In 2021, Eric Zhang was the CTO of Scalable, a company that helped small e-commerce businesses scale to sell globally.
This mostly meant mastering advertising through Google and Facebook, because they had all the user data. Google and Facebook knew what everyone did on the internet—the sites we visited, the things we googled—so they could show ads for Scalable customers to middle-aged women who liked dirt bikes. Or to young men. Or to people searching for “cute cat scarves.”
Then 2021 happened.
That year, Apple released App Tracking Transparency, which made it harder for apps to track users’ activity on other apps and websites. Suddenly Facebook found it harder to differentiate young men from middle-aged women who like dirt bikes.
Facebook lost billions from the change. Scalable was hit harder. They had to lay off staff to stay solvent.
As Facebook caught a cold that gave his company pneumonia, Zhang thought about user data in crypto, or Web3, which he’d been interested in for years.
In Web3, no one has a monopoly on user data like Facebook and Google do in today’s Web2 internet. The user data is all public—on different blockchains for anyone to see. If Fred from Minnesota buys artsy NFTs on OpenSea, OpenSea learns something about Fred. But you or I can see those purchases, too. We know almost as much about Fred as OpenSea, and we don’t have to pay OpenSea for that information.
This was a lightbulb moment. Zhang’s company was being hammered because it did not have direct access to user-behavior data. But even though this data was public in Web3, freed from Facebook and Google’s walled gardens, crypto companies weren’t using it for marketing and advertising.
It seemed like a huge missed opportunity. So, in 2022, Zhang and a cofounder, Andrew Li, raised money and launched a company, 2.5 Intelligence, to use blockchain data to improve Web3 marketing.
A year later, they still believe in their thesis. But they’ve also learned that there’s a major obstacle keeping it from working. To overcome it, they are making a big bet, taking the company in a direction that Zhang describes frankly as “batshit.”
A Startup Story
There was a good reason that Web3 companies weren’t using blockchain data when marketing their digital wares—it was hard to organize and use.
“I think people have a very idealistic view that blockchains are organized, elegant systems,” says Zhang. “But it’s messy, like any corner of the internet. Blockchain data is ugly. It's changing. It's inconsistent.”
To see whether crypto companies would find cleaned-up data valuable, Zhang and Li ran an experiment. They looked at blockchain data on NFT sales and made Excel spreadsheets showing which NFT owners had “diamond hands” (meaning they held onto their NFTs regardless of price jumps) and which had “paper hands” (the opposite).
“We went into random Discords and said, ‘Hey, would you like to buy this report for a few hundred bucks?’ And some people bought them.”
“That's literally how the company started,” Zhang adds.
More recently, 2.5 Intelligence used a similar approach to help one of their customers: GodHatesNFTees, a cheeky collective of NFT communities. Using data from 2.5 Intelligence, GodHatesNFTees identified their most diamond-handed users and rewarded them with tickets to a special raffle.
This was very different from Zhang’s Web2 past of using Facebook and Google to identify demographic groups and place ads in front of online shoppers. But Web3 is young, and Zhang never wanted to exactly replicate Web2 marketing for crypto. After all, Web3 offered new possibilities.
Consider the case of Amazon’s recommendation engine.
“If I go to Amazon, I have all these great recommendations because I've made a lot of Amazon purchases,” says Zhang. “But then if I go to eBay, it's not going to make good recommendations for me because I haven't made enough eBay purchases.”
But on Web3, eBay could see all those Amazon purchases and use them to give Zhang recommendations on eBay. Or, to use a crypto example, a just-launched NFT marketplace could use blockchain data on OpenSea sales to power its own recommendation engine before it facilitates a single sale.
“I think what's most interesting is not just how we can mimic Web2,” says Zhang, “but how Web3 data can open up these new paradigms for marketing.”
Make Something People Want
Today, the majority of 2.5 Intelligence’s customers are NFT collections such as Check VV, Art Gobblers, and Pirate Nation.
A number of them pay for intelligence reports and sales data that 2.5 Intelligence gets from blockchain activity. It’s a bit like a traditional company handing their user data to a consulting firm that will then make powerpoints and spreadsheets with insights like “Your customer base is growing faster abroad than in the United States.” Except 2.5 Intelligence draws on data from the entire Web3 ecosystem.
But as Zhang and Li talked with the heads of NFT collections, they learned that what many of 2.5 Intelligence’s customers really wanted was more ways to engage their audience. Which usually meant having more features in the Discord (chat app) that owners of their NFTs got access to.
That’s why, today, 2.5 Intelligence’s most popular tools are Discord bots. One popular bot, their NFT Mint Bot, tracks the release of new NFTs in real time and posts about it on Discord. Another, their NFT Whitelist Bot, makes it easy for NFT creators to collect the wallet addresses of their NFT holders and run giveaways.
Although Discord has become the go-to platform for NFT creators looking to cultivate community among their NFT holders—a place for fellow CryptoPunks or Squiggles to hang out online—its original and still-dominant purpose is for gaming.
This has led a number of crypto developers and founders to question Discord’s dominance in Web3 communities. But Zhang thinks this is misguided.
“I feel like Discord is sufficient and not going to be replaced anytime soon,” he says. “I guess I'm biased, but I do feel like more companies should be building their products not on the web, but within Discord.”
Center and the Discord Bots
As 2.5 Intelligence built their bots, they found themselves looking for a way to avoid a thorny problem: rendering NFT images on Discord.
“We want an image to represent this NFT pulled from the metadata. That's extremely difficult,” says Zhang. “It sounds silly, but it's totally non-trivial.”
“We initially deployed Center as a backup for Alchemy, because Alchemy had a similar API. But we noticed that the failure rate for Alchemy was very high, and Center would catch cases that Alchemy’s API just could not figure out.”
That’s how we at Center ended up working with 2.5 Intelligence. As we’ve written about before, our API makes it easy for companies to render NFT images. It navigates all the messy blockchain data so that, in cases like 2.5 Intelligence’s Discord bots, an image of the specific Bored Ape or Pudgy Penguin appears.
“So many of our bots show NFT images,” says Zhang. “And if they're showing a NFT image, chances are that data is coming from Center.”
Cooking Up The Bakery
Talking to customers led 2.5 Intelligence to build Discord bots. But Zhang doesn’t see bots, by itself, as a sustainable business.
“We have premium features,” he says, “but the 2.5 intelligence Discord bots are almost more like a public good—most of the functionality is free.”
So 2.5 Intelligence faced a quandary: Their bots were not a viable business, and few companies wanted to do the kind of marketing 2.5 Intelligence had in mind, which involved harnessing user data to target specific customers with ads.
Zhang sees two reasons for this.
One is that Web3 marketing is different from Web2. In crypto, people hawking NFT collections or DeFi products launch and grow with “hype marketing.” Rather than place ads in front of specific users à la Web2, they talk up their products on crypto podcasts, create tokens, and foment FOMO by posting on Twitter and Farcaster (now Warpcast).
The other is that “the Web3 population is just too small.” If more people made regular purchases on-chain, companies might want 2.5 Intelligence’s help finding and targeting specific demographics. But with every crypto project chasing the same, small group of Web3 users, all that matters is targeting crypto enthusiasts, and parsing blockchain data doesn’t really help with that.
“I think that’s the biggest problem for everyone right now,” says Zhang. “The growth in Web3 is not there right now.”
So rather than wait for the market to grow, 2.5 Intelligence has launched a project that Zhang describes as kind of wild. It’s their own foray into hype-based Web3 marketing, and an effort to expand the Web3 population and induce demand for their services.
It’s called The Bakery. It’s a Web3 project. But rather than a blockchain, it has a “Butterchain.”
Like any good hype marketer, Zhang plays coy about the eventual, full feature set of The Bakery. But the thesis behind it is that most people don’t really care about fully decentralized blockchains—and that crypto founders’ devotion to full decentralization is making the user experience so difficult or annoying that the space isn’t growing.
As Zhang puts it:
“Imagine something exciting is happening in crypto and you go to your friend, who has never used Web3, and say, ‘Hey, you’ve got to get in on this. This is super cool.’ They’re convinced, but they have to go to Coinbase, connect their bank account, maybe make some micro deposits. Then, days later, after they’re approved, they have to pay ETH for gas fees. And only then can they do the cool thing.
That's bullshit. What in Web2 has such high friction? Nothing. Nothing in Web2 with such high friction has been successful.”
In contrast, joining The Bakery and accessing its Discord just takes a few clicks. And you don’t need to pay thousands of dollars for your first NFT (since Zhang and co. regularly give away NFTs), and you don’t need to connect your bank account to pay gas fees in ETH (since you receive a chunk of $BUTTER, the currency of the Butterchain, when you sign up).
The Bakery doesn’t need users to pay gas fees because while the Butterchain is transparent, it’s not fully decentralized. So there’s no need to compensate nodes or users for verifying the Butterchain.
Instead of decentralization, 2.5 Intelligence is focused on making The Bakery fun and engaging. Which means, for example, running raffles for NFTs and hosting games on the Bakery Discord where users can wager their $BUTTER. Since The Bakery just launched a few months ago, 2.5 Intelligence is more focused on building and exploring than on monetization.
Zhang is aware that The Bakery does not sound like a page in a marketing startup’s playbook. But with the growth of Web3 users stalled—due to factors like rising interest rates and fears of a recession—at a size too small to support robust demand for their data-driven marketing, they’re determined to do something, to make an all-in bet to grow the pie.
“Even if we blow up,” says Zhang. “I’m not going to wait out the bear [market].”