Third-party cookies are being deprecated soon, and the digital advertising industry is scrambling to meet user targeting needs with a mix of cookieless solutions.
These cookies have numerous privacy and security concerns, and have led to companies restricting them (Safari, Firefox, Brave) and broad-reaching regulation (CCPA, GDPR). In response, the industry is doing a mix of cohort-based targeting, exploring alternative IDs, and making first-party-data transfer easier.
We don’t believe these alternatives meet brands, publishers, or users’ needs.
Cohort-based targeting, like Google Privacy Sandbox and Facebook’s Interoperable Private Advertising, is less accurate than targeting individual users. This reduced advertising efficiency leads to brands paying more and publishers making less.
Alternative IDs like Prebid’s SharedID or Unified ID 2.0 have not reached feasible scale because users don’t want to share a personal identifier, like their email or phone number, and have no incentive to do so.
First-party data solutions, like LiveRamp’s ATS, allows a website to contribute user data anonymously with the rest of the industry. This has similar privacy concerns to third party cookies and are at risk of regulation.
Brands and publishers need individual targeting at scale, but users don’t want to share personal data.
Replace third party cookies with identifiers that don’t use personal information.
Blockchain wallet addresses are pseudonymous and do not contain personally-identifiable information. They are a permission-less and open standard
Use advanced cryptography to share insights without sharing user data.
Encryption methods like zk-SNARK/STARKs allow wallets to generate proofs of ad engagement, browsing activity, and a specific cohort, while not sharing the underlying data. A blockchain allows for these proofs to be verifiably true.
Address the network cold start problem by providing tokens incentives to early users.
Token emissions can incentivize new participants from advertisers, publishers, and users to be early adopters and bootstrap the multi-sided liquidity needed for an advertising network to function.
- Cookies are replaced with blockchain wallets
- Browser stores ad engagement and browsing history encrypted in this wallet
- Ability to opt into sharing more data (eg. income, zipcode, gender)
- Wallet creation process is TBD. Initially, we may need web3 natives to install a Chrome extension that broadcasts a public address or proof to websites. In the future, this could be built directly into privacy-conscious browsers.
- Ability to easily create a new wallet to reset tracking
- Data management platform (DMP) are replaced with zero-knowledge proofs (zk-proof)
- Instead of sending a cookie per website visit, we send zero-knowledge proofs that provides insights on a user without the underlying data
- Zk-proofs can prove a user has visited this website before, has viewed an ad from this company, or is part of a specific cohort of users - without sharing when they visited the website, which specific ad they saw, or the data that proves a cohort. This source of truth that is more accurate than the ML-based modeling that DMPs rely on.
- These proofs allow any entity in the advertising ecosystem to answer questions about targeting, retargeting, and attribution.
- Brands, publishers, demand side platforms, supply side platforms, ad exchanges remain the same
- They use identifiers instead of cookies, and proofs instead of data management platforms
- In the future, an onchain auction can replace DSPs, SSPs and ad exchanges, by allowing advertisers to directly connect with publishers. For the time being, blockchain finality is too slow, with the fastest being Solana at 400ms, for programmatic ad buying, which is typically <100ms.
Web2 cookies implementation
Web3 desired Behavior
User visits product site A
Third party tracker (eg. Google) adds a cookie via tracking pixel on a product site A
A tracker associated with publisher site A is stored locally on device
Browser creates a blockchain wallet and mints an NFT. Metadata in this NFT is encrypted and stored. Zero knowledge proof is generated that shows that this user has visited product site A.
User visits publisher site B which has shows an ad from product site A.
Cookie is sent when you visit publisher site B. Tracker knows you past history of viewing product site A, and creates a bid request with a lot of other personal data. Bid request is sent to programmatic ad exchange, an ad for product site A is returned. Ad for product site A is shown
When you visit publisher site B, insights on you, without the underlying data, are shared. Publisher, or a data management platform, learns that you’ve been to product site A, but knows nothing else about you. Bid request is based only on relevant insights, rather than your whole history.
Zero knowledge proof is sent to publisher site B. Publisher site B knows enough about the user that they can generate a bid request. Ad exchange returns a targeted ad from product site A. NFT metadata is updated to show that publisher site B showed an ad from product site A.
User clicks ad for product site A.
When you visit product site A, the header in your request include the cookie showing that you came from publisher site B
Data is logged that a user clicked this ad
NFT metadata is updated to show that ad for product site A was clicked from publisher site B.
User makes a purchase on product site A
Third party tracker has code running at your transaction confirmation and attributes the sale to publisher site B (Note - that’s if the cookie hasn’t expired!)
Data about attribution and conversion are readable and verifiable by both product site A and publisher site B
NFT metadata is updated to show that product for product site A was purchased. Zero knowledge proof allows the publisher or product sites to find out this information without having to go through a third-party.
I believe zero-knowledge proofs will reach internet-scale sooner than expected, and that blockchain tokens allow us to bootstrap enough users, brands, and advertisers to overcome the network cold-start problem.
Zero-knowledge proofs have been written off as a practical solution. First, they require significant computing resources to generate a proof. Recent developments reduce the computing resources to a fraction of what was previously needed, such that application-specific zk-rollups can handle 10,000 transactions per second. Second, larger companies want to avoid regulatory scrutiny around crypto like Operation Chokepoint 2.0. I believe broader acceptance to crypto will come, but only a startup with little to lose would take that gamble.
I’ve also seen how blockchain tokens uniquely allow upstarts to challenge incumbents. I saw first-hand how one of our competitors, Blur, unseated the dominant platform, OpenSea, by providing token incentives for users to use their platform. A similar playbook would allow us to onboard advertisers, publishers, and users onto our standard, even without a major tech company partner.