TLDR:
Many web3 products are mostly web2, for which you can use your usual user analytics playbook. For on-chain data, analytics are in their infancy and bad at presenting quantitative results. A big step will be building Google Analytics that also sees on-chain data.
Web3 rebuilds the inner workings of the internet. Early crypto-anarchists promised a decentralised and anonymised future without central banks. It doesn’t seem that we’re moving this way. For now, at least whilst web3 remains such a small percentage of the overall market, the best way to grow is to build bridges between these worlds. This also helps to explain why almost all of web3 is not doing well on fulfilling its promise of anonymity, tamper-free ledgers and decentralisation.
There are at least three pressures that keep companies from fulfilling the crypto-anarchist vision:
Governments don’t like it
It’s harder to build in Web3
To build something people love, you have to understand them
I want to talk about an aspect of the last one: analytics.
For businesses, it’s really important to identify customer profiles. To understand who buys, where they are from, what age group they’re. You can then identify your most valuable customers. This is important, because 80% of your money will come from 20% of your best customers. Also, now you can do outreach and find more similar folk.
It’s also important to understand how people interact with your product. What stops them from checking out, which button placement works best, what big image and wording to use on your landing page, etc. For instance, at my previous company, Knit (co-founded with Orest), we had single-sign on for several social network sites, but it turned out 0% of our 10k users used Twitter, so we removed it. We also saw people preferred to talk in smaller groups, so built the UX around that. Another insight was that most of our most valuable customers came from Germany and the UK, so this informed out go-to-market strategy. We gained these insights through traditional product analytics.
So how is it done in web3? We had some great conversations about this with Elaf Deyab. She talked to a dozen or so people on how they do their analytics. I’ve written this mostly based on Elaf’s notes from conversations with experts. So let me share my first impressions of the space.
So wallets (web3 word for account) are anonymous, so it’s super hard to get things like demographic data. You can scan the blockchain and see all the transactions from a wallet, what was traded with which wallets. However, you don’t know who is behind them. And people might have dozens of wallets. And largely, that’s the point.
However, as a builder, you want to know who uses what you’ve built and how they use it. So how is it done currently in web3? Most of web3 is actually web2, which makes analytics way easier. There are two layers:
The back-end on-chain stuff, transactions, wallets, tokens, and whatnot
The front-end stuff, which is almost always just a web2 website
The former seems new and poses new challenges, for the former, just use your old web2 playbook. Companies will lie on a spectrum on these. Coinbase has value because of the front-end stuff. Ethereum has value because of the back-end stuff.
There is another spectrum - how deep you can follow the user in their journey. Coinbase can follow you 100%. They have to. A conversation with policy expert Marika Tedroff (ex-TikTok) revealed how important this is, both because of regulatory pressure and just to build a good product and community. They have KYC, one account per person, they can see all of your trades. You even give them your exact demographic data. They see each button you click and also that time where you put in the wrong card details three times. And that’s all from their web2.0 playbook.
Somewhere in-between will be many DAOs, where your actions can be tracked but you could remain anonymous. However, there is usually little incentive to do so. Think about LexDAO, a community around building smart contracts. No point in being completely anonymous, quite the opposite, you want people to be able to reach you and you want to build a reputation as an expert. So you opt for pseudonymity. It’s also probably wise to show-off some off-chain credentials like your Law degree from Harvard, so might just reveal who you are.
It gets even more interesting when people want to show off their overpriced monkey JPGs and it turns out it’s really hard to show off that you own one. Whilst you can buy an NFT and its unique, it’s not as obvious that it’s you who bought it. There isn’t much stopping me from displaying your JPG as my profile pic (shout out to Eike). So this is a case where people actively want to show off their identity, but can’t.
The other extreme sees completely anonymous products. Anonymity is a feature here. For instance, the jurors in Kleros’ dispute resolution might have an incentive not to be tracked, because a bad actor might take revenge on their decision. Or you’re using some crypto to buy prescription pills on the darknet. Many legit reasons.
Back to the ‘back-end’ on-chain stuff. The big problem is that once stuff happens outside of your product, you don’t know what happens to it. So you might not know what happened to a given token (imagine you’re a musician who's trading your music NFT - shout out to Mali, or you’re making a reality show with tokens to control the actors - shout out to Casey).
There are some scanners and basic analytics tools that can show you which wallets transferred what to whom. However, even as a product owner, you’d have a hard time finding out which people trade the most volume on your platform or with your tokens. These scanners are built around showing individual details, but not insights from big numbers. Also, you’ll have a hard time linking them to your web2.0 frontend because there are no integrated tools, as far as I know.
Also, it’s impossible to track their on-chain interactions through other frontends. Their MetaMask wallet (web3 equivalent Google’s single sign-on) might be connected to your product, but you won’t know they’re up to on other products, without going individually through their on-chain data.
The next step will involving resolving this journey analytics gap between:
Front-end: Google Analytics, Mixpanel
On-chain Data: Etherscan, Dune, Flipside
Ideally, you’d be able to track individual users and cohorts both front-end and on-chain (including actions done outside your wallet, combining the publicly available on-chain data with your private frontend analytics.
Is this a huge hit to anonymity? No. All of web3 is run on Infura and Alchemy, who are kind of the web3 equivalents of AWS and Google Cloud. After all, no one wants to run their own servers to access and write on-chain data, just like no one wants to run their own server to host a website. So these two know everything anyway. They could easily be the Google Analytics of web3 already. The question is - why aren’t they doing this yet and are they planning to?