Streaming Trust
Federated identity stores are like music CDs: large troves of data to be consumed from a single source. Self-sovereign identity is like streaming: share just what you want, just when it's needed.
In a recent discussion we had, Marie Wallace shared a wonderful analogy for verifiable credentials. I think it helps with understanding how credentials will be adopted. She compares traditional approaches to identity and newer, decentralized approaches to the move from music CDs to streaming. I'm old enough to remember the hubbub around file sharing. As this short video on Napster shows, the real winner in the fight against piracy was Apple and, ultimately, other streaming services:
Apple changed the business model for online music from one focused on sales of physical goods to one focused on licensing individual tracks. They launched the iPod and the iTunes music store and used their installed user base to bring the music companies to the table. They changed the business model and that ultimately gave birth to the music streaming services we use today.
So, what's this got to do with identity? Most of the online identity services we use today are based on a centralized "identity provider" model. In the consumer space, this is the Social Login model where you use your account at Google, Facebook, or some other service to access some third-party online service. But the same thing is true in the workforce side where each company creates great stores of customer, employee, and partner data that they can use to make authorization decisions or federate with others. These are the CDs of identity.
The analog to streaming is decentralized or self-sovereign identity (SSI). In SSI the source of identity information (e.g., a drivers license bureau, bank, university, etc.) is called the issuer. They issue credentials to a person, called the holder, who carries various digital credentials in a wallet on their phone or laptop. When someone, called the verifier, needs to know something about them, the holder can use one or more credentials to prove attributes about themselves. Instead of large, centralized collections of data that companies can tap at will, the data is streamed to the verifier when it's needed. And only the attributes that are germane to that exchange need to be shared. Cryptography ensures we can have confidence in the payload.
Identity streaming has several benefits:
Confidence in the integrity of the data is increased because of the underlying cryptographic protocols.
Data privacy is increased because only what needs to be shared for a given interaction is transmitted.
Data security is increased because there are fewer large, comprehensive troves of data about people online for hackers to exploit.
The burden of regulatory compliance is reduced since companies need to keep less data around when they know they can get trustworthy information from the holder just in time.
The cost of maintaining, backing up, and managing large troves of identity data goes away.
Access to new data is easier because of the flexibility of just-in-time attribute delivery.
And yet, despite these benefits, moving from big stores of identity data to streaming identity when needed will take time. The big stores already exist. Companies have dedicated enormous resources to building and managing them. They have integrated all their systems with them and depend on them to make important business decisions. And it works. So why change it?
The analogy also identifies the primary driver of adoption: demand. Napster clearly showed that there was demand for online music. Apple fixed the business model problem. And thousands of businesses were born or died on the back of the this change from CDs to streaming.
Digital credentials don't have the same end user demand pull that music does. Music is emotional and the music industry was extracting huge margins by making people buy an $18 CD to get the one song they liked. People will likely love the convenience that verifiable credentials offer and they'll be more secure and private, but that's not driving demand in any appreciable way. I think Riley Hughes, CEO of Trinsic.id, is on to something with his ideas about digital trust ecosystems. Ecosystems that need increased trust and better security are likely to be the real drivers of this transition. That's demand too, but of a different sort: not demand for credentials themselves, but for better models of interaction. After all, people don’t want a drill, they want a hole.
Verifiable data transfer is a powerful idea. But to make it valuable, you need a trust gap. Here's an example of a trust gap: early on, the veracity and security of a web site was a big problem. As a result many people were scared to put their credit card into a web form. The trust gap was that there was no way to link a domain name to a public key. Transport Layer Security (TLS, also known as SSL) uses digital certificates, which link a domain name to a public key (and perhaps other data) in a trustworthy way to plug the gap.
There are clearly ecosystems with trust gaps right now. For example, fraud is a big problem in online banking and ecommerce. Fraud is the symptom of a trust gap between the scam’s target and the legitimate actor that they think they're interacting with. If you can close this gap, then the fraud is eliminated. Once Alice positively knows when she's interacting with her bank and when she's not, she'll be much harder to fool. Passkeys are one solution to this problem. Verifiable credentials are another—one that goes beyond authentication (knowing who you're talking to) to transferring data in a trustworthy way.
In the case of online music, the solution and the demand were both there, but the idea wasn't legitimate in the eyes of the music industry. Apple had the muscle to bring the music industry to the table and help them see the light. They provided much needed legitimacy to the idea online music purchases and, ultimately, streaming. They didn't invent online music, rather they created a viable business model for it and made it valuable. They recognized demand and sought out a new model to fill the demand. Verifiable credentials close trust gaps. And the demand for better ways to prevent fraud and reduce friction is there. What's missing, I think, is that most of the companies looking for solutions don't yet recognize the legitimacy of verifiable credentials.
Got an idea about how verifiable credentials will be legitimized? Leave a comment:
Nice analogy!
New to Substack and found your article through a recommendation. It was written in 2023 and curious what your thoughts are on this today.
In my view, businesses spend enormous effort guessing who people are, while giving them very few mechanisms to directly express who they are.
Much of today's digital economy is built on inferred data. Organizations collect signals and derive assumptions. What if trust ecosystems also allowed people to become first-class publishers of their own interests, intentions and preferences?
Not replacing third-party attestations. Complementing them.
The infrastructure for inference is mature. The infrastructure for self-expression is almost nonexistent.