Flywheel v1.0 Litepaper, part 2
What's broken in crypto marketing and how to fix it
This is part 2 of the Flywheel protocol litepaper (part 1 is here). The protocol itself is launching in late August/early September (the actual ad system is live now, though only partly onchain). Feedback always welcome.
The headlessness (and footlessness) of current Web3 growth
Flywheel isn’t the first growth-oriented protocol in Web3. Layer3 and Boost have built well-thought-out onchain mechanisms for incentivizing users to engage with dapps. Similarly, Fuul (and Spindl) created a referrals product that mimics the original logic of programs like GMX’s.
A protocol isn’t a distribution channel though: a pixel-serving publisher is.
In the case of questing platforms like Layer3 and Rabbithole/Boost, that publisher is their own website (e.g. layer3.xyz). The protocols in question might be very elegant onchain machinery that rewards users for provable actions, but without an actual consumer experience they’re just invisible smart contracts on some chain. New and real users (rather than farmers botting quests) need to have their attention waylaid in some other consumer experience and onboarded to a new dapp. Without additional distribution channels, any questing platform is only as good as its own website, an audience upper bound set by the site’s daily uniques.
That other Web3 mechanism, referrals, is even more headless and distribution-constrained. Most of the referrals platforms like Chainvine are struggling or dead, and the only referral programs that seem to do well (e.g. GMX, Zora) are ones that have fanatically strong community and bake the referral mechanism natively into their protocols. Simply bolting on a referral program to an existing protocol, in our experience of running referral programs, almost never scales to meaningful size.
Quiz: Who’s the biggest referrer on the GMX referral program, probably the biggest such program in Defi?
It’s not an influencer or a whale: it’s CoinGecko. And the link isn’t even underneath a splashy ad or call-to-action, it’s on a completely unremarkable website link on the GMX token page. ($750k in referral revenue to date, on Arbitrum alone.)
So what if Flywheel, in addition to bringing publishers to an ads ‘buy side’ that it enables, also brought permissionless distribution to protocols like Layer3 or GMX’s referral program? The publisher-side ads slot in precisely with the existing referral economics: the publisher, like the referrer, gets paid a cut of user revenue. The only task for Flywheel (or specifically, a publisher partner on the supply-side of the market) is choosing which users and publishers to target. This can even be done mostly permissionlessly, as the sponsoring dapp doesn’t much care where their inbound comes from so long as users convert. Marketing composablity, FTW.
Conversely, the existing ‘Web3’ ad networks like Hype or Slise have no onchain manifestation whatsoever; they’re essentially Web2 style ad networks for crypto-related publishers. They offer little or nothing in the way of onchain measurement, and no economics beyond the risky CPM we’ve had since the late 90s. While such inventory may well make sense for general brand awareness, it doesn’t exploit the rich onchain data of the blockchain for either targeting or measurement. Ultimately, it’s just a rehash of technology that’s dated even by Web2 standards.
Flywheel can help both the headless buy side (quests, referrals, airdrops) and the footless sell side (normie ad networks) by presenting the missing (onchain) half of their respective flywheels. Quests and referral programs can get much greater distribution on more natural consumer touchpoints, while keeping their native economics. Web2 ad networks can aggregate the performance-oriented demand of crypto-native advertisers and (assuming they can convert users) get richly paid for their media.
The glue holding the whole thing together is the shared state of the blockchain coupled with crypto payment rails that seamlessly turn tokens into marketing budgets. That’s it. Spindl (the company) isn’t even essential to the turning of the proverbial (or literal) flywheel here. We simply show that it can be made to spin.
User Privacy
Thus far, we’ve harped on how Web3 is a close parallel to Web2 marketing, rebuilding most of the offchain ads stack onchain (with some concessions to the blockchain). One area where matters are not only dissimilar but inverted is that of user data and privacy.
In Web2, your real identity is online, but your transactions are private; in Web3, only a pseudonymous ‘you’ is out there in the onchain Internet, while every transaction is public and queryable. This inversion creates very different expectations around user privacy, one we’ve built into Flywheel.
(More broadly, we’ve delicately traded off the two privacy worlds inside Spindl measurement via privacy-safe Web2 SDKs. With Flywheel, we hope to avoid offchain personal data altogether.)
In the ideal Web3 ads system, only blockchain data is ever used in all but the most local, client-side decisioning, and even then anonymously (ranking ads by observed clickthrough rates, for example). Flywheel neither expects nor receives any offchain personal data, nor does it drop any cookies or fingerprint any devices. No user data leaves the client side at all, and outside of the publisher partner1, nobody knows anything about the user beyond what’s already on the blockchain. While Spindl, the measurement platform, joins some user session data to wallet addresses for the sake of offchain to onchain attribution, Flywheel never use offchain personal data in any way.
This simplification of the privacy question—really, just building the system along the lines of existing Web3 privacy expectations—is part of why we chose to focus only on crypto-native wallet-aware publishers from day one. We don’t fire pixels on websites, ingest user data, leave traces on the user’s computer or mobile device, or sell data: only what’s onchain is involved.
The offchain part of Flywheel is merely a way to serve pixels somewhere or do some particularly hairy math (like ranking): none of the data or value-transfer situation happens offchain. Everything that matters, from the value of a publisher to the success of an advertiser campaign, are all queryable via Dune. Sunlight is the best disinfectant, wrote Louis Brandeis, and the blockchain provides an unprecedented level of transparency into the often shady business of digital marketing (and perhaps more transparency than some current market participants would like).
We hope Flywheel will surpass Web2 ad tech both in effectiveness and openness. And all that without the often misguided regulation (ahem, GDPR and its endless pop-ups) that have only made matters worse for users and better for monopolistic incumbents. The dream of crypto is to decouple ownership and value from the ‘weary giants of flesh and steel’: governments, banks, and legacy rentiers of various flavors. Flywheel brings that same spirit into the attention economy of marketing. Most of our natively-onchain Spindl clients recoil at the thought of spending money on regular, spammy Web2 banner ads, both for performance and privacy reasons. Time to offer them something better.
If you’re an advertiser or (wallet-aware) publisher who’d like to get involved, please reach out.
For more marketing takes:
Perhaps not even then: publishers can poll the advertising partner’s ads API from the publisher’s server side, such that the advertiser doesn’t get to touch the user’s browser or mobile device. One of the advantages of blockchains is that wallets are global addresses; in Web2, the 50 pixels firing on every website are 50 outside parties reading or writing their pseudonym for you everywhere you go, associated with God knows how much user data, a gross duplication of effort and an intractable privacy problem. Web3 fixes this (actually).