraging their friends, investors, and random people on Twitter.

In 2008 Satoshi exchanged a number of emails about Bitcoin with his/her/their ​​cypherpunk friends including Hal Finney and Wei Dai, who in turn evangelized a bunch of early Bitcoiners on bitcointalk.org and in person. The Ethereum founding team went on world tours in 2014 to evangelize a small handful of early developers, and those world tours include the famous Devcon 0.

--

--