Creating Data Sharing Models That Work (Video)

Bloom co-founder, Jesse Leimgruber, outlines the problems with data sharing and presents solutions to create data sharing models that actually work. Approximately 200 people were in the audience. This event was filmed at the Underscore Blockchain Summit in Boston, MA.


My name is Jesse. I am the co-founder of Bloom. Bloom is a decentralized credit scoring and identity system. We did a token sale in December. And I’m here to talk to you about creating data sharing models that work.

When you think of data sharing, why is this important?

There is data on every single one of you guys right now. You have credit data on you. You have identity data on you. If you’ve ever signed up to an exchange, like Coinbase or Poloniex, you had to verify your identity to show them that you are able to deposit money and they’re able to send you money.

My co-founders of Bloom actually built the compliance and identity verification system that powers companies like Coinbase.

There’s a lot of problems with data sharing in the world. We recently saw the Equifax hack where 145 million Americans had their data leaked. Obviously the Cambridge Analytica case with Facebook, data was leaked without users really knowing what data was out about them.


Risks of Leaks:

Leaks from central parties that are holding your data. They are owning it. It’s clear, plain text often. And they are not always the best at security.

We’ve got silos:

Corporations do not really work well together. And they especially don’t work well together when they have to cross borders. To tie it back to credit, if you’re an immigrant, and you move to the United States, or if you leave from the United States and go somewhere else, your whole credit history resets.It doesn’t matter if you have $50,000 of debt or have paid millions and millions of dollars in mortgages, this data does not talk to each other.

There’s bias in understanding data:

When you trust a central party to make an interpretation of the data, there’s huge biases here.

In China, your political affiliation literally influences your credit score. And in the UK, whether you’re registered to vote influences your credit score. In the United States, we have lots of weird things that influence your credit score like if you’ve checked it or if you’ve closed an old account. And these things are mathematically proven to not actually matter, but because central parties are entrusted to interpret the data, that’s how we go.

Storage: Users don’t own it,

You don’t even know where your data is stored. 50 millions records are stored in Cambridge Analytica.


Central parties love to charge for access to data. That’s their business model.

So, I talked about some of these, but in credit, just because credit is … everyone has data on credit and it’s the first system that Bloom aimed to tackle, there’s risk of identity theft. You’re verifying your identity with plain text social security numbers. There’s cross border issues. Scores are based on historical information. There’s lack of data on a lot of populations, billions of people have no data on them so they’re unscorable. And this leads to lots of unnatural behavior.


What can you do here? If you want to build a data sharing model that actually works, what does that look like? Obviously blockchain has a lot of opportunity. This is why we’re here today. And now is the time. After the Equifax hack, things like the Cambridge Analytica issue, now is the time to really rethink what data and identity actually means.

So, number one, and this is in my view very important, we have to get away from corporations controlling the data, and governments.

1. Users should own your data.

Now, this can take a ton of different forms. People have talked about IPFS, they’ve talked about putting it on your device. But one way or another, users need to be the custodian of their own information. Because if the user controls it, the user can decide how that information is used. This is a screenshot of Bloom. You can set up an identity and you actually own all of your ID information.

2. Limit access and privileges to data.

Right now, there is literally, I guarantee you… Today, probably at least 10 times, your data and your personal information was mined and sold. Facebook makes between $5 and $100 a year per person off of selling your data. Credit bureaus charge, on average, about $15 every single time a lender pulls your credit report. There’s cookies and pixels that are mining your data and your behavior.

Your data is being mined and sold and it’s valuable.

As a US citizen, that data is actually worth hundreds of dollars a year per individual. So, we need to limit the access to that data. We need to create a system that you understand where it’s going and you decide where it goes. On Bloom, we do this through private encryption and hashing. Not all of the personal information is stored and you can selectively release information. If you don’t wanna release and entire credit report, you can just release whether or not you’re capable of paying back a debt or if you have a positive bank balance.

Now, number three is:

3. Allow for low-cost reusability of this data.

There are central parties that charge time and time again to access the same exact information. If you’ve ever used an employment verification service, an education verification service, or any verification service, either you or the company that you’re applying to is paying over and over and over again. But that doesn’t need to be the case. If you own the data, you can have low-cost reusability.

And then last, if we’re gonna have users own the data, and if we’re gonna have users decide how that data gets used, and if we’re gonna make it reusable, we need a good interface.

4. Create a Good Interface

Blockchain right now has a terrible, terrible interface.

Bloom: I’ll be the first to admit it, if you wanna set up a Bloom ID, you need Ethereum, you need Bloom Token, you need Metamask, you need to understand how to even, like somebody mentioned, onboarding Fiat earlier, you need to have Fiat to buy this. It’s a nightmare. It’s like a half hour to an hour process.

We’re working on a mobile application that will streamline the whole onboarding system, no gas fees, no dealing with Ethereum, just using an app like you normally would in the real world, something like Nerd Wallet or any of the other lending credit assessment or any of the lending apps.

To create data sharing models that work:

  • You need to remember that you need to put the user first.
  • You need to work on reusability of information.
  • You need to limit access to that information
  • Get the data out of the hands of central parties.

My name is Jesse. You can reach me at, and that’s my talk. Thank you.

Creating Data Sharing Models That Work (Video)
Share this